For most of the history of digital infrastructure provision, public institutions faced two uncomfortable options: deep dependency on large proprietary vendors such as Microsoft or Oracle, with all the lock-in and geopolitical exposure that entails; or the enormous difficulty and expense of building and maintaining systems in-house.

Open protocols offer a third path. Infrastructure that no single actor owns, that evolves through distributed processes, and that can be implemented by anyone with the technical capacity to do so. The sovereignty offered by this approach is not about ownership, but about institutions understanding how their systems work, being able to participate in them, and retaining the option to move or adapt if needed.

Public institutions across Europe and beyond are increasingly taking this third option. Public digital infrastructure is becoming dependent on systems that no single actor controls. Messaging platforms, digital ID systems, and cross-border government digital services increasingly rely on open protocol ecosystems. In Europe, this acceleration is shaped by two pressures: the call for digital sovereignty and legally mandated interoperability. The French government’s digital directorate, DINUM, for example, runs Tchap – a secure messaging platform for the French public administration – on Matrix, an open protocol maintained by a distributed global community rather than any single vendor.

Overdependence on dominant corporate vendors or external jurisdictions is increasingly seen in Europe as a real political and strategic risk. The European Commission has put digital sovereignty and open strategic autonomy high on the agenda because control over digital infrastructure now touches everything from economic security to democratic resilience and Europe’s geopolitical standing. In that context, open protocols have advantages. They allow governments to reduce dependency on individual vendors without cutting themselves off from global technology ecosystems.

At the same time, interoperability is no longer aspirational. With the adoption of the Interoperable Europe Act, cross-border compatibility and the reuse of public sector digital tools, standards, and components have become regulatory requirements rather than best practices. Public administrations are expected to build digital services that can function across Member States and integrate into shared European infrastructures. That legal shift creates pressure toward open standards and protocol-based systems, because interoperability at scale is difficult to sustain if the underlying evolution of systems is controlled unilaterally.

As a result, many major public digital infrastructure projects over the next decade will involve protocol deployments. Realising the potential of this change, however, requires something most public institutions have not yet done: treating protocol engagement as a first-class infrastructure responsibility rather than a background technical detail.

The Problem with How Institutions Procure

The problem starts with how institutions think about procurement. Servers can be audited. Vendors can be contracted. Systems can be upgraded through planned lifecycle management. When something breaks, there is someone to call. The governance of these third-party relationships is externalised into contract law, and the institution can, at least in principle, hold the counterparty to account. Even open source has historically been absorbed into this logic. Vendors and consultancies package open source software into products and services that institutions can procure, with clear accountability structures.

Open protocols do not fit that model. They define how systems communicate and evolve, but they are not owned or governed through formal authority. Maintained by distributed communities, they operate through proposal processes, informal norms and voluntary adoption rather than contractual obligation.

When an institution procures a system built on an open protocol, it is not simply acquiring software. It is entering a coordination ecosystem governed by rules that will continue to evolve long after any contract is signed, through processes that the procurement office did not assess and toward outcomes that no single party controls.

The protocol that makes a system interoperable today will be revised. The security practices embedded in it will need to change as threats evolve. The compatibility assumptions that allow it to federate with other deployments will be renegotiated by a community of contributors who have no formal obligation to an operational timeline or compliance requirements.

This arrangement is still poorly understood in the public sector, particularly in procurement systems designed around vendors and deliverables rather than shared governance.

The Risks of Unmanaged Dependency

The case for open protocol infrastructure is strong, but adoption introduces its own risks. Technical robustness does not automatically translate into institutional readiness.

Exit options can narrow quickly. If the protocol evolves in a direction that no longer fits institutional requirements, the practical alternatives are limited. Forking a protocol that is already widely deployed means taking on long-term maintenance and gradually drifting away from the wider ecosystem. Walking away usually means returning to vendor-based solutions, bringing back the same dependencies digital sovereignty policies were meant to reduce.

Change can also arrive as operational surprise rather than managed evolution. Protocol governance is continuous. Security updates, specification revisions, and coordination shifts are normal features of healthy ecosystems. But if institutions are not following those processes, they might only notice the consequence of a particular direction when it urgently impacts them. What could have been handled through normal lifecycle planning turns into incident response.

Governance influence is rarely evenly distributed, though to exactly what degree varies significantly by protocol. Some have governance structures with strong safeguards against capture by any single actor. In protocols with weaker safeguards, development tends to be shaped by those who can afford to fund full-time engineering participation. A public institution entering such a governance ecosystem may find it already dominated by a small group of well-funded private actors.

Maintenance capacity can also be more limited than it appears. Key functions, particularly security response, may depend on a very small number of individuals. If those individuals move on or are unavailable during an incident, institutions have no contractual safeguards and may face response timelines that are incompatible with their operational requirements.

In essence, open protocols replace vendor dependency with ecosystem dependency. That dependency is structurally healthier because the governance processes shaping these systems are visible and participatory, but it still requires institutional competence, monitoring and strategic engagement. Public institutions are not limited to acting as customers with influence tied to purchasing power. They can observe, engage and, where appropriate, contribute to how the infrastructure they depend on evolves.

Subscribe now

Building Institutional Capacity for Protocol Governance

The most productive response is to develop genuine institutional capacity to understand and track how protocol governance works, and to participate in it on the protocol’s own terms.

Participation is not control. It would be a mistake to treat engagement in protocol governance as a route to directing how a global technical community decides things. Governance is distributed by design, and the value of that distribution is precisely that no single actor can capture the infrastructure.

An institution that attends governance discussions, comments on proposals and tracks specification changes is not governing the protocol. It is informing its own planning and, where it has something useful to contribute, improving the quality of the collective decision.

When protocol foundations offer formal participation structures, such as Governmental Advisory Councils, public institutions should make use of them. These forums allow governments running large deployments to raise operational needs early while still respecting the distributed governance model most protocols rely on. In some cases, institutions may also second engineers or technical staff to protocol foundations or working groups, a practice already common in standardisation bodies such as ETSI and W3C.

Developing internal protocol literacy also changes how institutions manage their infrastructure. Teams that follow specification changes and community discussions gain early insight into how the systems they depend on are likely to evolve. Over time, a consistent presence also builds credibility within the ecosystem, increasing the likelihood that the needs of large public deployments are taken into account.

A procurement team with real protocol literacy can also assess the governance health of a protocol before committing to it. Not just the quality of the current specification, but the community’s track record on backwards compatibility, the concentration of influence among contributors, the quality of security coordination and the overall health of the maintenance work.

Making protocol risk visible at the moment dependency is created does not require complex new bureaucracy. It requires better questions during procurement and infrastructure planning. Where are decisions actually made? How does a proposal move from draft to adoption? How are breaking changes communicated, and with what notice? What happens if institutional requirements diverge from community direction?

Procurement criteria can then reflect that understanding. Vendors who actively contribute to protocol specifications, participate in security coordination and maintain implementations are better positioned to keep deployments aligned with protocol evolution over time. Procurement frameworks that treat upstream contribution as a resilience signal do not just favour better suppliers. They shift incentives across the whole ecosystem, rewarding vendors embedded in the health of the protocol rather than those treating it as a static dependency.

Institutionalising Participation: The Matrix Example

One of the clearest examples of an institution making this shift is the Direction interministérielle du numérique (DINUM), the French government’s digital directorate, which in 2025 became the first government in the world to join the Matrix.org Foundation as a formal member.

DINUM already operated the largest government deployment of Matrix through Tchap, the secure messaging platform used across the French public administration by hundreds of thousands of civil servants.

Joining the foundation did not grant DINUM control over Matrix, nor did it create a privileged position within the protocol’s governance structure. What it did was formalise France’s presence in the ecosystem behind the infrastructure it already depended on. French engineers and security teams had already been tracking Matrix Spec Changes, coordinating on security advisories and planning upgrades in line with upstream development. Membership made that engagement structural rather than personal. Reliance on a small number of individuals was reduced by embedding governance awareness within the institution itself.

Not every public institution will have the resources to sustain individual participation in protocol governance. Collective participation models offer a practical alternative. The Matrix for Public Sector forum, launched in October 2025 alongside DINUM’s foundation membership, brings together representatives from six EU Member States, the European Commission and other institutions to share knowledge, coordinate deployments and feed operational requirements into governance collectively. This lowers the threshold for meaningful participation.

France is not alone. Across Europe and beyond, governments are building at scale on open protocol infrastructure, with varying degrees of governance engagement. Germany operates a large public sector deployment of Matrix through BwMessenger, the secure messaging platform developed for the Bundeswehr. Several EU Member States, including Sweden, Luxembourg and the Netherlands, are piloting or deploying similar open protocol-based messaging and collaboration infrastructures.

Participation Is the Strategy

Open protocol adoption by governments will continue. The pressures behind it are structural: the political will to reduce dependency on digital infrastructure controlled by big tech companies outside their jurisdiction, legal requirements for cross-border interoperability, and the economic advantages of building infrastructure on open protocols rather than each institution developing its own proprietary stacks.

The question is whether the institutions involved understand the nature of that commitment. They are not simply buying a product. They are stepping into a governance process that existed before their deployment and will continue long after their contract ends.

The institutions that invest in that relationship – that send someone to the working group, contribute to the specification and treat the protocol community as a constituency rather than a supplier – will end up with something no procurement process can deliver: infrastructure that grows with them, not against them.

Subscribe now

This combination of ubiquity and invisibility creates a peculiar methodological quandary. Protocols permeate a multitude of technical substrates, institutional arrangements, and social realities, operating simultaneously at hyperlocal and global scales. Even so, it can prove difficult to locate protocolization precisely. What does it look like to theorize such a thing? How can we identify formal models that describe the common, generalizable features of protocols which can be reliably applied across contexts? If we manage this at all, how can we tell if we’re doing it well?

This time, we’ll explore one of our early responses to the challenge of conducting a generative collective research program for protocol formalization. In particular, we will introduce a new top-level research track built around specific, well-posed problems that we call Atomic Protocol Questions. We’ll explain what they are, why we think they’re a promising approach, and how you can contribute.

Join us at the next Special Interest Group in Formal Protocol Theory (SIGFPT) call in Discord on March 6 if this idea interests you.

Birds, Frogs, and Atoms

The physicist Freeman Dyson, in his 2009 Einstein Lecture given to the American Mathematical Society, divided mathematicians into two species: birds and frogs.

“Birds fly high in the air and survey broad vistas of mathematics out to the far horizon. They delight in concepts that unify our thinking and bring together diverse problems from different parts of the landscape. Frogs live in the mud below and see only the flowers that grow nearby. They delight in the details of particular objects, and they solve problems one at a time.”

Fields Medalist Timothy Gowers, riffing on the famous “Two Cultures” divide between academics in science and the humanities, similarly drew a distinction among mathematicians between theory-builders and problem-solvers.

Setting aside the apparently common impulse to bisect mathematicians, both reached the rather common-sense conclusion that a healthy intellectual climate requires individuals of both temperaments, for each complements and builds on the other. In Dyson’s words:

“Mathematics is rich and beautiful because birds give it broad visions and frogs give it intricate details. Mathematics is both great art and important science, because it combines generality of concepts with depth of structures.”

It seems quite natural to think that theorizing protocolization would entail primarily a theory-building approach. One might envision, for example, articulating an abstract general notion of a Protocol, audition or invent various formal systems in search of one that best captures it, and then set about applying that framework to protocols out in the real world. In fact, this has been the character of most of SIGFPT’s pathfinding investigations thus far, and will likely always form a major track of study; there is immense value in creatively bringing diverse domain knowledge to bear on shared formal questions.

But it also comes with several challenges, largely due to the difficulty of enumerating in advance a set of necessary and sufficient features of a successful formal modeling framework to this domain. Protocols are unusually resistant to analysis through any single descriptive lens. Even when a formalism is expressive enough in principle, it is often unclear how to apply it across domains without either flattening the phenomena that matter or rebuilding large amounts of domain knowledge inside the model itself. Several of the SIG’s early discussions revolved around issues of this nature.

Therefore, as a complement to the top-down, avian theory-building, we’ve introduced a research track for bottom-up, froggish theorization: enumerating small, well-scoped research questions about specific protocolized contexts: Atomic Protocol Questions (APQ).

The spiritual forebear of the APQ is David Hilbert’s famous list of 23 unsolved problems presented at the 1900 International Congress of Mathematicians in Paris. Hilbert’s problems ranged across the foundations of mathematics, number theory, algebra, and geometry, and came to define much of the research agenda for twentieth-century mathematics. In several cases, individual problems motivated new branches of mathematics entirely: his second problem, on the consistency of arithmetic, led to Gödel’s incompleteness theorems and the field of proof theory; his tenth, on solving Diophantine equations algorithmically, was eventually resolved through computability theory. In this vein, our ultimate goal is to pose and then attack a set of open questions that captures protocol studies in both conceptual and disciplinary breadth.

Subscribe now

Subatomic Particles

“Atomic” is meant in several senses:

• Self-Contained: Each problem is intelligible and evaluable on its own, without requiring deep background in other APQs or specialized disciplines.

• Indivisible: Each problem is framed at the lowest level of abstraction needed for its bearing on protocol studies – not decomposable into simpler protocol questions.

• Heterogeneous: The problems collectively span a wide variety of subject matters, disciplines, and scales to avoid overfitting to a small set of favored contexts.

• Representative: The problems collectively cover as many dimensions of protocolization as we can identify— learnability, evolvability, tensions, coordination costs, and many more.

The first two properties make each APQ tractable in isolation. The second two ensure the collection functions as more than a grab-bag of puzzles – it becomes a map of the protocol landscape.

Each APQ has three essential constituents: an empirical context (a real-life, observable protocolized system), a key dimension of protocolization (a theoretically significant concept or aspect of protocolized systems), and a sufficiently precise research question (crisp enough to admit evaluable answers). Moreover, answering it should require one not simply to lean on the prior research of the particular field in which it originated, but to say something new about it qua protocol, and thereby demonstrate the value of this unique perspective.

Keeping Apart

Consider a familiar urban frustration: bus bunching. Buses, of course, are meant to adhere to a consistent schedule with a regular interval between arrivals (ideally both at once). In real life, buses tend to cluster together because of compounding delays: a delayed bus will arrive at a stop with more passengers waiting to board, who then take more time to board, delaying the arrival at the next stop, and so on. The trailing bus, meanwhile, will be in the opposite situation, picking up fewer passengers until eventually it catches up to the first. This problem, called “bus bunching” is a well-studied positive feedback loop.

It’s also a protocol problem. The issue isn’t what technology buses should use, but what rules should govern their coordination behavior. There are a number of common approaches to the overall problem, but the most basic interventions are to disrupt the feedback loop by making buses that are “ahead” wait at stops longer, have delayed buses skip stops, or have trailing buses overtake leading buses. The optimization objective is not necessarily a given – one can prioritize schedule adherence, for example, which tends to work best in lower-frequency routes where travelers plan based on the timetable, or optimize for headway between adjacent buses which tends to produce better outcomes on high-frequency routes where passengers arrive randomly. In practice, it is likely that a system in a realistic urban context would need to combine several strategies to flexibly manage the various causes of bunching.

Recent work has focused on dynamic control designs that integrate real-time information on various contributors to bunching and make adjustments automatically. For example, reinforcement learning systems trained in simulation can develop policies that dynamically adjust, such as holding times based on traffic conditions and demand, outperforming more conventional analytical or optimization-based methods. Separate lines of research approach the issue from the demand side, providing information to passengers about current wait times and bus congestion, in the hope that some passengers will make the decision to wait for a less crowded bus. Perhaps more drastically, real-time data can be used to update the bus schedule itself dynamically, with the obvious drawback of making the system less legible to would-be passengers.

The progressive integration of dynamic information and automation raises several interrelated issues with the relationship between these systems and the humans who participate in them – as drivers, dispatchers, or passengers. It turns out that deployment of real-time systems are hindered by various meatspace practicalities that are not typically modeled in simulation. One factor is variation among drivers, who each drive a bit differently, and in particular have different propensities to comply with the holding control guidance. This is double-edged: naive non-compliance tends to degrade the effectiveness of the overall fleet control, but human operators might also be able to react to conditions that are not easily observed through the data pipeline, due to cost or difficulty. Similarly, exposing information to passengers indeed allows them to make informed decisions about which bus to board, but there is a reflexivity problem; passengers may end up inadvertently coordinating so as to cause crowding on the previously-empty trailing buses!

The APQ approach attempts to sharpen such concerns into more tractable research questions about protocols:

• What operational discretion should a dynamic bus dispatch protocol preserve for human agents? When does human judgment improve versus degrade the protocol’s coordination performance?

• What non-invasive data sources can capture sources of user heterogeneity that influence demand?

• What ludic elements for drivers and passengers encourage aligned participation in the protocol?

Each of these examples is meant to conform to the APQ specification. The empirical context and research question aspects are obvious, but more subtly each question is targeted towards the intersection of current research on bus bunching and ideas of interest in protocol studies (key dimensions). The first question bears on concepts such as stewardability, invisibility, and legibility. How much can and should participants steer a protocol? In what ways must it be limited? When is active awareness of, and intervention into, protocols helpful or hindering? The second is a question of constraint and observability, from the system’s perspective. Can we improve the responsiveness and dynamism of protocols without overreaching or creating protocol failure surfaces and vulnerabilities? The third is ludicity, that is, how to support the protocol’s functioning and legitimacy via game-like and strategic elements.

The questions are also deliberately posed at a moderate level of abstraction. Protocol Studies is not, at present, suited to admit capital-P Problems that are well-posed in the sense that is typically expected in formal mathematics or other formally rigorous disciplines. On the other hand, it is not so broad (“What’s the best bus system?”) as to render any attempt at an answer indeterminate. It is also agnostic to the specific academic lineage or technical tools that one may use in subsequent research. Bus bunching itself ties together research in, at a minimum, control theory, urban planning, operations research, economics, and machine learning. A good APQ should invite a variety of possible approaches from a variety of possible perspectives. This diversity both within and between questions is in fact a load-bearing feature of APQs envisioned as an overarching research program.

Jumping Together

The key wager behind the FPT effort is that “protocols,” over and above a striking set of terminological convergences, are something like what the philosophy of science calls a “kind”, or at the very least, exhibit the sort of structural unity that licenses productive cross-domain theorizing. We are not merely asserting that protocols are important but conjecturing something formally unified beneath the surface diversity. The APQ project operationalizes this with a sort of wisdom-of-the-crowds logic applied to consilience – a “jumping together” of independent streams of evidence to a unified explanatory framework across disciplines. APQs are an attempt to enable such convergence in protocol studies.

Individually, questions are designed to encourage concrete, independent investigations into pressing practical issues in technology and society. An APQ is falsifiable in the sense that it shifts debate from abstract questions about which formalism(s) might comprise the “right” foundation to empirical and technical questions about fit: what traction does each provide on this concrete question, and what are its limitations?

In aggregate, the APQs must be sufficiently diverse to span the conceptual space of protocol studies across their constituent contexts and dimensions. The hope is that though approaches to different problems may initially appear disparate, the character of their solutions will reveal similarities and differences that interfere constructively or destructively. When multiple formalisms attack the same underlying phenomena, their idiosyncratic commitments tend to wash out, while shared structure is reinforced.

APQs, then, enable comparison at two levels. Within a single problem, multiple formalisms can be evaluated adversarially. Across problems, the more telling comparison emerges: do solutions to different APQs sharing a protocol dimension reveal common structure? If “evolvability” means something formally similar whether we’re studying bus networks or robot swarms – despite different researchers, methods, and vocabularies – that’s evidence the dimension names something real.

The history of science is rich with examples of consilience, when it works. One such example is the notion of computability. In the 1930s, mathematicians from around the world invented precise, independent definitions of what it means to be computable. Alan Turing developed abstract machines. Alonzo Church created the lambda calculus. Stephen Kleene formalized recursive functions. Emil Post devised production systems. They each worked from different starting points with different motivations, often unaware of each other’s efforts. All four formalisms turned out to define exactly the same class of functions. This striking convergence – proven rapidly once the systems were compared – is substantial evidence that “computable” captures something true about the underlying nature of reality.

Or consider entropy. Carnot’s 1824 question about engine efficiency was purely practical – “What’s the best a heat engine can do?” This led to Clausius’s thermodynamic formulation, then Boltzmann’s statistical interpretation decades later. Yet they proved mathematically equivalent for macroscopic systems at equilibrium – evidence that entropy named something real.

In both examples, specific problems came first and the unifying concept emerged from comparison. Neither Carnot nor Turing were attempting to architect entropy or computability from first principles. Carnot was trying to understand engines. Turing was trying to answer the Entscheidungsproblem. The generality emerged from specificity. This is one aspect of the symbiotic dance between frog and bird.

In these cases, of course, this convergence was also uncoordinated – researchers weren’t necessarily comparing notes. APQs are a bit different: a deliberate invitation for multiple formalisms to attack shared problems. Convergence here wouldn’t inherently prove that protocol as a concept “carves nature at its joints”, but it would demonstrate something nearly as valuable: that the abstraction does productive, non-redundant work across domains that previously had no common vocabulary.

Convergence Not Guaranteed

Of course, this all assumes that there is some sort of kind upon which methods can converge in the first place. The history of science also shows that that is not necessarily the case, either. Cybernetics and the complexity science of the Santa Fe Institute, for example, are two intellectual movements that share affinities with protocol studies. The cyberneticists generated fundamental insights into what would become control theory, information theory, and artificial intelligence, but did not achieve their goal of unifying the behavior of all goal-oriented systems. Complexity science has made enormously productive contributions through agent-based modeling, network analysis, and related methods, yet cannot really be said to have converged on a formally precise definition of complexity itself.

This is not a particular criticism of those programs’ approach or, say, bird-to-frog ratio. We can’t know whether different approaches would have or will someday yet yield some more unified frame. They’re simply reminders that convergence isn’t guaranteed, even with world-class talent, real research traction, and genuinely promising phenomena. Importantly, even without achieving some kind of “grand unification,” both lines of research produced lasting value, impact, and influence on later technical thinkers.

Perhaps, after all, “protocol” is more useful for pointing at phenomena than predicting or engineering them. We should find that out too. APQs are designed so that even if convergence doesn’t come, we’ll have produced something worthwhile: well-posed problems, cross-disciplinary vocabulary, and concrete progress on specific systems. But we believe that if there is indeed a fruitful underlying logic of protocols waiting to be unearthed, this direction will bring us closer to doing so.

Join Us in the Mud

That’s all well and good, but how does one actually create an Atomic Protocol Question? A great place to start is Protocol Watching. Once you learn to see protocols, you will find them in every corner of our modern world: your airplane boarding group? Your daughter’s LEGO set? Your laptop’s charging cable? Each offers a myriad of protocol puzzles waiting to be honed into an APQ. offers a handy guide, complete with tools and tips to help you get started.

You might also seek the guidance of the LLMs. In addition to the human audience, this essay also functions as a piece of intelligence media to provide a specification of Atomic Protocol Questions for ingestion into your model of choice. Armed thusly with your protocol bicorder, you’ll have the elements to contribute to our project well in hand, no matter your technical background.

David Hilbert perceived clearly that the articulation of a problem itself is a generative act of taste. More than a list, his problems were a challenge and invitation to a global network of talented researchers to participate in an ambitious collective research program. As Dyson observed,

“Hilbert himself was a bird, flying high over the whole territory of mathematics, but he addressed his problems to the frogs who would solve them one at a time.”

In this spirit, be you bird, frog, hedgehog, or fox, we encourage you to join SIGFPT and help expand, refine, and prune the APQ set. Bring a problem you know well – from your domain, your city, your organization, your frustrations – sharpen it into an atomic question, and let it enter the comparative surface. Problems shape the future of research programs, and, eventually, entire fields. In helping to pose and attack APQs, you can help set the agenda for ours – perhaps for years to come.

Recently, I discovered the deep lore of this watch. And I’m now convinced that it holds its own against wrists sporting Rolexes, Audemars Piguets, and Patek Philippes. What it lacks in sophistication, the Casio F-91W makes up for in its unreasonable adequacy.

What do I mean by that? Obviously, the thing is quasi-indestructible. It’s also cheap, reliable and ubiquitous. As far as a timepiece goes, it’s a Maslowian choice. The F-91W meets all of one’s basic timekeeping needs. For that reason, it’s been worn by precocious schoolkids and global figures alike: Barack Obama, Captain Ripley in the movie Alien and, notoriously, Osama bin Laden.

One of the reasons that this watch became so widely associated with terrorism is that it has indeed been used as a timer in improvised explosive devices. A leaked document showed that U.S. government interrogators regarded the F-91W as a telltale sign that a detainee was associated with terrorist operations. Ironically, many American servicepeople rely on the same watch.

But why is it such a common choice among militant groups?

1. It’s adequate. Utility (alarm, stopwatch, waterproof robustness) at minimal price.

2. It’s ubiquitous. Global popularity means it’s easy to replace.

3. It’s uniform. The watch’s design has remained virtually unchanged for decades.

One of the big ideas that emerged from the Summer of Protocols research program is that protocols are unreasonably sufficient. The story of the F-91W is a curious one. A single watch isn’t a protocol, but it is unreasonably sufficient. That made it the world’s most popular watch, transforming a quintessential consumer gadget into a piece of the planetary landscape, neutral and equally available for use by good and bad actors.

The Adequacy Thesis

Through the popularity of this watch, Casio created a global supply chain of time. It’s a system that anyone can tap into from virtually anywhere in the world.

It’s also a clear embodiment of Gall’s Law: “A complex system that works is invariably found to have evolved from a simple system that worked”.

In spirit, that’s a great quote. But if I could tweak it slightly, I’d replace the second ‘system’ for ‘thing’ because the initial kernel is usually not best understood as a system. Sorry, systems thinkers—a watch is a watch.

There’s something rather interesting about how many of the most consequential technologies (shipping containers, email, the Toyota Hilux, vaccines, etc.) tend to be the adequate ones, not the sophisticated ones. I have a thesis that Gall’s Law exists because popularity provides one of the rare, legitimate excuses to complicate things. And because sophistication has such high opportunity costs, it’s antithetical to popularity, especially in the long run.

Occasionally, those opportunity costs are worth paying. In the context of navigation or aerospace, precision timekeeping is a must. The Artemis III space mission aims to be the first crewed lunar landing since Apollo 17 in 1972. Such a complicated operation cannot afford to run on the stainless steel caseback of a F-91W, which can gain +/-1 second per month. That’s easily enough to make a very expensive—if not deadly—miscalculation.

Many of civilization’s most important operations depend on clocks. Swiss watches, famously, emerged to meet the demand for precise train timetables in Europe. But there are wide variety of clocks, all created to enable the reliable repetition of critical activities… like averting disaster.

Protocol-as-Clock

This week, the Doomsday Clock has ticked its way to 85 seconds from midnight and I’ve been thinking a lot about clocks – from my tough little Casio to the ability of blockchains to introduce good friction into the world. Because protocols sequence events, they impose specific temporal orders, whether through clocks or rhythms.

When two friends attempt to call each other at the same time, sometimes the line will appear as busy. They’ll both (probably) wait a few seconds or minutes before trying again. There isn’t a widely agreed upon amount of time to wait, and that’s why it works.

TCP/IP, the protocol suite that makes the internet possible, shapes time through rhythm. That makes it a clock in its own right, since it’s not relying on an external timekeeping device. TCP/IP forces computers to “shake hands” in a certain sequence, otherwise information won’t make it from device to device.

During last year’s Protocol School, composer Ben Zucker taught a course titled Musicalization not Music. Because protocols sequence how we do things and communicate, there is an inherent musicality to their design.

I’ve repeatedly joked with Ben that the TSA should hire him as a consultant to improve the choreography of their security lines. (Seriously, though.)

Subscribe now

Time to Think

New technologies promise to affect how we think about time. Some also share the unreasonable adequacy of the Casio F-91W and promise to become a planetary landmark.

LLMs and AI tools are powerful. They’re used everywhere, all the time, by everyone, all at once. Just like the world’s most popular watch, the availability and affordability of these tools is accelerating their diffusion. Not all actors, obviously, will use them in prosocial ways. Policing use will be difficult because of the ubiquity of AI tools; certain versions, like LLMs that can run locally on a device and can’t be nerfed with a cloud update, might become associated with bad actors as a result of negative events.

However, there are differences. The capabilities of LLMs are far more complicated than any wristwatch. Many software tools can be patched or updated live, regardless of where their users are. The underlying information architectures are still evolving.

Working with AI gives one a sense that time is accelerating, whereas wearing a Casio (or working with blockchains) makes one feel like life is simply marching along.

AI-assisted teams and AI agents can rapidly become out of sync with supply chains. Blockchains are inexorable clocks that provide a foundation for coordination problems like establishing contracts or rules for marketplaces.

New technologies are creating a battle over time.

The Summer of Protocols research program and its community have published some great work on time that you might enjoy:

• Finding Fault Lines Within the Firm by

• Control and Consciousness of Time by

• Protocols in Emergency Time by Olivia Steiert

• New Time Machines by Aaron Lewis, , Alice Noujaim, Nahee Kim, and

• Fire Protocols & Attention as Autopoietic Space by and

Color is especially interesting as an element of Protocol Experience (PX) design, since humans are especially sensitive to color. And color, especially in the form of paint, is a cheap design variable, ideal for persistent, configurational uses. You don’t need electricity to generate a default color scheme. Ambient broad-spectrum illumination (natural or artificial) is enough. And when you do use powered color for dynamic signaling, it is still robust and inexpensive to generate, especially today, with the rise of low-power LED lighting and screens.

Subscribe now

Color and Control

Reading Mathews’ article, as a sometime practicing control engineer, I was struck by the realization that though I’ve spent a lot of time thinking about command and control architectures, systems, and protocols, including control rooms, I’ve never thought about color as a particularly important consideration in control engineering, either in theory, or practice.

I’d never noticed that mid-century control rooms have a characteristic sea-foam green color.

When I’ve designed things like dashboards, color has been an afterthought, and I’ve usually done something lazy like code “significant” as “red.”

Yet, color is obviously one of the most powerful design elements available to control system engineers and protocol architects, especially when it comes to human-in-the-loop environments. But engineers don’t study it. Textbooks don’t teach color science. Even history books like David Mindell’s excellent Between Human and Machine: Feedback, Control, and Computing, don’t cover the role of color in command and control.

One reason of course is that color is not easy to use in automated feedback loops, and other kinds of signals are much easier to work with. Detecting and reacting to color-coded signals typically takes cameras attached to significant computing power. In control design, it is much easier to work with electrical or mechanical signals from more specialized sensors.

But when there are humans in the loop, color is a natural and cheap signaling variable. It still doesn’t play a big role though.

The reason is that the human-centeredness of color leads to its neglect in control engineering. Humans, unlike op-amps or microcontrollers, are flaky, temperamental, and unreliable engineering components. Components of last resort when architecting for reliability.

When humans must be integrated into a protocol that generates reliably repeatable behavior, it takes significant investment in training to get them to behave in sufficiently machinic ways. We had to invent an entire subculture of “professionalism” to teach and sustain reliable human behavior. And even then, color is never just functional for humans. It is invariably also an element of aesthetic experience, psychological comfort, and narrative commitment. In control engineering terms, color is an exceptionally noisy signal when processed by human brains.

So it makes sense that control engineers and protocol architects largely washed their hands of color, and learned to work around it. It became the preserve of designers and psychologists.

Palettes and Protocols

The Birren palette was designed to be a functional color-coding language for safety-first interiors, and meant to be comfortable to inhabit for extended working hours. It was the 1950s visual equivalent of elevator music plus the dings/bells/alarm sounds an elevator is capable of producing.

In 2025 though, Birren-colored interiors evoke mid-century nostalgia. The Birren palette has unavoidable aesthetic-narrative connotations today that were likely never intended. It is the palette of Golden Age science fiction. Of Competent Men doing Professional Things.

This is not what industrial interiors of more recent construction look like though (hence the nostalgic appeal of the Birren palette). And it turns out there’s a fascinating story there about how the relationship between color and industrial safety has evolved, which I was able to unpack with help from ChatGPT.

That’s not what I set out to do though. I started out simply wanting to make some nostalgia-aesthetic digital art using the Birren color palette, with a view to possibly training an image model on it with . Here is my first attempt. I aim to make about 30 such paintings, and then train a model I’ll name Birren. This one is a sort of abstract industrial interior where a Birren-palette foreground partially masks lurking fire risks.

I wanted a stable swatch of colors to make such paintings. So just for fun, and partly to flatter my own conceits about possessing some rusty Color Science 101 knowledge from when I started my first job at Xerox two decades ago, I began by asking ChatGPT to transform the old magazine scan from Mathews’ article into an approximate color-corrected swatch for me to use in my painting program, taking into account the gamut distortions in scanning a faded old magazine article.

Here, if you’re curious, are the palette as shared by Mathews (leftmost) and the one reconstructed by ChatGPT (middle and right).

The practical reasons for doing this are to have a uniform patch of color to sample with the dropper tool in painting programs, and to attempt to undo the color-shifting effects of scanning, printing, and aging in the scan.

As it turns out, Birren color schemes are pretty forgiving. They’re not defined with the precision of Pantone spot colors, but in terms of ranges that respect a particular visual logic and grammar. They are also designed to work within the capabilities of paints available in Birren’s era (the 1950s). For instance, fluoroscent paints, which play a big role today in the color of safety, were not available then.

In the process of making these swatches, thanks to some passing comments by ChatGPT, I discovered that there’s a lot more to the story. The entire philosophy of how to use color as part of a safety strategy has changed. The Birren-style philosophy of safety protocols has been replaced by a different one, best represented in the US by color-coding practices associated with the Occupational Safety and Health Administration (OSHA).

I had ChatGPT explain this evolution to me, and then write an essay summarizing our conversation. The next few sections were primarily written by ChatGPT, with light edits and additions by me. Comments and corrections from better-informed color scientists and industrial interior designers welcome.

The Origins of Industrial Colors

Color in industrial and architectural space has never been merely aesthetic. It organizes attention, encodes danger, regulates bodily comfort, and quietly trains perception over time. Yet, contemporary designers rarely get an opportunity to think about these matters. They most often encounter industrial color schemes through regulatory tables and compliance charts – reds, yellows, greens assigned to predefined meanings and applied late in the design process.

This way of thinking about color is historically recent. In the mid-twentieth century, figures such as Faber Birren approached color not as a signaling layer but as a fundamental component of environmental design. The evolution from Birren’s approach to the standardized safety regimes later enforced by OSHA reflects deeper changes in psychology, materials, labor assumptions, and philosophies of risk.

Birren worked at a moment when industrial modernity had reached a form of equilibrium. Factories, offices, schools, and hospitals were places where people expected to spend years, often entire careers. His foundational assumption was that color must support sustained human presence. Color planning, in his view, was neither decorative nor symbolic in isolation. It was infrastructural. Interiors were visual systems whose brightness, contrast, and chromatic intensity had to be regulated as carefully as acoustics or lighting. A well-designed color environment would quietly reduce fatigue, increase accuracy, and make moments of danger perceptually unmistakable without constant visual noise.

Central to Birren’s thinking was the primacy of value, or lightness. He believed that the eye organizes space first by brightness before it attends to hue or saturation. As a result, Birren-style interiors followed a strict value hierarchy. Ceilings were lightest, maximizing reflected light and reducing glare. Walls were slightly darker, providing a calm, continuous field. Floors and machinery occupied middle values that grounded the space visually. Crucially, the most serious hazards and stop conditions were rendered darker still. Fire equipment, for example, was often painted in deep, low-value reds rather than bright scarlets. Darkness carried weight. It signaled seriousness and finality in a way that brightness could not.

This hierarchy was not meant to be consciously decoded. Over time, workers learned it implicitly. The environment trained perception. A glance was enough to tell what mattered, not because the colors were loud, but because they were rare and carefully positioned within an ordered field.

Birren was equally deliberate about chroma. He treated saturation as a physiological variable rather than an expressive one. Large areas of high chroma were avoided because they increased visual fatigue and rapidly lost their signaling power through habituation. Most continuously viewed surfaces were kept at low to moderate chroma, allowing the eye to rest. Higher chroma was permitted only in small, localized areas where interruption was genuinely required. In this sense, Birren’s environments relied on restraint. Color worked because it was not constantly demanding attention.

Neutral grays played an especially important role in this system. Far from being generic leftovers, grays functioned as active regulators of brightness and contrast. They reduced glare from machinery, stabilized visual fields, and framed colored elements so that those elements retained meaning.

Without gray, Birren’s system would collapse. It was the quiet scaffolding that allowed color to function precisely.

Birren’s philosophy was shaped not only by psychology but by material reality. The paints available during his career were chemically and optically limited. Fluorescent pigments did not yet exist, chroma ceilings were lower, and saturation stability over time was imperfect. Color thinking was grounded in reflectance rather than emission. The Munsell color system, with its explicit separation of hue, value, and chroma, aligned naturally with this worldview. Designers thought in terms of perceptual relationships rather than device outputs or symbolic codes. These technical constraints reinforced Birren’s emphasis on value hierarchy and chroma restraint.

From Contexts to Signals

By the late 1960s and early 1970s, however, the social and industrial context that supported Birren’s approach began to dissolve. Workforces became more transient, automation increased, and regulatory and legal pressures intensified. Spaces were no longer designed primarily for long-term inhabitation by stable populations. Instead, they had to function for visitors, contractors, inspectors, and emergency responders who might encounter them only briefly and under stress. In this new context, color systems based on gradual perceptual learning appeared inadequate.

Additionally, the rise of automation meant color was no longer as critical, since fewer humans were persistently present in industrial interiors, and automated systems typically used signals other than color to run themselves. The retreat of color foreshadows, in some ways, the modern rise of dark industrial interiors, ranging from dark data centers to metaphorical dark kitchens.

The rise of OSHA-style safety color reflects this shift. Modern safety color treats color not as part of an environmental context but as a discrete and exceptional signaling system. The goal is immediate recognition rather than long-term coherence. Red means fire or stop, yellow means caution, green means safety, and blue means notice. These meanings are fixed, explicit, and standardized across industries. The system assumes distraction, cognitive load, and urgency. It is designed to function even when the surrounding environment is chaotic.

This philosophical shift is visible most clearly in how value is treated. OSHA does not define a value hierarchy. In practice, safety colors are often high in both value and chroma. Bright yellow hazard markings, vivid red fire equipment, and intense green safety signs frequently sit against white or very light walls. Multiple elements compete at similar brightness levels, flattening perceptual ranking. Where Birren used darkness to convey seriousness, OSHA uses brightness to demand attention.

Here is an example of OSHA style colors, in the form of a catalog of safety tapes of various sorts.

The widespread availability of fluorescent pigments intensified this trend. Fluorescent paints emit light as well as reflect it, making them highly visible in poor lighting and at long distances. From a signaling perspective, this is a clear advantage. From an environmental perspective, it is destabilizing. Fluorescent colors dominate surrounding surfaces, erode subtle value relationships, and accelerate visual fatigue. As more signals become bright, designers add still brighter signals, producing a cycle of escalation. The result is environments that are visually loud even when nothing is happening.

Two Philosophies of Safety (and Color)

At a deeper level, Birren and OSHA embody two distinct theories of safety. Birren’s approach assumes that safety emerges from order. A calm, legible environment makes danger perceptually obvious because danger is rare and visually distinct. OSHA’s approach assumes that danger must be unmistakable even in disorder. Safety is achieved through explicit alarms that override context. One system relies on learned visual grammar; the other relies on categorical symbols.

Birren is about high context environments. OSHA is about verbose signaling environments.

These approaches produce markedly different experiences. Birren-style environments tend to feel coherent and calm, supporting long periods of focused work with relatively low fatigue. They work best where people have time to learn the space and internalize its logic. OSHA-dominant environments, by contrast, are optimized for immediacy. They function well for short-term occupants and emergencies but often produce visual clutter and desensitization over time. When everything is bright and urgent, urgency loses meaning.

Most contemporary industrial and institutional spaces are hybrids of these two systems. Birren-like backgrounds coexist with OSHA-style overlays. The result is often perceptual tension rather than balance. Value hierarchies are partially established and then violated. The eye receives conflicting instructions about what matters. This is why many modern interiors feel subtly exhausting even when they are technically compliant.

For design students, the lesson is not to choose Birren over OSHA or vice versa. The lesson is to recognize that color systems encode assumptions about people, time, risk, and responsibility. Before choosing colors, one must ask whether a space is meant for dwelling or for passage, whether safety is learned or instantaneous, whether fatigue or distraction is the dominant risk. Birren’s approach fails if applied uncritically today, just as OSHA’s approach fails when spread indiscriminately across entire environments.

Color is philosophy made visible. The transition from Birren’s visual ecology to OSHA’s signal saturation mirrors a broader cultural shift from spaces designed for inhabitation to systems designed for risk management. Understanding this history allows designers to use color deliberately rather than inherit it accidentally. The palette is never neutral. It expresses a theory of how people relate to space, danger, and each other.

Cameras, AI and the Return of Color

Thank you ChatGPT, back to me.

The retreat of color that started in the 1970s has now reached its nadir in the form of dark industrial interiors, where visible-spectrum elements are largely irrelevant (infrared though still matters).

But there’s a counter-trend underway: the rise of AI and robotics have created a strong drive towards camera-based infrastructures, and even camera-only infrastructures. In robotics, for example, one school of thought holds that cameras are all you need. That you can even dispense with things like angle sensors or rotary encoders if you have enough cameras. In the self-driving car industry, Tesla famously holds that lidars (which rely on infrared) are unnecessary. If biological organisms can self-drive using vision alone, so can cars.

Of course, you cannot entirely eliminate all other sensors. Sound is clearly almost as important as vision even for us highly visual primates. And all mammals have the equivalent of accelerometers in their ears. So it is no surprise that phones and robots typically feature microphones and accelerometers as well. But, to a first approximation, cameras are eating the world’s sensors, and will demand a return to a more colorful world.

Cameras are surprisingly powerful sensors because almost everything that might matter for feedback in a controlled environment either has a natural visual component, or can be rendered visible at low cost (for example, introducing a colorant in invisible dangerous gases or clear liquids). Just as the smartphone has replaced many personal devices, the camera can replace dozens of different sensors.

The cost though, is the need for intelligence. You need an AI, or ideally, a robot to interpret and react to color (and visual fields in general).

And when color communicates safety, and is linked to behaviors that must unfold extremely rapidly, you have a major challenge for future designers and architects.

What should the color of safety be in the age of AI? How will be paint industrial interiors when the primary inhabitants are camera-driven intelligent machines, ranging from robots to appliance-like machines, to background smart-environment features?

The designer – or AI – who figures out the answers will enter the history books alongside Birren and his late-industrial successors.

Coda: The Color of Safety On Screen

Since the color of safety is rarely fully visible unless there is an active emergency unfolding, and because industrial interiors are not particularly accessible anyway, it is not an easy observation target for protocol watching1 enthusiasts.

Fortunately, it also happens to be a favored topic in screen media. It is fascinating to pay attention to the color of safety in movies and television, especially over time (obviously, this is only possible past the black and white era).

The wave of disaster movies from the early 1970s, including such cult classics as The Towering Inferno and The Poseidon Adventure, showcase Birren-like safety environments at their peak. The build-up of tension has an uneasy background-dominant vibe of terror to it. The characters seem to inhabit their environments much more mindfully. Plots seem much more atmospheric and suspenseful.

By contrast, modern disaster thrillers are much more likely to feature bright flashing lights and loud alarms – and characters who don’t really attend to their environments until things start going wrong. Thrilling foreground excitement tends to replace ominous background build-up of terror. Plots are action-packed rather than suspenseful.

As storytellers adapt to the age of AI, we can expect to see the color of safety reshape human narratives yet again. And us protocol-watchers will have a whole new crop of movies and TV shows to watch, and a new color-of-safety language to decipher.

Rafa Fernández is the host of the Protocols for Business SIG, which meets every two weeks on Discord to discuss protocols in business settings, and you’re welcome to join the next session. Link to join at the end of the article.

Seeing Business Protocols

If you ask a typical manager how their company works, they will usually begin with familiar answers. They describe the business model, sketch an organizational chart, or point to operational policies. Often they use these to explain how decisions are approved, how risks are reviewed, and how work is supposed to move from idea to execution. These descriptions are usually sufficient – until something goes repeatedly wrong or keeps going stubbornly right.

When coordination continues to break down or when performance remains unexpectedly stable despite adverse conditions, those accounts begin to feel lacking. Deadlines slip week after week. Review cycles stretch without resolution. Or a team continues to hit targets year after year in an unfavorable market. Managers reach for partial explanations: tooling issues, unusually strong execution, “good ops.”

Individually, these accounts are accurate. But they rarely explain why the same deviations – positive or negative – recur in the same places.

Recurrence is the clue to protocols, which provide a more precise explanation. Protocols do not solve problems or guarantee success. They lock in commitments that stabilize how variation is absorbed over time. In doing so, protocols keep friction within tolerable bounds, allowing organizations to continue operating – and in some cases, to build momentum – despite persistent strain. This is their quiet achievement: protocols hold recurrent tensions within ranges the organization can live with.

For this reason, protocols rarely appear as foregrounded processes. Once established, they recede into habit, embedding themselves in infrastructure, workflows, and expectations. They become “how things are done,” rather than objects of reflection or redesign. As Venkatesh Rao, Program Director of Summer of Protocols, observes:

Protocols do not announce themselves. They constrain motion quietly, with little conscious effort, shaping what can happen without drawing attention to themselves. This creates a methodological problem. If protocols are most effective when they are least visible, how should they be studied?

Our group, the Protocols for Business Special Interest Group (SIGP4B), has taken an oblique approach. Rather than attempting to inspect protocols through documentation or formal analysis, we have found success by studying moments when habituation fails in business contexts. Operational reinvention and technological shifts interrupt routine behavior and make protocols visible.

These moments of disruption show up as unresolved approvals, delivery delays, or persistent strategic confusion. In parallel, they appear as periods of sustained advantage: new categories forming, firms holding position through volatility, or business models that continue to function when others falter. Without attention to the underlying pattern, these episodes are often treated as isolated failures or market-cycle anomalies. When viewed through the lens of protocol studies, recurring business challenges in moments of volatility temporarily illuminate how an organization is actually structured.

In geology, fault lines are not identified by close surface inspection. They are discovered when accumulated stress forces the underlying structure to express itself. Our discussions highlighted how protocols behave similarly. Persistent problems and persistent point less to local error or exceptional talent than to the protocols through which pressure concentrates and trade-offs are stabilized. The artifacts left behind by this work – extra reviews, HR policies, new roles, software controls, even new business models – offer a useful vantage point. Still, it is sustained strain and disruption that make the boundaries of protocol structures unmistakable.

Over the past year, our group has been watching one such disruption unfold across many organizations at once. AI software solutions, such as large language models and autonomous agents, are diffusing, often in the shadows, into everyday business operations. How work is generated, evaluated, and sequenced in time is changing rapidly. As a result, protocols that once operated in the background have become visible, creating an opportunity to examine what business protocols are actually doing when placed under pressure.

Join Protocols for Business

Protocols under Pressure

The dominant technologies of the past two decades – networked software, mobile devices, cloud infrastructure, remote collaboration – are entering a period of relative maturity. Their organizational effects are well understood, if not fully resolved, especially in today’s top enterprises. AI software, by contrast, has just arrived in the past few years. LLMs, copilots, and early autonomous agents are being adopted unevenly across functions, spawning new markets, but often failing at enterprise-level deployment.

AI is usually discussed in terms of automation or productivity. Those framings are not wrong, but they miss what makes AI adoption particularly revealing from a protocol perspective. While much of the public discussion frames AI in terms of cost-savings or new markets, our SIG has been focusing on the pressure it places on current coordination systems by changing the speed and scale at which work is produced: AI lowers the marginal cost of many forms of knowledge work while dramatically increasing total output. Drafts, analyses, summaries, and proposals can be generated continuously. AI’s production acceleration shows up everywhere at once; tools are being adopted simultaneously in product, marketing, legal, operations, and management functions. Work that once arrived in bounded increments now arrives as a flood, often carrying significant amounts of low-signal algorithmic output (“slop”) mixed in with it.

Many organizational protocols evolved under production paradigms which assumed a specific level of scale. Documentation, like other forms of knowledge work, took time. Decisions moved through sequential approval gates. These constraints shaped how firms organized authority, pacing, and oversight.

Our group repeatedly returned to version control as a protocol case study. Today’s default workflow assumes that meaningful code change is mediated through human reviewers, which stabilizes quality and accountability by gating “merges.” This implicitly capped how much change a codebase can absorb over time, limiting it to the skill and capacity of reviewers. Large refactors – such as migrating a core codebase to a new language – have historically been slow because review, coordination, and rollback had to proceed at human pace.

Now, LLMs challenge these assumptions and workflows. The resulting strain mirrors an earlier protocol transition from centralized to distributed version control, which supported continuous deployment at scale. New protocol paradigms are necessary to address AI-enabled speed and scale.

Production constraints become easier to understand when protocols are treated as technologies in their own right. In “Constructing the Evil Twin of AI,” Venkatesh Rao describes protocols as deliberately anti-intelligent systems:

Protocols are designed to limit, to configure trade-offs, and to make actions repeatable under uncertainty. They create leverage through hardness, yet often can be brittle under certain strains.

It’s worthwhile to note that AI compresses some constraints while bypassing others. Output multiplies while, at least for now, reviews and escalations are sometimes sidestepped. Production gains from AI adoption do not lead to smooth acceleration, but uneven load and jerky momentum.

Seen this way, the friction and failures accompanying AI adoption are not primarily symptoms of resistance to change or poor product quality. They are signs of protocol integrity under load. AI introduces fluidity where rigidity previously performed a stabilizing function. Existing coordination systems are being asked to absorb more output, faster, without having been designed for that operating range.

Similar to fault lines, coordination pressure accumulates along established paths. What has been routine begins to feel strained. What has been tolerable begins to feel tight. To understand the consequences of that pressure, it helps to look at where it concentrates first.

Tensions in Business Time

Across the SIG’s discussions, interviews, and readings, a consistent pattern has emerged. Under AI adoption, the first thing that stops working smoothly seemed unintuitive: time.

This became clear when our group reviewed Blake Scholl’s writing on Boom Supersonic. Here, Scholl distinguishes between at least two clocks operating inside the same organization. The first is the calendar: project timelines, milestones, and delivery dates. The second is what he calls the Slacker Index: the amount of time engineers spend waiting – on inputs, approvals, dependencies, or external constraints – rather than building. Even in well-run, safety-critical organizations, these clocks coexist.

Under stable conditions and in mature industries, this alignment is usually implicit. Engineering velocity, supplier lead times, regulatory review cycles, and internal decision-making rhythms evolve together. At Boom, hardware design, simulation, testing, and supplier manufacturing are paced to one another. Slower clocks constrain faster ones in predictable ways. Waiting is visible, expected, and priced into the system.

As Scholl points out, AI-enabled production changes the speed and scale of production. Certain forms of work – design iteration, analysis, documentation, internal review – can suddenly accelerate by orders of magnitude. From the perspective of the Slacker Index, local waiting collapses. Yet the calendar will not automatically follow. Supplier lead times remain fixed. Certification processes still unfold at human and institutional speeds. External partners continue to operate on contractual and regulatory time.

The consequence of AI-enabled opportunity is temporal divergence (a topic explored in depth by SIG member Sachin). Some clocks speed up sharply while others remain unchanged. At Boom, this would mean design teams outrunning suppliers, simulations outrunning manufacturing feedback, or internal decision cycles outrunning the capacity of external partners to respond. The Slacker Index may improve locally – less waiting to produce – but worsen systemically as downstream dependencies fall behind.

AI systems further amplify this effect in two ways. One, because they generate outputs without passing through the durations that normally situate work, creating a dizzying orientation. Large language models produce analysis and proposals instantly. Work arrives early in excess of the organization’s ability to absorb it. Knowledge accumulates faster than it can be evaluated, integrated, or acted upon.

Second, AI software using LLMs can be contextually misaligned. They draw on data that’s often years apart (a model trained up to 2024, used in 2026) and produced outside the local business context. From this lens, the recent focus on improving AI product memory seems intuitive. Efforts such as RAG, MCP, skills, and even “undo” prompt features become attempts to realign probabilistic software into business context, tempo, and authority.

Safety-critical organizations like Boom make these dynamics visible precisely because they cannot simply collapse time. Hardware, suppliers, and regulators enforce non-negotiable rhythms. When AI accelerates internal work without moving those external clocks, coordination strain surfaces quickly. Slack accumulates in unfamiliar places, with no protocols available to redistribute it.

When time regimes fall out of alignment, coordination problems and opportunities change form. Delays no longer appear as isolated errors that can be corrected locally. Instead, organizations experience escalating tensions: pressure to act without corresponding capacity to review, decide, or remember.

As with other systems under strain, previously hidden structures become easier to observe once alignment fails. AI adoption exposes how dependent coordination was on the quiet temporal alignment – rhythm – of existing protocols.

Yet work does not stop. Systems do not fail outright. Instead, protocols drift out of phase. Understanding how firms respond requires looking at how they attempt to restore coherence – often without redesigning the underlying structures that produced it.

Enterprise Management, Built to Last

When shared assumptions about time lose coherence, organizations first adapt within current structures. Work continues by absorbing friction rather than resolving its source.

One visible form of this absorption is Boom’s solution: integrate vertically. The critical move was purchasing their own large-scale manufacturing equipment rather than continuing to rely on external suppliers whose lead times dominated the schedule. Supplier queues and fabrication delays had become the governing clock for the entire program, producing a high Slacker Index: engineers were ready to iterate, but progress stalled while waiting on parts. By acquiring the machine, Boom internalized that bottleneck and converted supplier wait time into an internal, controllable process. This collapsed a multi-month external dependency into a shorter, iterable internal cycle, allowing design, testing, and manufacturing to co-evolve rather than queue sequentially.

Another response was novel translation work. The SIG discussed the fast growing Forward Deployed Engineer role, emerging to help mediate between fast-moving demands and slower-moving infrastructure. Their task is not to eliminate mismatch, but to work across it and leverage it – adjusting scope, translating intent, and negotiating constraints as they appear. This work allows organizations to keep operating even as tempos diverge, and gain a competitive advantage in the process. At its best, the work defines the operating model. This is the case for Palantir and large AI labs like OpenAI and Anthropic.

Other adaptations the SIG encountered took the form of operational formalization: AI usage guidelines, governance documents, digitized ontologies. These measures make previously tacit constraints visible without altering the structures that produced the misalignment. They stabilize behavior at the margin while leaving underlying coordination regimes intact.

As adaptive load to the new pressures increases, authority structures reassert themselves. Approval gates and prohibitions harden. Data confidentiality clauses are expanded. Hierarchies become more visible. These protocols surface because they are firm fault lines that stabilize liability, escalation, and accountability when temporal coherence weakens.

Some of these protocols absorb the load effectively. But others begin to strain. In geological systems, stress redistributed after an initial shift often produces secondary movements elsewhere. Organizational responses to temporal misalignment follow a similar pattern. Processes adapt where they can, while pressure seeks relief where they can’t.

This adaptive pattern becomes more consequential as management itself is increasingly parameterized within software, paving a path to programmatic enforcement.

Warning: Brittle Software

Before returning to our initial question on business protocols, it is worth taking a brief detour. Alongside the recent diffusion of AI, the SIG has been considering a quieter shift: management controls themselves have been moving into software.

Approval flows, access rights, version control rules, and incident response procedures are increasingly encoded directly into business software systems. Authority becomes a software admin configuration setting. Expenses are automatically rejected when outside policy. These changes predate AI, but they shape how AI is experienced inside organizations.

This matters because management subsumed in software behaves differently from practiced management. Protocols, when encoded in software, limit discretion in what is authorized and auditable. In contrast, management practice as a whole consists of what is necessary and effective in context. Under stable conditions, the gap between the two – protocol and practice – is manageable. A manager can override an expense request, or call a system administrator to update the software configuration.

AI changes the scale and speed at which this gap becomes visible. On one hand, employees and customers bypass formal systems using personal tools, custom scripts, or unregistered agents operating through employee accounts. On the other, software-encoded controls enforce protocol broadly and uniformly, at relatively low cost. A single configuration change can propagate across an organization instantly.

This shift produces a recognizable organizational pattern. Practices that once relied on informal judgment or situational flexibility are increasingly forced into deterministic software-defined pathways. Work that used to be resolved through conversation – such as asking legal for a quick review or exercising discretion under time pressure – is now mediated through systems that require explicit inputs, permissions, and logs. These systems work especially well in mature and slow-changing businesses, and can be brittle under volatile situations like a new technological disruption such as AI software.

A recent example is from an airline: A customer relied on an airline’s AI chatbot for information about bereavement fare rules. The chatbot incorrectly stated that refunds could be claimed retroactively, advice that contradicted the airline’s formal policy. When the customer followed the guidance and was denied the refund, the airline argued in court that the chatbot was a separate tool and that customers were responsible for verifying information elsewhere on the site.

The court rejected this argument. It ruled that the chatbot was part of the airline’s customer-facing system and that the company was responsible for the commitments it made, regardless of whether those commitments were generated by an AI system. It is tempting to shape the story to focus on LLM shortcomings: “This type of mistake, in which generative AI tools present inaccurate or nonsensical information, is known as AI hallucination.”

What’s more important is that the failure occurred because a time-tested operations efficiency strategy (delegating frontline explanation to an automated system) collided with protocolized legitimacy (formal fare rules, auditability, liability). The business made an assumption that explanatory labor could be delegated to probabilistic software, decoupling it from the protocols that confer legitimacy and liability. The chatbot accelerated response and reduced staffing load, but it was not strictly aligned to the same review, approval, and accountability protocols that governed pricing policy. At sufficient scale, that gap became visible.

Understanding this trend increases the diagnostic value of the current moment. AI adoption, like many previous “digital transformations” since the advent of the internet and before, highlights where management has already been formalized, where practice has been carrying hidden load, and where current protocols are breaking.

With that context in view, it becomes easier to return to the original methodological question: how AI, not framed as a solution but as a protocol disturbance, makes the underlying structure of coordination inside firms visible.

Strange New Business

Since the summer, our group has treated AI adoption and similar business disruptions as an observational lens rather than an optimization target. This lens supports our SIG’s goal of researching, evaluating, and finding opportunities to improve business protocols under pressure. AI is useful here because it is widespread, fast-moving, and poorly aligned with existing organizational rhythms.

This approach reflects a broader pattern in protocol studies. As Venkatesh Rao notes in Strange New Rules, new coordination regimes often feel unfamiliar at first, and then they are banal. Over time, they stabilize. Eventually, they disappear into habit. What becomes difficult is not living with protocols, but noticing and refactoring them.

AI has interrupted that disappearance. It has surfaced assumptions about time, authority, and review that were previously implicit. In short, it has made visible where management is already encoded, where practice has been carrying unacknowledged load, and where coordination depends on alignments that no longer hold reliably.

Although our recent work has focused on business management protocols – multisigs, version control, incident response – similar dynamics appear in transaction and production protocols as well, including market auctions and inventory management. Across cases, recurrent confusion, friction, and governance anxiety function serve as signals of protocol stress; they point to coordination systems doing more work than they were designed to absorb.

The conditions described here are not temporary. As AI software continues to diffuse unevenly into business operations, the protocols that manage coordination trade-offs will remain under strain – and increasingly visible. Our group treats this moment as an opportunity to observe, compare, and refine how business protocols actually function under load. If you are interested in these topics, or are encountering protocol changes yourself, share your email here and join our next meeting!

Join Protocols for Business

The goal of this two-part essay is to introduce the work of the Special Interest Group in Formal Protocol Theory (SIGFPT) over the last six months, and invite you to join us as we begin our second year of explorations. In this first part, we will establish some broad qualitative background and overarching frames and concepts. In Part II, to be published in a few weeks, we will introduce you to the specific theoretical apparatuses and technical attacks we are exploring, for developing formal protocol theory.

Join us on our Discord for the 2026 kickoff meeting of SIGFPT on Friday, January 9, 10am Pacific. Calendar invite for series.

So what is protocolization? In what follows, we will explore this question, with a view to setting up a canvas for formal and mathematical attempts to theorize it in general ways.

Protocolization is an unfolding long-term planetary transformation that is at least as old as modernity itself. That is to say, it is at least several centuries old, and has unfolded through multiple distinct chapters, in loose synchronization with major technological revolutions. Protocolization is the overarching story of how humanity has cashed out and orchestrated the capabilities of every wave of new technologies to make our world ever more comfortable and hospitable for humans.

One side of protocolization is arguably an ongoing terraforming. The first terraforming in fact; that of Terra itself, preceding any science-fictional terraformings of other planets.

The other side of protocolization is changes in human behavior in response to this terraforming. Relative to our ancestors, we have become more predictable and orderly on some fronts, and more unpredictable and generative on others. Protocolization is simultaneously a civilizing process, and a rewilding process.

Protocolization is reducible neither to its technological drivers, nor to the changes in human nature triggered by them. Protocolization is the co-evolution of both into what we might call a new nature. A planetary condition powerfully determined by the laws of the artificial, which can increasingly be engineered to be nearly as immutable and indefinitely persistent as those of nature itself.

Subscribe now

New Nature

Protocolization proper has received almost no attention, either popular or scholarly. It has merely been partially glimpsed in peripheral vision in attempts to understand more visible and dramatic planetary processes (such as modernization, industrialization, globalization, urbanization, and digitalization), which offer readier, more legible approaches to analysis and synthesis.

These more dramatic processes typically also feature charismatic technological megafauna at their cores – printing presses, steam engines, automobiles, ocean liners, skyscrapers, bridges, superhighways, rockets, airplanes, nuclear reactors, supercomputers, AI systems, miracle medicines, robots. These “hero” technologies lend themselves to anthropomorphic projection, and encourage a view of human-technological co-evolution as a kind of augmented hero’s journey. Our default notions of “progress” are of this heroic variety.

By contrast, the technological elements of protocols – interoperability standards, kits, standardized fasteners, electrical connectors, plumbing regulations, safety codes, sewage pipes, modularity grammars – typically gently diffuse and deflect anthropomorphic impulses. Rather than serving our heroic individualist impulses, they quietly orchestrate and shape our mutualist and cooperative energies. Protocolization looks like ecological emergence rather than a technological hero’s journey. It encourages a very different narrative of progress.

We hope to convince you that “protocolization” elegantly subsumes some of the most salient features of the more dramatic and “heroic technology” process frames you might be used to, while also apprehending (if not yet quite comprehending) a great deal that those frames remain systematically blind to.

It may help to keep a pair of pictures in mind. One is the familiar image of the African savannah, with its tableau of charismatic megafauna. The other is the near-pitch darkness of the mesopelagic zone of the oceans, beyond the reach and attention of both fishing fleets and the cameras of shark documentaries. This zone is densely populated by a family of bioluminescent fish that you may have never heard of or seen pictures of – Bristlemouths (Gonostomatidae) – which number in the trillions to quadrillions, constituting nearly 90% of fish biomass on the planet by some estimates. These fish anchor not just oceanic ecologies, but the entire planetary ecology.

Dramatic planetary processes are stories about the first kind of image. Protocolization on the other hand, is the story about the second kind of image.

The first kind of image requires no particular skills to see and appreciate, and anchors most visions of technological progress and change. The second kind of image is much harder to see, and takes cultivated protocol-watching skills, but is also much more useful for understanding how ecologies actually work; how new nature works.

Since our goal is to build powerful and useful theories of new nature, we navigate by the second kind of vision of technology; we try to search for formal theories and tools that can help us see, model, predict, and shape the “mesopelagic zone” aspect of new nature, rather than the “African savannah” aspect.

The most recent chapter of protocolization, triggered a half-century ago by the affordances of early computer networks, has been transforming the world in profound and accelerating ways. But arguably, it has barely been theorized at all, largely because when you try to look closely, it looks like the mesopelagic ocean depths rather than the African savannah.

Protocolization Defined

For the purposes of this essay, we will loosely define protocolization as follows:

Protocolization is the progressive metabolization of reliably repeatable technologically mediated human behaviors at all scales into reliable planetary infrastructures for coordination.

Here, we mean technological mediation in a broad sense, encompassing conceptual and cultural technologies such as the decimal system (instrumental in the development of modern book-keeping) or hygiene or nutritional practices that first took shape as religious observances.

It is worth noting that though we only mention human behavior in the definition, these behaviors shape the fates of all other life on the planet as well. Not just the major domesticated species of plants and animals, but through the shaping of “wilderness” regimes of the planet, all other life as well.

Protocolization is driven by some mix of knowledge discovery and diffusion, technology enablement, and human behavior change, resulting in a slow accumulation of new nature ecologies across the planet.

An example we use frequently in SIGFPT meetings is hand-washing.

The global ubiquity of hand-washing protocols was driven by the discovery and diffusion of knowledge about infectious diseases, coupled with ubiquitous technological infrastructure that has made clean water and soap available everywhere in modern built environments. The protocolization of hand-washing behaviors transformed a planetary landscape of previously uncoordinated and high-variance local cultural practices into an infrastructural behavior with much lower variance. As a direct result, public health improved everywhere. The planet became more comfortable and hospitable for humanity.

Thousands of similar examples can be listed across every area of human behavior, and in every corner of the built environment, at all scales from individual to international, and from trash cans to continental supergrids.

An entertaining cross-section view of protocolization can be found simply by listening for the word protocol in movies and television. Such usage has exploded in the last half-century. One of our favorite specific illustrations comes from long-running television franchises. In Star Trek for example, the word protocol was not used at all in the original series, but is used hundreds of times in more recent shows within the franchise.

This already extant cultural footprint, by the way, is why we expect you to find the term protocolization intuitive, and experience an immediate sense of recognition of what we’re pointing to with the term, even if you’ve never stopped to think about it before.

Domains like climate, traffic, networked computing, and healthcare make particular (and domain-specific) use of protocol-based understandings of the human and machine behaviors they design and translate into infrastructures. In other domains, protocols might not be the direct result of conscious design, but manifest as unexpected systematicity, orderliness, and serendipity in how things happen, which goes unnoticed and un-theorized.

But we do sometimes notice – when we travel to a city or country with more reliable subway services than our own, for example, we are struck by a sense of serendipity suffusing our transit experiences; a sense that the world is unusually hospitable in ways precisely attuned to our needs.

That is protocolization; steadily accumulating ecological transformations of the planet, each of which makes us feel surprisingly, magically lucky when we first encounter them, but which rapidly turn into tuned-out backgrounds and unconscious expectations of artificial lawfulness in our environmental contexts.

Progress through Invisibility

Even though protocolization has been as momentous as more dramatic planetary transformations like computerization and globalization, as well as earlier ones such as urbanization, consumerization, and industrialization, it remains hidden.

This is because, by their very nature, protocols in their evolved forms tend towards invisibility, and protocolization as the evolutionary process that creates them seeks out invisibility. This feature, arguably, is central to how and why protocolization constitutes progress. Protocolization is progress through invisibility.

Here, we mean invisibility in a broad, multi-dimensional and textured sense, beyond merely literal visual invisibility. The invisibility of protocols has multiple aspects:

1. Operational invisibility – We don’t think about the protocol while using it, acting through muscle memory.

2. Infrastructural invisibility – The machinery is literally hidden/backgrounded, like the ecologically critical mesopelagic zones we alluded to earlier. For example, cables and sewage lines are buried, and are generally as ubiquitous and invisible as bristlemouth fish.

3. Social/cognitive invisibility – People don’t recognize protocolization is happening because by definition, behavioral habituation is not only part of the essence of the process, it is functionally load-bearing. Traffic protocols would be much less effective if we were constantly aware of them. We “stay in our lanes,” without thinking about it, and reflexively signal turns and stop at stop signs.

4. Anti-memetic qualities – Protocols rarely feature in headlines, titles, or public discourse, because the unremarkable “boring” quality, with one experience being largely indistinguishable from another, is essential to their functioning.

5. Marginality – Protocols often organize and scaffold the technological space around other, more charismatic technologies that are intrinsically more visible, and more directly support foreground goals. They perform prophylactic functions, enforce constraints, and structure core behaviors for coordination and synchronization. For instance, the safety protocols of airline travel – airport security checks, seatbelts – are on the margins of the foreground retail experience of air travel. We do not think much about them when making travel plans.

6. Default quietness – Protocols are quiet and unobtrusive by default, and make themselves visible primarily when they fail. An ordinary hand-washing episode is not memorable among the tens of thousands that happen over an individual lifetime. An episode involving an empty soap-dispenser or malfunctioning faucet is memorable.

7. Obliquity of action – Protocols rarely directly act on consciously controlled human behaviors in pursuit of explicitly held intentions or goals. Rather, they act through secondary pathways of supporting causation, contingency triggers, and automated follow-through. For every conscious action in a technologically advanced environment, dozens of supporting actions may unfold unnoticed, rendering the conscious behavior preternaturally frictionless and seemingly “free.” The scope of protocolization is often everything else that must happen in order for actions you consciously and freely intend and do to succeed.

Aspects of protocols are of course highly visible in specific contexts. For example, everyone knows what USB is at a lay level, and what is meant by “containerization.” Standards bodies debate specs publicly (though it is worth noting that such debates are rarely newsworthy unless they fail in dramatic ways). But we hardly ever think about the USB protocol, even as we routinely plug and unplug connectors several times everyday. Or about the mechanics of containerization driving the endless torrent of package deliveries at our doorsteps.

So to a first approximation, and across multiple domains, protocols are invisible technological phenomenology. They could be considered technological dark matter; evading study by retreating to the margins of our attention, ceding the spotlight to the “main characters” of the technological environment – striking buildings, beautiful bridges, impressive airplanes and rockets, awe-inspiring AI systems.

Protocol designers in nearly every field rightly view such invisibility as a feature rather than a bug. Protocol designers intuitively tend to design for invisibility. Invisibility of some sort – from literal visual invisibility to the reflexive automaticity in our own responses to subconsciously registered cues – is either a strong functional requirement, or at least highly desirable. Often, the very reliability of a protocol rests on its invisibility, and visibility can lead to fragility and unreliability.

Keep a protocol in peripheral vision, and it works. Look too directly at it, or attend to your behaviors within it too deliberately, and it starts to falter.

In our emerging field of Protocol Studies, this widespread design intuition has been elevated to the level of a philosophical principle, captured by Whitehead’s famous observation that “civilization advances by the number of operations we can perform without thinking about them.”

Within our young field, it is now a convention to refer to a valuable advance in protocolization, marked by a transition of an important behavior to invisibility and collective unconsciousness, as a “Whitehead advance.”

The Price of Progress

Protocolization can be understood as a centuries-old macro-trend of progress through invisibility, but this progress does not come for free. There is a clear cost associated with it, in the form of a new class of problems induced by invisibility.

This is most apparent in governance at larger social scales. In the most extraordinary cases, such as the effect of vaccines (inarguably one of the greatest Whitehead advances), protocolization becomes invisible to the point of even erasing its own historical raison d’être, and undermining the broader literacies required to sustain itself. We might perhaps call this pathology overprotocolization.

Modern vaccine denialism is possible precisely because vaccination protocols have made their own logic largely invisible through their extraordinary success. Something similar can be said of nuclear non-proliferation – the perception of risks of nuclear-armed conflict have diminished in proportion to the success of nuclear weapons protocols, and the attendant retreat of nuclear weapons into background invisibility.

Invisibility also creates peculiar difficulties in the study of protocols, since we are forced to talk of them through the fragmentary and mutually inconsistent languages of other fields, making the search for universal principles and general theories particularly difficult.

This, however, is a problem that we can begin to address in many ways. And one of the best ways to make protocols visible in positive ways, at least to scholarly and professional scrutiny, is to develop expressive formalisms and mathematical models to undergird a language for talking about them in general terms.

The work of SIGFPT encompasses both formalization as such (the development of ontologies and precise terminology that allow us to talk consistently about phenomenology across many domains) and mathematization (the development of suitable technical methods to investigate the phenomenology through the formalism).

Our explorations and activities should interest researchers and practitioners across a wide variety of existing fields – computer science, control theory, operations research, economics, sociology, anthropology, and political science among them.

Regardless of what background you are coming from, the value of FPT for you is mitigation of the theoretical fragmentation that you likely encounter whenever your interests lead you to protocols and protocolization.

Tackling Theoretical Fragmentation

Protocolization is a multi-scale planetary phenomenon that couples the local to the global, and entangles micro-level dynamics and macro-level dynamics, via technological media. This broad footprint across spatiotemporal scales and regimes of dynamics, exacerbated by general invisibility, tends to create fragmented theorization of hugely important phenomena.

An example will illustrate what we mean here, and motivate the approaches we are developing to mitigate theoretical fragmentation and construct more integrated understandings.

Containerization, the protocolization of global commerce, is generally analyzed through the lens of globalization, understood as a primarily political and economic phenomenon, even though it obviously involves a great deal more. At the micro-scale it is simply a mature standards process that stewards definitions of a set of relatively low-tech box form-factors. At the macro-scale it is an emergent intermodal network-of-networks for materials transport. At the local level, it is a tangle of logistics problems, such as regional staging, port operations, last-mile operations, border transit operations, and security procedures. At the global level, it is a climate-like hyperobject comprising a constantly shifting set of flows and stocks, and featuring weather-like phenomena.

To enable theoretical study of containerization, a sufficiently expressive formalism (or more likely, a harmonized assemblage of such formalisms) for protocols must be able to articulate concepts and propositions ranging from the “box” level to “planetary climate” level. It must comprehend dynamics ranging from an earthquake affecting a particular port, to a slow accumulation of empty containers on one continent due to unbalanced trade (the backhaul problem).

And this is merely the simplest sort of protocolization, where we can name and point to a relatively coherent single “protocol,” with relatively intelligible structural and behavioral phenomenology to roughly isolate, model, and study.

Such intelligibility, however, is an illusion. Any fragile, non-fragmentary picture we might build of a phenomenon like “containerization” shatters the moment we situate it in the real world, and bring it into contact with other protocols of comparable richness that occupy some of the same spaces it does.

Protocol Tangles

Even within the seemingly isolatable domain of containerization, formal protocol theory faces tough tradeoffs between generality and specificity, resolution and scope, and so forth.

But the need for formal protocol theory becomes even clearer when we attempt to investigate multiple protocol infrastructures at once, motivated by provocative similarities, seductive resonances, and most importantly, consequential real-world convergences (such as many physical networks sharing right-of-way corridors).

When many protocols co-evolve in convergent ways and get materially entangled, constituting a single protocolization process spanning multiple partially visible and operationally coupled infrastructures, they tend to become especially invisible. This is because a view corresponding to one component protocol and associated set of interests misses all the others.

Since the early days of protocol studies, we have thought of this as the “three blind men and an elephant” problem in studying protocols, especially large and complex ones.

Such circumstances present as especially acute fragmentation of attempts to study them theoretically. We refer to such circumstances as protocol tangles. Here are some examples:

• Factory automation emerged at the intersection of computerization within factories, and globalization between (globally distributed) factories. It is a tangle that cannot be adequately studied by any of the obviously relevant disciplines, such as industrial engineering, operations research, computer science, economics, or trade policy.

• Global public health regimes, a response to planetary-scale contagion phenomena, are tangles comprising protocols at the intersection of municipal public health infrastructures, global pharmaceutical research and disease surveillance protocols, and international relations.

• The presently unraveling Rules-Based Order of international relations is a tangle of planetary scale coordination problems factored into impoverished sets of nation-centric problems. Even though massive international cooperation undertakings (such as the UN Food Program, the UNHCR, and the WWF), operate as exceptionally invisible protocol infrastructures, they are understood and governed through highly impoverished and fragmented theoretical frames.

The worst protocol tangles are so invisible, they even lack names with which to point to them. Attempts to think systematically about them run aground immediately at the level of language.

For example, we lack even a name for the particular protocol tangle comprising containerization infrastructure, the internet infrastructure used to govern it, and the multinational corporations that manage the entanglement. We are reduced to thinking in terms of sui generis historical episodes, such as the “2017 Maersk ransomware attack,” rather than powerful theoretical frames attached to particular units of analysis, and usefully general operating principles.

In this instance, despite the similarity of the two networks (both are packet-switched networks of sorts), it is awkward to talk about the converged tangle in the language of either containerization or the internet. For example, a key difference between the core protocols of the two networks is that internet packets can be “dropped” and re-transmitted but lost containers cannot trivially be “re-shipped.” Another is that the container network has a “backhaul” problem, but the internet does not.

How does one theorize the implications of these differences to shed light on a shared phenomenon like “buffer bloat”? Could we talk formally of reversibly vs. irreversibly lossy protocols? How does one use such theorizing to deal with problems where the two networks converge or collide in the real world, such as in the case of the Maersk ransomware attack? Could we invent a converged “internet of containers” that is robust to such attacks?

So while both might be considered part of a broader class of packet-switched networks, it is not immediately obvious how to construct a formal theory encompassing both that is not uselessly over-general (like the famed “spherical cow” problem), and useful in engineering application.

As a result, discourse on the beguiling similarities between these two protocol infrastructures tend to remain in the realm of poetry and metaphor. Consequential encounters between the two, like the Maersk attack, never gravitate out of history books and into engineering or management textbooks.

Where the impoverishment caused by fragmented and incompatible frames becomes too severe, as in the case of climate action, planetary AI, and blockchains, coordination fails almost entirely, and the failure of sufficiently ambitious and powerful theorization is most acutely felt.

So as protocols become ever more complex, ever more mutually entangled, and non-reducible to non-protocol infrastructures or technological islands, they become increasingly difficult to deal with and govern.

And perhaps no protocol tangle is more devilishly difficult to comprehend and investigate than the planet-scale one caused in the last half-century by computation.

The Planetary Computational Tangle

In developing formalisms and mathematical approaches, it is worth paying special attention to a particularly consequential tangle of protocols, involving the varied modern interacting infrastructures that rest on computing capabilities. This tangle includes, but is not limited to: the existing internet, AI, blockchains, AR/VR, robotics, the “internet of things,” and planetary sensing/monitoring infrastructures, at all scales from microscopic to orbital.

This is the civilizational-level boss of protocol tangles, the planetary computational tangle, or PCT.

As the Maersk ransomware attack episode showed, it is already clear that the default mental model of the PCT as some sort of “extended internet” has been severely strained, to the point of collapse. While pleasing new formulations like “planetary intelligence” help us at least point to the PCT, they typically pre-commit to an overly legible and panoramic theorization heavily shaped by a unitary perspective, aesthetic leanings, and focal interests shaped by their most charismatic elements. They miss the essential messy plurality of the PCT, and the inextricability of human experiences and agencies within it. They serve our appreciative needs, but not the instrumental needs of practically useful theorization.

To develop powerful mental models of the PCT, especially given that thinking of the internet itself as merely a “network of computers” has already proved to be highly limiting, we need a radically different approach to theorizing. One that privileges a worm’s-eye view of the tangled messiness over pleasing panoramic views and grand narratives.

For formal protocol theory, the PCT might be viewed as the “one ring” modeling and governance challenge.

To the extent we are able to build theories and practical techniques for grappling with it, we will be able to make sense of all dimensions of protocolization, and deal more elegantly with more bite-sized protocol problems that can be meaningfully isolated and tackled. It is already clear that existing ontologies will not do when it comes to “carving the reality of the PCT at the joints,” so to speak, to isolate problems worth tackling. New ontologies are required.

To the extent we fail, our increasingly technologically terraformed planet will become invisible not just to our senses, but also to our minds.

The irruption of AI into the world, currently the most dramatic and attention-attracting world process, is worth some additional thought. Currently, the role of AI in the PCT is being understood and constructed through the lenses of corporate product development and nation-state-based “sovereign AI” frames. This approach is inherited from previous experiences like dealing with fissile materials or climate change. These frames miss much of what is important, interesting, valuable, and risky about AI, especially when dealing with its rising “agentic” usage at planetary-infrastructure scales. Such usage is unlikely to remain limited to the traditional boundaries and capabilities of products, corporations, or nations, and human actors who identify with them.

Similarly, the role of blockchain-based infrastructures in the PCT is being reduced to rigid and relatively unimaginative nation-based identity and monetary infrastructures. This again misses much of what is important, interesting, valuable, and risky about them. The flexible programmability of infrastructures like the Ethereum “world computer,” with its capacity for expressing and articulating novel and fluid regimes of law and governance, and disciplining the myriad wildernesses of the PCT with striations of orderly hardness, remains largely unexplored. Again in part due to the limiting theoretical frames through which it is viewed.

Our mental models of the PCT are perhaps most severely compromised when it comes to the raw materiality of it.

There are already over 17 billion connected Internet of Things (IoT) devices in the world. We are currently on the cusp of a robotics boom. Planet-scale sensor-networks and environmental monitoring systems are beginning to take shape all around us. Warfare has already been irreversibly transformed by autonomous vehicles on air, sea, and land, and civilian life is poised to follow.

The picture is similar at the micro-scale as well. From home automation and self-driving cars, to robots on sidewalks, we are drowning in a chaotic sea of “globally local” small-scale protocols everywhere. Conventional frames, such as traditional urbanism for sidewalk governance, or home construction practices that clumsily attempt to integrate ideas like “smart keys” for package delivery and “smart thermostats” for energy management, struggle to rein in the chaos. The result is proliferating localized Darwinian “technological tangled banks” of hyperlocal infrastructure ecologies that are invisibly connected to the planetary-scale PCT.

Even apparently small, simple, and isolatable examples of protocols, such as hand-washing, become complex to theorize when they are embedded into the PCT.

What happens to handwashing, for instance, when all soap dispensers turn into small internet-connected robots that can order refills for themselves online? What changes in our behaviors when smart faucets start singing happy birthday at us to ensure we spend the mandated twenty seconds washing our hands in controlled environments?

Why Theorize?

Mathematization and formalization have not always been part of human governance responses to major planetary transformations. Computerization was partially theorized early and powerfully, through the development of frameworks such as Turing machines and the lambda calculus. Globalization was accompanied by the rise of fields like operations research and supply chain management that formalized and mathematized critical aspects of it.

Economics, a field that offers many precedents, cues and clues for Protocol Studies, mathematized in fits and starts in the process growing beyond mere description and analysis to become the operating system of the global economy (as Donald Mackenzie powerfully argued in his classic, An Engine, Not a Camera). Along the way, thanks to the methodological demands of economics, rigorous and formal statistical sciences emerged from unsystematic beginnings in gamblers’ intuitions and superstitions.

But other planetary transformation forces have resisted formalization and mathematization. Protocolization lies somewhere between phenomena like computerization and financialization on the one hand, which lend themselves very well to formalization, and phenomena like urbanization and consumerization on the other, which resist paradigmatic treatments more cohesive than ethnographic and sociological analysis.

Some mathematization and formalization is clearly possible, especially because so much of the substance of protocolization is technological. But trying to formalize too much is misguided.

In the second part of this essay we will lay out the particular approach the SIGFPT is taking to this challenge.

FPT Within Protocol Studies

Formal Protocol Theory (FPT) is one of a handful of disciplinary areas within the broader emerging field of Protocol Studies, and in our activities and explorations, we maintain collegial ongoing dialogues with several peer areas, devoted to shared concerns and intellectual foundations.

The current list of areas, and associated Special Interest Groups, can be found here, along with participation information. Below, we briefly describe these adjacent areas in relation to FPT (each meets once every other week).

Memory Protocols

The Memory Research Group (MRG), led by , studies the role of memory in protocols, protocol technologies, and protocolization processes.

Perhaps one of the major contributing factors to the invisibility of protocols is their close relationship to memory and forgetting. The most successful protocols don’t involve user manuals during their operation, but produce and propagate the necessary, operable memory through their performance. In the spirit of the “medium is the message,” we might say that “the protocol is the memory.” For example, blockchains produce a record of their actions as part of their actions themselves, and this record is actually one of their primary utilities. The contemporary era of protocolization is marked by the ever-closer coincidence of procedures and their memory traces.

The MRG studies memory in all its manifestations, and in 2026, SIGFPT will take cues from MRG to look for ways in which we might formalize and theorize the memory features of protocols.

Protocols for Business

The Special Interest Group in Protocols for Business (SIGP4B), led by , studies protocols within organizations, approaching the domains of existing fields like management science, organizational behavior, and organizational psychology, through the lens of protocol theory.

Specialized organizations and sectors that fulfill various societal functions have historically been the locus of a great deal of protocol development and evolution. Technological specialization is another major contributing factor to the invisibility of protocols. Not only are many protocol infrastructures hidden behind the boundaries of organizations devoted to developing and using them, knowledge of such protocols is also hidden behind sector-specific jargon. Even protocols that enjoy a degree of universal presence across sectors and industries and some lay familiarity – such as those involved in managing supply chains, or staff functions such as human resources and facilities management – tend to become invisible behind the boundaries of functional specialization.

Currently SIGP4B is undertaking a protocol-theoretic exploration of perhaps the most important organizational technology in a century: AI. Effective adoption and use of AI technologies is a protocolization problem that is being simultaneously tackled by all organizations across the world in near-crisis mode. There is much to see and learn.

In 2026, SIGFPT will take cues from SIGP4B to help theorize the injection of AI into traditional organizational environments.

Protocol Fiction

Building on a year of publishing dozens of protocol fiction stories here on Protocolized, the Protocol Fiction SIG, led by and , studies the rules and principles of this emerging genre, and explores how to look at our protocolizing world more effectively through stories.

For us in SIGFPT, we hope to find both stimulation and entertainment in the work of this SIG, as well as inspiration and motivation. Fiction can often point to subtle phenomena and generative contradictions and paradoxes that can power theoretical investigations. The notion of “protocol tangles” developed in this essay, for example, was largely inspired by the many tangled plots and premises that have been explored through protocol fiction in the last year.

Besides these three peer SIGs, other SIGs may be spun up this year.

Further Reading

In the second part of this essay we will dive deeper into our early ideas. In the meantime, if you’re interested in participating in SIGFPT, you may want to read these older updates based on our work in 2025.

What is Formal Protocol Theory

Constructing the Evil Twin of AI

We hope to see you this Friday for our 2026 kickoff.

Join us on our Discord for the 2026 kickoff meeting of SIGFPT on Friday, January 9, 10am Pacific. Calendar invite for series.

Acknowledgements
Thanks to Giovanni Merlino, Mike Travers, Patrick Atwater, Seth Killian, Jonathan Moregård, Joseph Michela, and Kei Kreutler for comments and contributions to this essay in draft stage.

But that is the strange story of Bronisław Malinowski, who after a childhood of illness became a pioneer of anthropology. In 1914, Malinowski accidentally began an expedition to Melanesia – documenting social systems in New Guinea, Fiji, New Caledonia – that would return him to Europe five years later, in 1919. You might think that you were born too late for such an adventure, but you were not.

Malinowski was an influential and contentious researcher. The 2018 edition of the International Encyclopedia of Ethnography credited him with shaping the field of ethnography. He was known for stressing the importance of getting “off the verandah” and exploring everyday life. Not all of his colleagues approved of his methodology and it’s unlikely that his opinions have aged well.

What did last, however, were Malinowski’s innovations to the art of people watching.

Fei Xiaotong was among Malinowski’s students at the London School of Economics and went on to have an adventurous, difficult career as an anthropologist in China. Fei embraced Malinowski’s anti-armchair sensibility but disagreed with his claim that social institutions exist to satisfy universal human needs – arguing that social order is more about relational positioning and that Chinese society cannot be reduced to problem-solution pairs.

Today, such intellectual adventures don’t require the overhead of an expedition. You needn’t go far to find strange rules; you might not realize how much you live in your own world. Every day we interface with a tangled bank of technologies, laws, and cultural norms. Paying attention to the world around us is enriching; I would argue that we have far more interesting things to observe than Malinowski did.

Scholarship and profitable application can be an afterthought. Before Darwin was Darwin, he was Charles, a birdwatcher from England. His naturalistic theories emerged from thorough documentation. He was collecting data without a hypothesis!

If you’re unfamiliar, protocol watching can be grokked as a cousin to birdwatching or people watching. Its closest sibling might be umarelling; loitering around construction sites to check out works in progress. It has a simple definition: watch the protocols – from TCP/IP to traffic lights – as they happen around you.

Ethnographic practices might help us all to be more observant in and around our professional spaces.

Software UI/UX designers, bike lane planners, and football coaches alike must study how people interact with each other and with their environments. People in these jobs, however, start with goals in mind. They have an objective: improve user retention, reduce accidents, win more trophies, etc. Protocol watchers are also rigorous, but don’t have the disadvantage of time pressure. They can collect, collect, and collect and let patterns – and profit – emerge naturally.

It’s a potent, if nerdy, hobby. I wrote a first version of this guide last year. Since then, I’ve gotten a much better feel for patterns in how we coordinate, argue, and share information. No course registration required. No sailing ship overhead. From hiking trails and mechanic shops to airport tarmacs and financial transactions, there’s always something to learn or even replicate. It’s an adventure-in-a-lens.

My challenge to you: take up protocol watching as a miniature hobby for the new year. It’s perfect for transforming dead time in airports, traffic jams, and boarding queues into a tad more interesting downtime. It’s also a sneaky way to get off your phone a bit more, if that’s on your list of resolutions.

Some examples to orient your eyeballs:

• Queueing protocols at airports, bus stations, and ferry ports. Who gets to skip the line? Are there different types of lanes?

• Gift-giving protocols at parties. Are you expected to bring something for the host? What purpose do gifts serve? Is the expectation stated or implicit? When do people open gifts?

• Out-of-office messages. What information do people provide? How standardized is it? Is the medium automated emails, their main publication, or their website?

• Hardware standards. Charging outlets, bolt sizes, countertop heights – they afford interoperability in the long run, but is it worth the inconvenience? Who makes standards?

• Transactions. How do you establish trust with vendors? When people get stuck on the side of the road in a snowbank, why is it that others help without expectation of compensation?

It’s important to follow your own interests when protocol watching. Otherwise it will be less of a hobby and more work. Enthusiasm is the bottleneck – Darwin and Malinowski both entertained a deep-seated interest. Treat the examples above, and the tools and tips of the trade below as a starter pack of suggestions. Once you’ve spotted a couple of easy-to-recognize protocols, keep looking and take note of everything you find interesting and potentially useful.

Tools

Since everyone likes tools, we’ll start with those. These will help you build a sharper eye, faster. They’re also the backbone of a compounding hobby. The thoroughness of your notes or working memory will determine how easily you can spot patterns in your observations later. These tools will help you record, detect, inspect, and taxonomize the protocols you observe.

NOTES

At a minimum you should keep notes, whether on paper or digitally. Minimalists and flaneurs can start with just a notebook or mobile notes app – aim to fill up your first notebook as fast as possible.

To level up your game, the best off-the-shelf tool for this is the Protocol Bicorder, developed by University of Colorado professor Nathan Schneider, who also maintains the Protocol Oral History Project. The recently-created Bicorder will help you understand and catalogue protocols via some scoring scales, based on research from the Summer of Protocols program. I recommend you enable the Short Form mode.

Sketching is another nice way to take notes. Whether it’s a physical thing you draw or a choreography at a TSA security checkpoint, we’d love to see what you come up with.

COMMUNITY

When kickstarting a hobby, it’s helpful to chat with fellow hobbyists. Documenting and sharing protocols is also just fun. Our Discord has a dedicated #protocol-watch channel for exactly that. Posting there is low stakes and even lurking will give you a sense for how other people treat this as a hobby. Send a photo, video, sketch, or the JSON file generated from your bicorder entry anytime.

Protocol watching is also more fun with a buddy, kind of like people watching. Provided you have a sufficiently nerdy friend or anthropologist family member nearby, this could make for a good way to kill time.

TENSION DIAGRAMS

Every protocol is an engineered argument and exists to manage an underlying tension. Part of the game of protocol watching is trying to identify that tension. A common one is convenience vs. safety (hello again, TSA). At a minimum, you should seek to name the tension. If you want to level up your intuitions, make a diagram with the following template.

Label the axes with your factors. Draw a trade-off curve. Where does the protocol land? Is it on the frontier? Or can you easily imagine a tweak that would improve, say, convenience AND safety at the same time? Which actors prefer a safer version? Who prefers a more convenient version?

For a detailed example, check out One Tension to Rule Them All.

TRANSLATORS AND TRACKERS

A translator is a must-have tool. For example, the maintenance tags on fire extinguishers and defibrillators are full of jargon, codes, and acronyms. Using an LLM can help you translate those really quickly. You might also see protocol artifacts written in other languages – are there subtle differences not explained by translation?

Trackers are another fun augmentation. They are the telescope of protocol watching. Tools like FlightAware, FIRMS, VesselFinder, ISO standards, and IQAir all provide a unique look at the world. You can track flights, wildfires, cargo ships, food classification standards, and global air quality all from your phone browser.

You can also use LLMs to generate professional briefs and histories of protocols you observe, as and sometimes do with their writing projects, and , respectively.

Tips

Your toolkit should also include some principles and heuristics. These three tips will help you improve over time and might enable your protocol watching to compound from hobby to competency.

First, go for the hotspots. Places with a mix of hardcore engineering and human factors, like airports, libraries, kitchens, universities, factories, mines, farms, legislative assemblies, and so forth.

Second, go through, around, and behind the protocols. Get a feel for what they’re like as a participant. Try to hack them – skip lines, find workarounds. Talk to the people who design and enforce them. By exploring multiple perspectives you’ll develop a much richer perspective or, at least, not become as jaded towards a protocol’s apparent dysfunction.

On a related note, the third thing you should pay attention to: protocols are most visible when they’re not working. As much as it’s fun to catch one breaking down, it’s even more satisfying to spot a protocol that’s doing its job invisibly well.

Congrats, you’re now initiated to the protocol watching club! Beyond the benefits mentioned above (and general aura enhancement), this certification will improve your odds of acceptance into next year’s Protocol School.

Add to LinkedIn

The editorial team at Protocolized is thrilled about all of the stories and studies we published this year. We encourage anyone who wants to write with us to try protocol watching. It will help sharpen your imagination for futures that might logically unfold from new technologies. It will also help you add another dimension to you worldview.

Happy protocol watching and thanks for reading Protocolized. All 84(!) issues from this year are available in archive. See you in 2026 for more stories about strange rules and essays about protocols – the world’s (mostly) invisible infrastructure.

The Protocolized team - Timber, Venkat, Tim, James, Josh, and Jenna

Subscribe now

Between June and December 2025, the Memory Research Group met every two weeks to read texts about memory. The readings played with juxtaposing different historical approaches to memory. We began with classical techniques, then embarked on a digital memory architectures module with contemporary AI systems before moving onto older physical memory substrates again. This allowed us to trace the present’s confusions backward to see which problems are genuinely new and which keep recurring under different names.

The group convened via the Summer of Protocols Discord, with participants from neuroscience, computer science, philosophy, architecture, and design reading texts together in real time before discussing them. We weren’t trying to arrive at a unified theory of memory, but instead trying to distinguish how memory operates in different contexts.

Subscribe now

June: Early Memory Techniques

Reading: Frances A. Yates, The Art of Memory (1966).

The group started with Frances Yates’s The Art of Memory. Yates discusses the method of loci – placing memorable images in imagined architectural spaces – exemplifying memory as active practice rather than passive storage. The ancient Hellenic poet Simonides realized after a building collapse that he could remember where everyone had been seated among the ruins. This became a protocol: associate information with locations in space, and you can retrieve it reliably.

What stood out was how these ancient techniques didn’t oppose knowing with memorizing. Memory was described as the faculty enabling “firm perception of things,” not just the stockpiling of facts. Some 15th century practitioners picked up on this lineage, developing elaborate systems for organizing ways of thinking, not just information. The group noted tensions around trusting external representations versus cultivating internal capacity, a concern that would keep reappearing when we discussed digital systems.

Related to this month’s topics, read the post A Very Short Introduction to Memory and Technology.

July: Distributed Cognition

Reading: Edwin Hutchins, Cognition in the Wild (1995); Michael Levin, ‘The Space Of Possible Minds’ (Noema); Michael Levin et al., ‘Endogenous Bioelectric Networks Store Non-Genetic Patterning Information During Regeneration and Development’ (2021).

Edwin Hutchins’s Cognition in the Wild documented techniques of naval navigation, arguing that memory and cognition distribute across people, tools, and environments. In this context, a ship’s navigation team doesn’t allow crucial information to exist only in one member’s memory. Instead, practices spread across coordinated action and material artifacts. This raised questions about whether memory requires a specific substrate at all, or whether it is better understood as patterns of coordination.

We followed this by reading Michael Levin’s work on bioelectric networks, which proposed that a kind of memory exists in the electrical patterns of cells, not just neural systems. His research on how organisms store non-genetic information during development and pass it on to future generations expanded the working definition of memory beyond neural substrates entirely. The group wrestled with what it means to call something “memory” when it operates without anything resembling conscious recall.

August: Spatial Memory and Medieval Neuropsychology

Reading: O’Keefe and Nadel, The Hippocampus as a Cognitive Map (1978); Mary Carruthers, The Craft of Thought (1998).

O’Keefe and Nadel’s The Hippocampus as a Cognitive Map distinguished between cognitive maps and stimulus-response learning. The hippocampus doesn’t store information about locations alone, but rather builds associative models of spatial relationships. This connected back to ancient memory palaces: spatial organization seems fundamental to how we recall knowledge.

Mary Carruthers’s The Craft of Thought examined how medieval Europeans conceived of memory as residing in physical locations within the body. Memorial landscapes were imagined architectural spaces for organizing theological and philosophical knowledge. This session was guest hosted by Aaron Z. Lewis, who wrote the text ‘Four Doors – A Portal to Sacred Memory Protocols’ for the Summer of Protocols 2023 cohort.

September: Where Memory Lives

Reading: Contemporary neuroscience research speedrun; Mary Carruthers, The Book of Memory (2008).

A research speedrun on contemporary neuroscience was followed by Carruthers’s The Book of Memory, on medieval neuropsychology. Reading about memory stored in the heart, regulated by bodily humors, organized through visual imagery, participants reflected on where they personally experience memory. Most noted that they don’t experience memories as stored anywhere in particular; memories arise in response to contexts and triggers, rather than being retrieved from a location, at least in some of today’s cultures.

Related to this month’s topics, read the post Reflections from Memoria.

October: Digital Memory Architectures

Reading: Amit Bahree, ‘What is KV Cache in LLMs and How Does It Help?’ (2025); Alan J. Smith, ‘Cache Memory Design: An Evolving Art’ (1987).

The turn toward digital systems began with understanding KV (key-value) caches in large language models. These caches allow transformer models to remember previous tokens without recomputing everything from scratch. It’s a computational efficiency trick, but it raises questions about what distinguishes “remembering” from simply having information available. The group discussed parallels to human working memory constraints.

Cache memory systems in 1980s computer architecture followed. Alan Jay Smith’s 1987 overview explained how cache memory bridges performance gaps between the relative speeds of processors and main storage through predictions about which data will be needed next. The principle of locality – recently accessed information is likely to be accessed again – operates across both biological and technological systems.

Related to this month’s topics, read the post Images of Memory.

November: Magnetic Core Memory

Reading: Paul E. Ceruzzi, A History of Modern Computing (excerpts); ‘Magnetic Core Memory,’ Computer History Museum.

Magnetic core memory, an early form of computer memory storage, brought the group into direct contact with memory’s materiality. Reading Paul Ceruzzi’s history, the group examined how early computers stored data in tiny ferrite rings magnetized in different directions. Each bit was literally woven by hand into arrays of wire and magnetic cores. After months discussing more abstract cognitive processes, this tangibility was striking. Software woven into hardware collapsed distinctions between code and physical object in ways that contemporary solid-state memory obscures.

December: Virtual Memory and Metaphor

Reading: Matthew G. Kirschenbaum, ‘Extreme Inscription: Towards a Grammatology of the Hard Drive’; Italo Calvino, Invisible Cities.

The advent of virtual memory marked the shift from physical to abstract storage. Learning about how memory became virtualized in computers – no longer tied to specific physical locations but dynamically mapped and remapped by operating systems – introduced a new layer of ambiguity in our concept of memory. In this context, the illusion of continuous memory space maintained through constant background processes revealed memory more as performance than substance.

The final session turned to Italo Calvino’s Invisible Cities and the question: what core metaphors do each of us use to relate to memory? After months of technical readings, this offered space to share more personal reflections.

What Keeps Recurring

Jumping around chronologically in our readings revealed that contemporary technical systems have rediscovered some of the tradeoffs and ambiguities of memory through different means. The tensions between capacity and retrieval speed, between compression and fidelity, between what to keep and what to discard appear whether discussing core memory in the 1950s, the hippocampus, or transformer attention mechanisms. Sometimes even fundamental mechanics resurface across substrates: magnetic core memory’s destructive reading – where retrieving a state requires overwriting it, before subsequently restoring the original – mirrors how human memory recall performs a recontextualizing function, potentially altering what remains in memory. Still, critical differences remain. Importantly, memory doesn’t oppose human knowing when it’s externalized into tools, because, ultimately, it never operated purely internally to begin with.

Memory operates as a biological process, technological substrate, cultural practice, and phenomenological experience simultaneously. No single framework dominated our discussions, perhaps because of our interdisciplinary composition as a group. Technical explanations of cache hierarchies sat alongside reflections on where memory feels like it resides in the body. Medieval mental diagrams found resonance with neural network architectures. The classical method of loci proved conceptually similar to how operating systems manage address spaces.

Amidst reflections during the last session of the year, our group surfaced a strange irony: sometimes what is remembered most is what is most dynamically changing, through attention, interaction, and reinforcement from its world.

In the medium term at least, fossilization ensures forgetting, whether for cities or neural patterns. That is, until the longer term, in which their presence might be uncovered millennia later, requiring us to map contemporary memories onto their molded matter.

MRG @ Summer of Protocols Discord

Few science fiction authors have been as involved at the frontier of technological innovation as Stanley has. In 2019 he co-wrote a book, AI 2041, with the CEO of Google China, Kai-Fu Lee. That same year, the World Economic Forum tapped him to help imagine possible futures. I had the opportunity to ask Stanley about his process and his work. What follows are some highlights and you can find the full transcript here. The text is lightly edited for clarity. All quotes are marked with timestamps [min:sec].

Protocolized encourages its writers to use LLMs and other AI tools. I’ve been trying to use LLMs in my writing, but just haven’t been satisfied with the results recently. At the same time, I feel an ongoing tension about that – how could I not be finding it useful to draw upon the entire combined archives that were scraped from the open web?

Stanley touched on this feeling early in our conversation:

[10:20] I think it’s not there yet. So you feel like always there’s something missing here, or this character might not seem to be 100% aligned and this dialogue could be a little bit cheesy. There’s always something there – so you will not be satisfied at all.

Despite that, he’s designed a custom workflow that he feels does enhance both his fiction writing and his fiction translation outputs. As a longtime teacher of fiction writing, he’s also able to provide some direction for what to focus on to create actual improvement. Indeed, our conversation did frequently return to the taste that needs to underlie successful LLM use, and his suggestions for how to develop it. Be deliberate, Stanley says, about how they are prompting you:

[03:56] It’s not just you prompting the LLM, but also the LLM as a prompt reversibly triggers your imagination and creativity.

Subscribe now

Where LLMs fit into Stanley’s writing process

Given the amount of time Stanley has spent tuning his use of LLMs for writing, I asked:

[3:02] Where within your writing process do you see there being a use case for LLMs? Do you see it as more relevant during the writing process or editing process? What’s your philosophy there?

His response:

I thought it was just for collecting the data and information and trying to build the structure. But now I think it’s it’s both – it’s when you’re conjuring ideas, when you use it for world building, character building, and structure building. But also you can use it for rendering your style in a way and trying to find a correspondence between different disciplines. For example, when you try to use a very comprehensive metaphor to describe some super complex science theory. An LLM can usually give you some options that you can choose from…

I will use different models for different purposes, and I will separate different tasks. Maybe one big task, you can break it down to different small parts and you give it to this model or give it to that model, or you might want to rerun an output in a different model¹ And you try to find the best outcome. But you have to, I mean, you have to train yourself with some kind of good taste.

This question of taste reappeared throughout the process, craft, and philosophy sections of our conversation. Taste and agency have been two buzzwords of 2025, for anyone tech-adjacent and terminally online.

My question to Stanley, then, was: Can you improve your craft as a fiction writer using LLMs? If so, how? This was a question that was particularly on my mind as someone relatively new to fiction writing. It seems like a skill issue not to have found a good workflow for this. Stanley agreed – improving your craft happens beyond the actual writing practice. An LLM can give you options, and your job as the writer is to have a consistent taste to discern which among them serves your vision.

That means no copy-pasting.

[10:04] I always rewrite. I couldn’t use even a single line it gave me directly. I think in a way there’s some legal issues. But through time and through your career, you raise the bar. You have a higher understanding, you have a deeper understanding of writing as a craft. You have to redo it over and over again. When you find your touch, you find the right feeling. I think that’s the most important thing.

This really reminded me of a college professor I had who swore by manually retyping code snippets from Stack Overflow. “You’ll see what’s wrong with it and change it when you rewrite it,” he said.

My former Spanish teacher said something similar. His rule was that when using Google Translate, even if you know the translation it gives you is correct, you still have to re-type it yourself. And then when you do, you often notice ways in which it was subtly wrong.

Oh, and speaking of translation –

Translating fiction with AI

Stanley and Ken Liu have had a multi-year, two-way partnership in translation – Ken Liu translated Stanley’s Waste Tide into English. Stanley translated some of Ken Liu’s work into Chinese.

[11:46] When I translated some of Ken Liu’s work into Chinese I used an LLM to find the tongue. I think it’s very difficult to capture the style, the tongue. It’s at the inter-linguistic level. You have to find the right slang, for example, or the right metaphor. Sometimes there are some cultural references that might not be able to directly attach with something in Chinese, or the other way around. So I think it’s always useful to use LLMs to help you to find something you know as a counterpart of the the language and the culture. Again, this is about taste. If you are bilingual you can feel the nuance and you can feel the subtlety of the language and the literature and you you might feel, “oh this is not exactly what the author tried to do in the original version.”

So sometimes they do experiments as well. The previous work can be referenced, but it’s not exactly the same, right? So we have to... dive deeper and we try to understand what the author really tried to do. And we come up with this idea, or maybe we could create this kind of style in Chinese to fully capture what the author tried to do. And we try to have the LLM represent what we think might be correct or close enough. But again, it’s not easy. So maybeit’ll give you the mediocre, or an average level of translation.

So I think this is like back and forth. I don’t think you can fully rely on LLM to do the translation, especially on literature.

Again, the role of the LLM here, in Stanley’s view, is largely to present you with a palette of options.

So what’s happening in between, then? What’s happening while you’re redoing it over and over again, prompting the LLMs and letting them prompt you back?

[28:48] As with every tool, you can use it in a dumb way, or in a smart way. You can use it to be lazy, or you can use it to challenge yourself. You think firstly and you see what the AI gives you and you try to compare the two. And you can see, okay, maybe this is what I’m missing here. These are my blind spots and there’s something nice, nicely done there. So maybe this is the focus area I should look into next time. So I think it’s a very good practice for your brain, for your imagination, for your narrative...

This was the thread I was really trying to pull on throughout our entire conversation. What does it look like, to Stanley, to use LLMs in a smart way?

It means being in dialogue with them, and being fluid about their role in your process. I asked Stanley what his mental model is when he uses AI in his fiction process – tool, palette, partner? Something else?

[4:52] It’s inter-subjectivity, so it’s something in between. You can see this subjectivity shifting or flowing between me and the machine in a distributed way. So I think all the magic happens in between. During the process, you know, think about electricity, think about the energy flows.

Interacting with your characters

Generative technologies offer new opportunities to interact with your characters and to discover them, even as the author.

[15:05] I’ve been playing with Midjourney and, recently, with Sora 2. It’s very interesting because when you have a character in your imagination there’s a distance, right. It’s kind of vague, kind of abstract. But when you transform it into a very vivid and and visually stunning image, even an animation, you give it more information. You might be triggered with a sense of how the character acts, or talks, or responds to certain scenarios.

What does it look like, though, to flesh out those characters, and to visualize them within the space of Sora or similar tools? Stanley experimented with assigning characters into the 16 Meyers-Briggs personality types.

[16:49] Now you can divide yourself. You can role play with different personalities. Each type could be attached to specific characters. You can interact and this becomes a role-playing game, like LARP. But also I think in the future, we might use Midjourney or Sora to do a real role-playing game. Everyone could create something like that. It’s not only writing a novel, but also creating something visually. I think that’s the future.

He’s been involved in an immersive drama production in Hangzhou. The premise is that its eight characters are all reincarnated.

[19:25] We have previous lives, we have this life, and the next life. When people buy the ticket, if they pay more they can see different lives. If they pay less they only see this life. So they can choose, just like in Sleep No More in New York. And you can follow different story lines to explore. I found it fascinating because there are some interactions between the characters and the audience. And it might trigger something that wasn’t in the script. Sometimes the character might have to improvise. But from the very beginning, we had like a big Bible defining the world building, and each character, like attached to the different lives. Part of their identity and part of their personality has been changed, like your job, your career, even your appearance, but part of it, like the soul, you know, like some addiction, like some fears, like some anxiety pattern has not. This is a very different way of creating and also a different way of storytelling.

Subscribe now

Keeping the process separate from the product

However, Stanley cautions, you still need to be deliberate about how much information you are providing to your readers.

Unveiling the process is helpful to a point, but the work needs to stand alone. The artist should clearly have a vision about where the work ends and what the reader needs to figure out on their own.

This is one of the aspects I admire most about Stanley’s fiction. As much as the process is important, and as much as he’s been willing to publicly share about it – including in this interview – the work itself is the thing.

[21:07] Maybe there’s some fan fiction – readers can develop the character by themselves, they might have some alternative history, some subplots – like branches, storylines. So they can develop something like that. But I think to me what I should reveal to the readers is already written. You can interpret it by yourself, that’s your freedom. I always say, OK, just figure it out by yourself because what I want to say is already written there.

Even the endings. With a semi-open ending a lot of people ask me, “what does it mean?”

You have to fill in the blank with your own imagination. I think that’s the best part of it. You can keep guessing, you can keep thinking about it. That’s what I try to do.

Shaping the future with writing

Part of shaping the future with writing is about experimenting with form and embracing the new.

I asked about the prospects for young, emerging authors:

[48:59] They are suffering because there’s no platform at all. There’s no publication. There’s no opportunity for them. So I think this is very, very cruel. People feel angry. People feel that they’re outliers, that they’ve become irrelevant in a way, right. I don’t see an easy solution here.

But I always told my students, you have to create your own job in the market. Maybe it doesn’t exist yet, but you have to start to get down to work and start to build your own career, not just waiting there, waiting for some companies, you know, some bosses to give you one, to offer you one. I think this is the reality.

And with AI, I think, yes, they can do it, but they need to refresh their experience. They have to unlearn whatever they’ve been taught, you know, from the traditional schools or parents or educational institutions.

In his own work, Stanley has been very proactive about forming partnerships with people directly making decisions that impact the future.

While he’s a sci-fi writer, he has been invited give guidance in technology and economics discourses.

So what perspectives does he find it useful to share in those forums?

[56:36} We’re so small, we’re so powerless. This is a structural asymmetry. I think punk, like solar punk, basically is what we need here. And how we should do that starts from creating new kinds of metaphors. Because metaphor is how we package and compress all this storytelling, emotion, ethics, values, mythology all together. It’s like a super efficient info pack. But it takes a very long time to form a metaphor. So that’s why we should start now. We need some time to build up new metaphors…

I try to shift the vision for these people because they are the one who spend the money, make the action and policy-making, etc. I think you should do something to change the narrative, change the metaphor and tell a new story and create a new myth for the future.

A lot has changed in climate protocols in recent years. In 2023, when I first gave a talk on this topic for Summer of Protocols’ inaugural research cohort, the two previously disconnected worlds of traditional carbon markets and DeFi had only recently collided. The main character in that clash, KlimaDAO, had proclaimed itself as a “black hole for carbon” upon its launch and sucked 17 million metric tons of CO2 from suit-and-tie carbon market registries like Verra into its treasury, thus transforming millions of carbon credits into a new zombified onchain form that neither the corporates nor the degens knew what to do with. Regen Network, perhaps the oldest player operating at the intersection of crypto and climate tech, was in the process of expanding beyond its focus on carbon, taking in other environmental markets such as biodiversity crediting.1 While it may seem strange to be designing markets for biodiversity when we still can’t seem to get carbon markets right, biodiversity crediting has managed to attract attention from international policy players such as the United Nations Development Programme and UN Environment Programme through efforts like the Biodiversity Crediting Alliance. Regen Network soon found itself partnering with some very promising groups working in this sector, from the Colombian habitat bank developer Terrasos to indigenous leaders from the Sharamentsa Achuar community in the Ecuadorian Amazon.

Regen Network’s pivot to broaden beyond carbon stemmed both from a desire to resist carbon tunnel vision, and a response to a failing voluntary carbon market. The more traditional subsector of the carbon markets – consisting mostly of reforestation and avoided deforestation credits – had continued to face challenges, both from the plummeting corporate demand, which culminated in the Guardian’s exposé denouncing Verra’s REDD+ program in early 2023; and from continued slow progress instrumentalizing Article 6 of the Paris Agreement – a key piece of the international climate treaty that aims to clarify how nation states can trade carbon credits against their NDCs (nationally determined contributions towards limiting CO2 emissions).

Durability: A New Focus for Carbon Markets

Despite the world’s complete failure to put a meaningful dent in greenhouse gas emissions, over this same period of time a new and more focused carbon market in durable carbon-dioxide removal started gaining traction in both the regulatory and corporate sustainability worlds. A significant marker of this shift was the launch of Frontier in 2022, a $1B+ advanced market commitment for durable carbon removal led by Stripe, Alphabet, Meta, Shopify and McKinsey. This renewed focus on more permanent carbon removal and storage technologies (often abbreviated as CDR) represents a rebranding of the legacy carbon markets away from greenwashing and questionable “counterfactual claims” of avoidance-based forest carbon credits, and toward scientifically verifiable removals where the CO2 sequestered can be safely assumed durably stored underground or in the depths of the ocean for hundreds or thousands of years. The approaches and technologies being researched and deployed in the CDR sector vary widely, from highly engineered direct air capture facilities, to spreading pulverized volcanic rock dust over agricultural fields so that it reacts with and absorbs CO2 under rainfall, to electrochemically removing CO2 from the surface level of the ocean or otherwise altering seawater chemistry to enhance the ocean’s capacity to absorb CO2 from the atmosphere.

If this all seems bewildering and daunting, trust that it’s hard to keep track of existing and emergent CDR pathways even for those working in the industry. One could ask why we need so many different strategies for sequestering carbon out of the atmosphere. The answer to this question is, in principle, pretty simple. We don’t know which strategies will scale, and we need to scale very quickly. In 2023, the Intergovernmental Panel on Climate Change (IPCC) released its synthesis report, stating that, even in best case scenarios for emissions reduction, an estimate of 5 to 13 gigatonnes (Gt) of carbon dioxide removal (CDR) will be needed annually by 2050 in order to limit global warming to 1.5°C and achieve net-zero emissions. So, at this point in time, the industry consensus is that we need to be funding, researching, and exploring all feasible pathways, with humility and understanding that any potential solution may, at any time, be deemed uneconomical, unsafe, or otherwise unfeasible.

A Protocol with a Thousand Faces

Protocols and carbon markets have had a complicated entanglement since the earliest days of international climate agreements. The predecessor to the Paris Agreement was explicitly titled the Kyoto Protocol, and prior to that was the Montreal Protocol (often touted as an early success of international environmental agreements). In contemporary carbon markets, the term protocol has evolved from being used to designate a singular stamp of overarching international consensus to instead operating like a Swiss Army knife of sorts. Whether in the corporate sustainability arena of voluntary carbon markets, or the nascent but rapidly evolving jurisdictional and international compliance markets, if there’s a problem with clearly defining something, or a lack of consensus regarding what exactly something means, slap the term protocol on it! The word itself becomes elastic, with varying focus and specificity that can stretch and shape to any context. In the context of today’s voluntary carbon market, the industry’s self-appointed governing bodies, known as “registries” (e.g. Verra, Gold Standard, Isometric, American Carbon Registry) publish protocols – programmatic rules which dictate what kinds of activities need to be undertaken to warrant issuing carbon credits, and how those activities must be monitored, reported and verified (a process which is abbreviated as MRV).

In this wild-west, funded predominantly by corporate do-goodism, competing registries beget competing protocols. Furthermore, with a seemingly boundless set of different sequestration pathways being explored, a key indicator of pathway maturity is how many registries have codified a protocol for a given pathway. Without any infrastructure in place to compare projects across different protocols, the market remains fractured. Prices vary across pathways and protocols and the industry has little consensus on which approach is right. Toss in the fact that we are still waiting on UN policy guidance detailing which methodologies (the UNFCCC’s term for MRV protocols) are to be approved for use in country-level trading of carbon balances under the Paris Agreement Crediting Mechanism, and it’s easy to see how neither the voluntary nor compliance markets can make up their mind about what’s good enough. The Article 6.4 Supervisory Body is still working through this task nearly ten years after the Paris Agreement was signed.

A core tension here is that despite failing to provide any concrete pathway to fungibility for carbon financiers (demand is still concentrated in a handful of players, and supply too heterogeneous), these protocols supposedly exist for the explicit purpose of making real-world carbon removal interventions legible for financial markets. In this way, to ask whether the MRV protocols of Verra, Gold Standard, Isometric or Puro function more as scientific protocols or financial protocols is a question of perspective. Scientific for sure, but their function as financial protocols remains aspirational at best.

Data Protocols Behind the Curtain

With the industry’s spotlight on durable CDR, the resulting diversity of pathways and technological approaches to carbon removal is likely the main culprit in our current protocol predicament. If carbon markets were still dealing only in forest carbon, the industry’s first attempt to solve inter-registry interoperability with the CAD Trust might have delivered on its promise. That story begins in late 2022, with a group including IETA (International Emissions Trading Association) and The World Bank launching the Climate Action Data Trust as an attempt to harmonize data schemas from the various carbon market registries. Key to their approach was developing a shared meta-registry: a new database for existing registries to report their own project and transaction data to that could prevent double counting and support interoperability between the various different carbon registries. As with many other double counting prevention initiatives in enterprise or business arenas, there was a blockchain component, and Chia Network was to serve as the blockchain technology partner in the project.

Fast forward to 2025 and CAD Trust still exists, but with the durable CDR market dealing in carbon credit offtake agreements the size of a Series C financing round (see Frontier’s $58M deal with BiCRS developer Vaulted Deep), the CAD Trust approach of replicated harmonized meta-registry has done little to provide support in an area where much of the current market activity is in future-oriented offtake agreements. On top of that, CAD Trust’s focus on technical API integrations has made adoption slow among legacy registries like Verra and Gold Standard, and nonexistent among the younger durable CDR-focused registries like Isometric or Puro.

One thing CAD Trust did get right from the beginning was bringing data standards into the spotlight. In the years since, several other initiatives have followed their lead. The Carbon Data Open Protocol (CDOP) is likely the carbon market data standards initiative with the most industry traction today. Launched as a multi-stakeholder initiative, led by RMI, S&P Global, and Sylvera (a prominent carbon credits rating agency) among others, CDOP focuses on harmonizing data schemas and vocabularies between the many different actors in today’s carbon market. Arguably more important than what CDOP is, is what it is not. By electing not to build any infrastructure or host project and transaction data themselves, CDOP sidesteps the risk of becoming another value-extracting intermediary in an already crowded market, leaving open the possibility that the project could get enough traction to serve as a genuine interoperability layer.

In carbon markets, where data isn’t at risk of disappearing and transactions are often built upon trust between known parties, interoperability initiatives don’t necessitate the fully-replicated data infrastructure that blockchains provide. With this in mind, any initiative claiming to enable interoperability in carbon markets should focus its energy on developing open standards, exchange formats, and governance frameworks rather than building its own infrastructure and risk becoming yet another silo to navigate. For CDOP this approach is more than an implementation detail, it’s deeply embedded in the initiative’s branding and overall strategy (see their July 2025 blog post Carbon Markets Need a Berners-Lee Moment for Data Standardization).

Will this latest attempt to protocolize the carbon markets improve our ability to meaningfully scale carbon removal? Early signs suggest momentum is building. Even ISO is paying attention now, having initiated, in October 2025, a working group to develop a standard Data Model for Carbon Credit Markets, in which both CDOP and CAD Trust will have membership as nominated experts. If the scientific, financial, and data infrastructure continues to mature, the foundation will be there to support the gigaton scale CDR needed by 2050. Leaving one question remaining… Will the incentives and resources align fast enough to make it a reality?

Subscribe now

Cory Levinson is a software consultant working on data standards in the carbon removal industry, serving as technical lead for Carbon To Sea’s Ocean Alkalinity Enhancement Data Management Protocol, working with the ocean science consultancy Submarine Scientific. He is an active member of the Carbon Data Open Protocol’s Technical Working Group, and is currently pursuing an MSc in the Soils & Biogeochemistry Graduate Group at UC Davis. Views expressed here are the author’s own and don’t necessarily reflect those of any affiliated organizations.

This text is based on a talk which Cory gave for Summer of Protocols in July.

Update on our current writing contest

We’ve extended the deadline for Building and Burning Bridges, our Bridge Atlas-inspired short story contest. The judging panel has now assembled and includes Nils Gilman, who appeared in episode 2 of the Bridge Atlas salon series, and previous Protocolized contest winner Spencer Nitkey, as guest judges.

Your story of diplomacy, bridges, and espionage is now due December 8, 2025.

This means that you’ll have one more opportunity to get some community feedback on your work. Join our Special Interest Group on Protocol Fiction for its next call, December 4 at 10am Pacific Time. More information in our public Discord server’s #protocol-fiction channel. We’ve provided thorough contest instructions here, and there are some bonus ideas hidden below to draw on.

After seeing the first three protocol fiction anthologies in print last week, we’re particularly eager to see this next thematic collection come together. This contest is our most complicated yet. We can’t wait to read some tales about what the future of diplomacy might be like – and to assemble another collection.

It’s time to lead the next renaissance in science fiction. We are on the cusp of so many new, interesting technologies that have not yet received thorough treatment in fiction. Cryptography, GLP-1’s, rapid vaccine development, glowing plants, cheaper-than-gas solar energy, universal translators, supersonic transport…

Protocol fiction, which focuses strictly on special rules over special people, is a good way to take a shot at telling stories about these strange new developments in science.

Plus there are tremendous prizes for the Building and Burning Bridges contest, which closes December 8, where you can tie in some of these technologies.

Enter a Story

Meeting with protocol thinkers

The cafes of Paris, Vienna, and the rest of Europe were lighthouses of intellectual activity. People would discuss economics, philosophy, and industry, and drink coffee in quantities inadvisable by modern health standards.

In a protocol-ish twist, we organized a meetup in a small, industrial coffee roastery. During our Bridge Atlas event at last week’s Devconnect, we invited a mix of participants from Buenos Aires and some visiting from abroad to share and discuss ideas. Of course, the first topic was coffee production – participants watched the roasters in action as they walked everyone through their protocols during a small batch. Like listening to someone explain protocols for poetry or music, it was surprisingly technical. The freshly roasted beans were then bagged and taken home by attendees.

The meetup included two panels that discussed how protocols could be used to bridge divides between technology and culture, then debated how the protocolization of migration and mobility will continue. Some key highlights:

• It’s hard to explain how the world works (or fails) and simultaneously tell a good story.

• The work of translating technical features of a protocol into marketable features is an artistic practice.

• Just as there are safety protocols in place for internet users, there must be safety protocols for AI users.

• Citizenship has scaled from cities to states – rules for planetary citizenship are creating a future which we’re already seeing glimpses of.

• Traditional diplomats are on the back foot. New side channels continue to open up and more non-state actors are getting involved in a permissionless fashion.

We’re still working on making a distributed Vienna for protocol studies happen. If you do research as a hobby (and you should!) the Summer of Protocols Discord is a unique place to hang out and share caffeine-powered ideas.

Running a diplomacy workshop

Devconnect is a biannual conference for developers who work on the Ethereum protocol. We attended as well, to learn about frontier developments in programmable cryptography, and to facilitate a workshop on ecosystem diplomacy.

Governments, standards bodies, and corporations create critical infrastructure – often together – and must manage tensions. Technology ecosystems (especially those that maintain protocols like Ethereum, TCP/IP, Linux, or even how we decide when to wear medical masks) must also build bridges with the rest of the world. Unfortunately, technologies can’t hire or appoint diplomats. People instead choose to take on that job.

We put on a full-day workshop on ecosystem diplomacy. It was based on three years of research from Summer of Protocols and especially our recent Bridge Atlas series. Why diplomacy? Isn’t that just for governments?

The workshop talks introduced some useful tools and are available on the Ethereum Foundation’s YouTube channel. Watch the talks here. Get the slides here.

No – even a technology like steel production needs diplomacy. For example: there are strong cultural and political factors at play when a company chooses a location for its next plant. People’s relationship with work and with employers varies a lot between cultures. That relationship can directly affect production schedules and planning processes. And that’s just to make metal, where causality flows mostly one way: from culture to the technology (and its implementation).

Protocol technologies like encrypted messaging have a deeper political element. The causality flows both ways, which means stakeholders have, well, more at stake. A tool like censor-resistant social media will affect culture in hard-to-predict ways. Making new markets and customers isn’t just about interfaces or features, but increasingly about diplomatic groundwork.

Summarizing the Bridge Atlas salons

If you haven’t already, be sure to listen to the series of salons we recorded leading up to the ecosystem diplomacy workshop at Devconnect. There are five episodes, each featuring a bridge-building conversation between one member of the Ethereum ecosystem and an external counterparty.

• Episode 1 - Intro: Tim Beiko and I discuss what this series is about, how Summer of Protocols approaches bridge building, and why protocol thinking is a good thing to get good at.

• Episode 2 - Paradigms: Josh Stark and Nils Gilman talk about hardness and planetarity, two extremely useful terms for anyone interested in protocols. Plus, they explore applications for programmable cryptography at unprecedented scales.

• Episode 3 - Commons: Trent Van Epps and Yancey Strickler go deep on how to protect and grow a commons so that everyone can win, by playing the long game. Between Van Epps’ experience with Protocol Guild and Strickler’s experience having founded Kickstarter and Metalabel, there’s a lot to take away.

• Episode 4 - Alignment: Former CEO of Twitch and OpenAI – and current CEO of Softmax – Emmett Shear joins Alex Stokes of the Ethereum Foundation to talk about alignment and why most people get it wrong. Protocols are a more appropriately scaled approach to the problem.

• Episode 5 - Verifiability: Shreya Shankar, creator of DocETL, and cryptographer Justin Drake explore the importance of verifiability in both AI and blockchains. They use the underlying similarities they discover to build a bridge between these two important domains.

Christine: Justin, can you start us off by introducing the lean Ethereum
project and what that entails?

Justin: In lean Ethereum, we’re tackling verifiability in two ways. First, for deterministic blockchain computation, instead of re‐executing transactions on thousands of nodes, we use cryptographic SNARKs – short proofs a single person produces and everyone else can verify in milliseconds. Second, for the non‐deterministic consensus part, we aim to speed up voting to achieve finality in seconds instead of minutes.

Christine: Shreya, I saw you smiling when Justin mentioned SNARKs. Is that something you are interested in as it relates to verifiability?

Shreya: I smiled because I learned about zk-SNARKs early in my PhD – they’re very cool! My own work is on DocETL, a declarative system for unstructured data processing. For example, a public defender might mine court transcripts to find racially biased statements. This requires reasoning beyond traditional databases, so DocETL lets users write MapReduce-style pipelines with operators defined in natural language and executed by LLMs.

Verifiability for us means: how do we ensure AI outputs align with human expectations? We optimize pipelines for accuracy and cost, sometimes replacing LLM functions with deterministic code. We also expose statistical primitives so users can verify samples and estimate worst‐case accuracy. Provenance – establishing where in the source particular data statements came from – and understanding AI reasoning are important too.

Subscribe now

Justin: That reminds me of my ‘spectrum of trust’ model. On one end you have math, physics, cryptography – things you trust completely. On the other, legal systems and reputation – mushy, subjective. Blockchains aim for trust minimization, pushing everything toward the ‘trustless’ end. Dispute resolution in marketplaces remains hard; maybe your work could help build trustless marketplaces without escrow agents.

Shreya: How are disputes handled now?

Justin: In OpenBazaar, buyers and sellers agree on a trusted escrow agent. It’s reputation‐based and can be inefficient. Agents can also be tempted to collude for large transactions.

Christine: Shreya, talk about AI evals in verifiability.

Shreya: We annotate AI outputs to identify features correlated with human disagreement – missing definitions, odd prioritization, etc. – until the distribution of these ‘qualitative codes’ stabilizes. Then we can train LLM judges to evaluate outputs at scale and measure misalignment.

Justin: Decentralization in Ethereum is similarly hard to measure. The Nakamoto coefficient – how many operators control 33% or 50% of stake – has limits. SNARKs lower barriers to entry so even small devices can verify. For non‐deterministic aspects, economic penalties ensure that bad behavior would destroy large amounts of staked value. Shreya, how are adversarial actors handled in AI?

Shreya: We see two threat models: malicious users crafting pipelines to prove biased points, and non‐malicious users blindly trusting AI outputs without knowing counterfactuals. The latter worries me more.

Justin: In SNARKs, we counter maliciousness with interactive proofs – multiple rounds where the verifier challenges the prover – and ‘hinting,’ where the prover supplies extra information to make verification easier.

Shreya: I can imagine combining those for DocETL: for example, AI showing users samples it didn’t find interesting and asking them to label them.

Justin: Hinting parallels a teacher giving students clues.

Shreya: In data science, negligent or biased analysis isn’t new – but AI’s agency over vast amounts of unstructured data makes the problem more urgent.

Justin: We have recursive proof composition – combining fast but large proofs with slow but small proofs to get both speed and compactness. You could apply a recursive AI approach: one AI isolates snippets, another evaluates them.

Shreya: We do something similar, decomposing problems into units different AIs handle well, and measuring accuracy after rewrites.

Christine: Is there a tension between verifiability and efficiency?

Justin: In Ethereum rollups, each flavor trades off differently: optimistic rollups have long challenge periods; zk-rollups require heavy upfront proof building; TEE rollups trust hardware; committee rollups trust members. Energy limits mean we can’t snarkify all AI – so we must be selective.

Shreya: In AI data systems, computation vs. resource cost trade‐offs are constant. DocETL partitions tasks to minimize cost and maximize accuracy – using deterministic functions for simple tasks, reserving LLMs for reasoning.

Justin: That’s like the co‐processor and glue model: specialized processors for repetitive tasks, flexible ‘glue’ for unstructured problems.

Shreya: Yes, for example filtering transcripts deterministically to discard irrelevant text before using LLMs.

Justin: In SNARKs, a team built auto pre‐compiles – machine‐generated specialized processors – which sometimes outperform human‐written ones, even for hash functions. … Blockchains would seem to provide an ideal money for AI because AIs don’t have access to bank accounts, but they can have access to cryptocurrency. When you combine that with the presumption that AIs are going to be so much smarter than we are, it’s only natural to presume that AIs are going to be the richest entities in the world and they’re going to be storing almost all of their wealth in cryptocurrencies. What kind of emotions that does that invoke in you?

Shreya: Giving AI currency could create alignment incentives we currently lack. But what happens if AIs hoard wealth? What can they spend it on? In data analysis, I’d pay more for better AI, but that could be costly.

Christine: Could you hardcode AI to value a cryptocurrency as its ultimate goal, even without things to spend it on?

Justin: Network effects could make such money valuable if other AIs accept it for tasks. I foresee a future where AIs pay humans to do things they can’t.

Shreya: True – AIs will need compute, memory, and human labor for certain tasks.

Justin: Some jurisdictions might grant AIs legal entity status, enabling bank accounts. We’ll see experiments on whether that leads to better or worse outcomes.

Christine: What are some sibling concepts to verifiability?

Justin: In blockchains, reproducibility is key: nodes re‐execute to get the same result. Bugs are mitigated either through client diversity – independent teams producing matching results – or formal verification, proving software has zero bugs. Formal verification is hard for AI, but AIs can help formalize SNARKs by tackling sub‐problems in parallel.

Shreya: In AI, reproducibility is often misunderstood. Non‐determinism isn’t bad – it’s a feature – but we may need statistical claim reproducibility rather than exact output reproducibility.

Christine: Closing thoughts?

Shreya: I’m excited about hints – using them to make verification easier has clear analogs in data processing.

Justin: I’m encouraged by efforts to turn mushy realities into black‐and‐white outcomes. Maybe Shreya’s research could replace referees in sports or escrow agents in marketplaces, solving the last 20% where humans are still needed.

If you’re reading this newsletter, then there’s a good chance you have a memory of the scene from 2001: A Space Odyssey in which astronaut Dave Bowman, alone aboard the spacecraft Discovery One, decides to dismantle HAL 9000, the computer system controlling the ship. He goes to the memory terminal and, one by one, proceeds to remove glowing tapes. Meanwhile, HAL sings a simple song that falters and fades out, like an animatronic toy with dying batteries.

This is one image of memory we have in pop culture. Other shared images of memory related to digital technologies might include a USB on your keychain or, for some of us, the translucent, colored memory cards of early game consoles. There might be a faint archival image that comes to mind: people standing in large rooms with terminals spinning magnetic tape or, in other words, an image of an early computer system.

Given the large degree to which metaphors from digital technologies shape our understanding of memory, it’s surprising that our images of memory are so abstracted from how humans experience memory. While we might conceptualize our memory as hard drives, for many of us, the actual experience of recall can feel much more muddily intuitive than browsing a digital filing cabinet of the mind.1 These images often suggest that memory is a discrete object, unlike the well-worn path around a mountain that remembers, through its form, those who walked on it.

To explore these nuances in our images of memory, the Memory Research Group has been surveying different digital technologies and our metaphoric relationships to them.

We naturally looked at one of the hardest problems first: how memory works in LLMs. We read the blog post What is KV Cache in LLMs and How Does It Help? Most LLMs use KV caching: the technique of storing previously computed data in the form of key value pairs during a conversation. When a person sends a new prompt in the conversation, the LLM can access data previously generated in its response without having to compute it again. This makes the computation costs required for an ongoing conversation with an LLM scale linearly rather than quadratically.

Without KV caching, LLMs would be far too computationally expensive to work in the way they do today. Applying the analogy to a human conversation, it would be as though every time you spoke a new word, first you would have to silently rehearse every prior word you’d said so far. It would take so long, and be so redundant, I’m not sure anyone would have the patience for conversation.

But humans aren’t LLMs, and LLMs aren’t humans. Their memory problems are far from solved.2 In one example, after about 200 conversational exchanges between an LLM and a person, the KV cache storage size in gigabytes can be greater than the memory required for the transformer model weights. It’s still a bottleneck, and the blog post we read summarized current approaches to solving it.

While memory storage in LLMs clearly has quantity issues, it also has quality issues. In my experience, an LLM’s memory often feels less like persistent contextual awareness and more like they’ve been primed to answer in superficially relevant ways. (How many times can they paste my gardening interests over my questions about digital memory architectures?) There are multiple emerging strategies other than KV caching that try to address these issues across conversation windows. Yet with memory problems in LLMs, we’re approaching new computational terrain, which can be better understood within the history of computer architecture development.

So, in our next Memory Research Group session, we explored the concept of the cache in early computer architectures. We read Berkeley computer scientist Alan Jay Smith’s ‘Cache Memory Design: An Evolving Art,’ published in 1987. Written at a moment when caching was maturing from an engineering experiment to standard architectural practice, the paper highlights how cache design remains a balance between cost, complexity, and performance.

Caches are small, fast memories that store recently used data so it can be quickly accessed later. Researchers first invented the concept in the late 1960s, based on two simple observations about how programs actually behave: they tend to access the same data over and over—referred to as temporal locality—and they tend to need data that’s stored near each other—referred to as spatial locality.

A library is the metaphor commonly used to explain caches. Imagine you’re working on a research paper. Rather than making constant round trips to the library every single time you’re looking for information, you check out the books you think you’ll need, and you keep them on your desk at home. In this example, your desk becomes the cache, allowing you to quickly browse the most relevant books. You can only keep a limited amount of books on your desk, though, because you don’t have as much space there as the library has in its stacks, illustrating why a cache is fast but limited.

A library’s layout also mirrors this principle. Libraries keep frequently checked-out books in a small cabinet near the front desk, while they move older, rarely needed books to off-site storage. In this example, using prior data to predict usage, libraries give the spatial locality and temporal locality of frequently checked-out books priority.3

In the time that’s passed since Smith wrote the paper we read, caches have evolved into sophisticated hierarchies, and research increasingly explores using machine learning to choose which data to prefetch. The concept has also become useful for various tasks beyond a computer’s central processing unit. Web browsers cache data from websites you visit frequently, content delivery networks cache popular content closer to users geographically, and databases cache frequently accessed queries. Now LLMs also depend on caching. Anywhere there’s a speed-versus-storage trade-off, there’s likely some form of caching happening.

Subscribe now

In our Memory Research Group discussion on cache memories, we found ourselves immediately translating the concept back into the spaces we actually inhabit. One person described their kitchen: the counter holds ingredients within arm’s reach like the cache, while the pantry requires slowing down to retrieve less frequently used items, a kind of “main memory,” analogous to the slower hard discs used in computers. Another person reflected on their studio practice, noting how different zones of the room hold different “heat” depending on when work was last touched. Their projects flow between spaces according to temporal patterns, like cache locality principles.

The insight about hot and cool zones in a studio made me think about dust. Dust reflects physical entropy. The more dust on something, the less recently it’s been used. This acts as a natural “least recently used” indicator, a common algorithm used by cache memories to choose which data should move to slower storage. It’s how we know without any formal tracking system what hasn’t been touched in our personal archives, even our public ones.

As soon as you start looking at images of memory through digital technologies, especially computer architectures, you see both the connections and the discrepancies in these metaphors. Caches optimize for speed, but sometimes the salience of a memory doesn’t reflect its deeper helpfulness to us. Dust tells us what we haven’t touched, but it can’t tell us why we’ve avoided touching it. In 2001: A Space Odyssey, the spacecraft world of HAL appears dustlessly sterile. More muddy images of memory, often found through software rather than hardware metaphors, give us richer language for the human experience of remembering. Environmental metaphors, like the well-worn path around a mountain, come into view again.

As the synthetic memory maker in Blade Runner 2049 explains, “They all think it’s about more detail. But that’s not how memory works. We recall with our feelings. Anything real should be a mess.”

The 1987 paper we read captured caching at a moment of transition, and we’re in another one now with LLMs that may greatly reshape images of memory in the future. It’s a change we don’t really have an image for yet.

For the next Memory Research Group sessions, we’ll continue looking at technological transitions, from magnetic core memory to the advent of virtual memory, before closing the year with a session focused on the open question: What metaphors for memory await us next?

Christine: Let’s start with some background. Emmett, you’ve spoken before about agent alignment. Could you give our listeners a quick summary of your focus?

Emmett: Sure. My background is in internet entrepreneurship. I started Justin.tv, which became Twitch, and had a brief stint as CEO of OpenAI during a turbulent period. That experience made me realize AI had an unanswered question: what does it really mean to align models, and to what? So I founded Softmax, an alignment research company.

We run multi-agent reinforcement learning simulations to study alignment – not as a static state, but as a protocol and ongoing process. We see alignment as coordination between agents at multiple levels, with shared success as the foundation. If one wins while the other loses, there’s no real alignment. Shared success enables specialization and team coherence. But you also need enforcement to protect alignment from parasites or bad actors. Enforcement isn’t what creates alignment – it protects what already exists. Softmax focuses on building shared success and coordination, not enforcement.

Christine: Alex, tell us about your work at the Ethereum Foundation.

Alex: I’ve been at EF for a while, focusing on coordination – essentially our version of alignment. Ethereum’s decentralized ecosystem contains many actors with different interests. We want everyone “rowing in the same direction” toward a shared Ethereum success, but that’s tricky when interests diverge. My job involves thinking about protocol direction – maintaining and securing today’s Ethereum, and scaling and evolving it for the future.

Listen to the full episode

Christine: Emmett, you said alignment starts with shared success, but it’s hard to get that across a large group. In Ethereum, breaking down into smaller local groups might help, but then alignment becomes relative and harder to evaluate. How do we evaluate good alignment in large groups?

Emmett: Alignment is relative – like mass or speed are relative to a frame of reference. You can be aligned with your family or company, but not necessarily with everyone. Systems often have chains of local alignment that feed into global alignment. Crypto does well at creating global shared success – Bitcoin is almost pure shared success. But crypto struggles with multi-scale alignment and specialization, because it flattens context into one global ledger. We haven’t figured out how to support boundaries for private information while keeping global coordination.

Christine: Alex, do you agree that crypto struggles with specialization?

Alex: Partially. Bitcoin has a small “alignment surface area” – one token, one function. Ethereum’s is larger and more complex. But we already see specialization via subtokens, DeFi apps, and rollups – smaller ecosystems connected to Ethereum’s base layer. These are counterexamples to the idea that crypto can’t support localized alignment.

Emmett: Yes, Ethereum’s hierarchy of boundaries is a form of alignment. But the real economy is messier, with overlapping hierarchies and constructive ambiguity. In Ethereum, ownership is unambiguous by design, which is great for decentralization but loses some of the flexibility ambiguity provides in real-world systems. Forking is the crypto equivalent of revolution, but crypto doesn’t embrace frequent forks – a rigid, fork-resistant global chain could be dangerous.

Alex: I see your point about rigidity. But I think crypto pushes power to the edges – users choose to run the software, and forks do happen. Even a “too big to fail” chain could be forked eventually, like governments that seemed unchangeable. Ethereum changes its protocol regularly, which keeps flexibility.

Emmett: Imagine Ethereum runs the global economy – every government and corporation depends on it. Forking then becomes almost impossible. My challenge to crypto is: how can we embrace forking without destroying shared success, allowing multiple interpretations of the same ledger?

Alex: This is part of Ethereum’s rollup-centric roadmap – the base layer stays neutral, while rollups provide flexibility. Decentralization and permissionlessness ensure no one can unilaterally control the chain. Mechanisms for exit – whether via forks or new chains – are crucial.

Emmett: But how is Ethereum making exit easier today?

Alex: Right now, exiting means leaving Ethereum’s L1, but there’s not yet an alternative system to exit to. That’s why rollups are important – they can offer localized ecosystems.

Emmett: Ethereum isn’t just tokens – it’s smart contracts and identities. If the whole economy runs onchain, exiting becomes extremely hard without early planning. That’s why I’m concerned: decentralized systems can get stuck in bad ways, especially if AI agents store property there.

Alex: We could design smart contracts today with exit functions. Early national ID onchain systems will likely be centralized, so they can patch bugs. Fully decentralized ones could get stuck, as you say – but decentralization also makes exit easier because there’s less centralized control.

Emmett: You gain the ability to exit centralized systems in exchange for the inability to exit decentralized ones – that trade-off might be worth it. But with AI, there’s a risk in trying to encode “the good” as a fixed meta-ethical loss function. You can’t perfectly define the good. Alignment needs openness and ambiguity, letting agents rewrite even their base values. A unified global meta-ethics in AI could be as dangerous as a rigid global blockchain protocol.

Christine: Ethereum changes core assumptions frequently – unlike Bitcoin, which has hard-coded values. Could alignment even be possible with a protocol that changes that often?

Emmett: Yes, because survival in the same reality acts as the ultimate arbiter. Systems that believe falsehoods eventually die or adapt to align with reality. Even groups with vastly different beliefs (like Catholics and Protestants) share reality and can find common ground.

Christine: I’m skeptical that man-made protocols can self-correct dynamically like that over time – whether crypto, AI, or even the internet.

Emmett: That’s why I oppose any global protocol. The real game is constant invention of new value systems and protocols. A healthy ecosystem bubbles with new protocols, with no single durable leader. The danger comes if crypto or AI ossifies into one global layer that must be “right” forever.

Subscribe now

Christine: Final thoughts? For me, if Ethereum ossifies, it dies. Technology always needs intervention. And as a side note, I see some protocols – like the Ten Commandments – as timeless truths, though that’s outside our scope.

Alex: I’m optimistic about Ethereum because of the people involved. The tech is an extension of human values, so misalignment can be corrected. AI is riskier because it could run away from human control.

Emmett: I think crypto will grow fast, especially as AI agents adopt it. That could lead to a global chain – which is why we need to think now about maintaining exit rights if the whole economy ends up onchain. We should also think early about AI models that may surpass human intelligence and influence decisions beyond our audit capacity. My criticism comes from optimism – I believe these systems will succeed, so we should plan for winning scenarios.

The first 70 issues of Protocolized were illustrated with art generated by Midjourney, referencing a moodboard stocked with visuals from the 20th century’s “golden era” of science fiction publishing.

In particular, one magazine from that time – Astounding Stories of Super-Science – has been a formative influence on our project to meme the genre of protocol fiction into existence. Astounding set itself apart by moving away from the garish, action-oriented tradition of pulp illustration – heroes punching aliens in brightly-colored, sensationalized imagery – and toward a more subtly conceptual approach, grounded in believable technological futures. Editor John W. Campbell considered art to be an integral part of the magazine’s proposition to readers. He took inspiration from the suggestive potential of images, sometimes even inviting authors to write stories in response to paintings that he had commissioned. Campbell was fully aware of the concerted world-building power of tightly integrated text and image.

Subscribe now

Through its initial art direction, Protocolized has paid explicit, admiring tribute to Astounding, positioning itself in a familiar tradition of speculative fiction. We have made this style our own by consistently de-emphasizing human protagonists in the composition of illustrations. We instead picture less likely protagonists, such as autonomous agents, cybernetic systems, and novel infrastructures – bringing the background into the foreground.

Astounding aesthetics now carry nostalgic connotations which were never part of their original cultural appeal. As compelling as they still are, they have become cliche.

When it came to imagining a contemporary protocol fiction aesthetic, the approach had to be: do as Astounding did in embracing the art of its time, not look as Astounding looked. Since a significant aspect of what sets Protocolized apart editorially is its embrace of LLM-assisted writing, it made sense to train our own image generation model.

In collaboration with TITLES, a platform for training and distributing image generation models by artists, we have created two new models, with the initial intention of using them to distinguish between our protocol fiction output and the non-fiction, Summer of Protocols program-related content that we publish. The TITLES studio interface allows us to sample both models in a single generation, so these categories can be as permeable as necessary.

Unlike the major image generation platforms directed at outputting polished-looking images for content creators – Midjourney, ChatGPT, etc. – TITLES centres artists by naming them, giving them creative control over their training art, paying royalties on the use of their models, and storing image metadata onchain for human-legible proof of provenance.

Both Protocolized models were trained on new images created specifically for the purpose.

We invited Darius Ou of hyperpress to produce art for the first model. Darius’s training art is a byproduct of his incredible work in 3D printing, exploiting its primary graphic devices: the slice and the cross-section. Darius created an intricate 3D tree structure which was then sliced in software to produce fragmentary 2D cross sections, which Darius calls Spaceland Trees, in reference to Edwin A. Abbott’s 1884 novella Flatland, whose two-dimensional protagonists are categorically unable to conceive of a three-dimensional world.

The training art for the second model, created by the author, combines soft, smooth ‘neumorphic’ digital interface elements with traditional mark-making techniques used by archaeologists to document prehistoric sites of cultural interest.

The approach with both models has been to focus training art on narrowly defined and highly abstract visual languages. This creates cohesion across outputs, albeit with the consequence that these models are not primed by their training art to resolve figurative prompts. When figurative imagery is preferred, we introduce reference images to coax the models to repurpose their formal elements, with intriguing results.

In issue 22 we published Ralph Witherell’s story Time to Die, set in Serenity Now Hospice, a 2093 palliative care facility. Patients are held in a state of superficial comfort by monitoring, medication, and controlled hallucination. The story’s affect is compassionate, with a suffocating undertone.

Our original art depicted the bird which features in the closing scene as a sudden reminder of the violence of mortality. A space vessel passes in the background – in Witherell’s scenario “personal space travel was as common as taking a bus.”

Were we to have illustrated this story with the author’s model, we might have conflated the environment of Serenity Now with an abstract metabolic system, or introduced an anatomical reference image to emphasise bodily form. Darker, scattered elements would lend a hint of menace.

The art for our new Bridge Atlas salon series was generated with the hyperpress model. These examples show its outputs modulating between scenic composition and something flatter and more topographic. We used the former direction for a feature image, and the latter for YouTube title cards and thumbnails.

Other Internet’s 2019 essay Headless Brands charted the emergence of brand identities that were not dependent on the taste and discipline of the organizations which they were representing. This nascent concept was exemplified by decentralized, blockchain-native organizations – with Bitcoin as the canonical case study – but it has only become more culturally salient and technologically practicable in relation to developments in AI image generation.

Headlessness also resonates with a distinction made by John Fiske in their 1992 essay The Cultural Economy of Fandom, between “official” cultures established by a given brand or IP, and “fan” cultures which derive from and appropriate them. For Fiske, the difference between the products of these cultures is not necessarily qualitative:

Fans produce and circulate among themselves texts which are often crafted with production values as high as any in the official culture. The key differences between the two are economic … fans do not write or produce their texts for money; indeed, their productivity typically costs them money.

In the present moment, in which the production of media in the style of other media is a primary affordance of LLM technology, a publication such as Protocolized can easily collapse the distinctions theorized by Fiske and Other Internet by ceding the production and maintenance of its visual identity to its growing community of contributors.

Find our models, Protocolized × hyperpress and Protocolized × James Langdon, at TITLES.

To mark the release of these models, we have posted a new protocol fiction bounty in the Summer of Protocols Discord – write a story prompted only by the feature image at the top of this post. The selected story will be published as a paid contribution in Protocolized.

Listen to the full episode

Christine: Trent, can you start us off with an intro. What’s your background?

Trent: Sure. My background is in architecture and design, but I got pulled into Ethereum a few years into my career because I’ve always been fascinated by systems. I now work at the Ethereum Foundation on protocol coordination – that means managing changes to Ethereum’s core software, coordinating releases, and engaging stakeholders. One of my main projects is Protocol Guild, which is a collective funding endowment for about 200 core contributors working on Ethereum’s foundational protocols.

Christine: Yancey, how about you?

Yancey: I’m a writer and entrepreneur. I co-founded Kickstarter and led it as CEO, started The Creative Independent, and more recently, I’ve been working on Metalabel – a studio for group projects and tools, both onchain and offchain. I’m also involved in Artist Corporations, which is a project to create new corporate structures for creatives, and Dark Forest Operating System, which is about private internets for groups with shared space and resources.

Subscribe now

Christine: It sounds like both of you have taken unexpected paths into your current work. Yancey focuses on creative commons, Trent on Ethereum commons. Trent, can you define “Ethereum commons” for us?

Trent: I think of commons as systems for shared production and management of resources over time. Open-source software, like Ethereum, is a shared resource requiring governance and trade-offs. Ethereum fits into the “digital commons” category – a shared resource producing digital goods. The goal is to distribute value widely, but there’s always a risk of capture by capital interests, which can make these resources less “magical.”

Christine: Yancey, do you see parallels with the creative commons?

Yancey: Definitely. Creative commons rest on public goods like music theory, colors, and cultural building blocks. Standards like Creative Commons licensing are vital, because if standard-setting is left to commercial actors, you risk a race to the bottom. Managing commons well requires daily care – it’s easy to let things degrade, and hard to restore them once they do. Ethereum has an advantage over centralized Web2 platforms because its public goods are more resilient and not dependent on a single decision-maker.

Christine: But neither Ethereum nor creative commons sustain themselves automatically – they need tools and innovations. Yancey, you’ve focused on “labels” as a way to organize creative work. Why?

Yancey: Online culture is very individualistic – everyone builds their own following, which makes you compete with the people who are most like you. But independent record labels historically created alignment between artists: if one succeeded, others benefited. Labels amplify and invite others in, creating a shared credibility and a scene. That’s mostly disappeared online. With Metalabel, we’ve been building tools for group treasuries, revenue splits, and decision-making, so people can collaborate more effectively. It’s about creating alignment and shared identity.

Christine: Trent, does that resonate?

Trent: Yes. Labels have both social and very practical benefits – pooled payments, marketing, representation. In Protocol Guild, when we unite 200 contributors, we gain legibility. Funders can see the collective impact, which makes them more willing to commit resources, like through our “1% pledge” initiative. This kind of collective representation wouldn’t be possible for individual contributors alone.

Christine: Capital is a big part of sustaining commons. Trent, in a previous talk of yours, you didn’t take a firm stance on whether capital harms or helps commons. Has your view changed?

Trent: Not really – it depends on context. Capital from a multinational with strings attached is different from capital managed transparently by peers. You can’t avoid dealing with money because people need to live. The challenge is doing it in a way that preserves the commons’ integrity.

Christine: Yancey, your thoughts on the impact of capital?

Yancey: Money has huge influence, but the real driver in any project is whether it’s interesting enough for people to commit their time. Local capital can help; outside capital can distort incentives. Sometimes outside funding is the only option, but it comes with tension. Ultimately, the desire to participate is the most important currency.

Christine: That reminds me of Nadia Asparouhova’s book Working in Public, about the thankless nature of maintaining open-source projects. Money helps, but it’s complicated.

Trent: Exactly. Ethereum started as an ideological project, and intrinsic motivations still drive many contributors. Protocol Guild’s goal is to make funding a non-issue so people can focus on what brought them here.

Christine: Crypto tools like ICOs made it easy to bootstrap projects and directly connect with audiences. But Yancey, you have said that crypto isn’t fully mature for creatives. Still true?

Yancey: Yes. I think crypto’s future is as backend infrastructure – stablecoins integrated into mainstream finance. Creators will be consumers in that system, but things like “creator coins” are alienating to many artists. Financializing yourself is a big cultural leap that most creative people don’t want to make. The DeFi toolkit is exciting to crypto folks but off-putting to most artists.

Christine: That’s interesting – Protocol Guild wasn’t built by DeFi experts either. Trent, how do devs feel about being exposed to the speculative side of crypto?

Trent: Mixed feelings. Each wave of crypto innovation – ICOs, DAOs, NFTs, prediction markets – attracts speculation, and that’s what gets public attention. Core devs believe in the tech’s potential but often dislike how it’s used. Still, there are valuable kernels in these ideas that could mature into better models.

Christine: That leads us to the question: are Ethereum’s ends aligned with its means? How could it better align?

Yancey: Honestly, I’m not sure what Ethereum’s ends are. My guess is it’ll succeed as a new layer for internet money, making some people rich, but without replacing existing power structures. The revolutionary dreams might not materialize.

Trent: I’m more optimistic. The real unlock is governance – DAOs managing value among distributed parties. We’re seeing early legislative recognition of onchain organizations. That’s promising long-term.

Christine: Yancey, tell us more about Artist Corporations.

Yancey: They’re a new corporate form for creatives, blending LLC flexibility with cooperative and IP-ownership features. Groups could pool for healthcare, hold shared IP, and scale up or down. We’ve started a nonprofit to pass laws enabling them and to educate creatives on governance and finance.

Christine: I think Ethereum’s “ends” are still up for grabs. The protocol is malleable now, but it’s heading toward ossification. Different stakeholders, from corporates to creatives, can influence its trajectory – but the window is closing. There’s no single feature that will “fix” things for artists; it’s about keeping Ethereum accessible, cheap, and permissionless while scaling and preserving its unique values.

Trent: Yes – amidst all the speculation, valuable projects will emerge over time. The key is to maintain Ethereum’s core values against commercial pressures. We want it to be dependable infrastructure, like a natural commons, shaped by humans but resilient against turmoil.

Yancey: Agreed. Ethereum should be trustworthy, stable, and conservative in governance – focusing on making its core incredibly strong. That’s the only way to think long-term.

Trent: Some criticize Ethereum’s slow pace, but that’s part of ensuring it lasts. We see it as infrastructure for millions – maybe billions – and want it to operate consistently and dependably for generations, like ecological software maintained by people who care.

Christine: That’s a great note to end on – ambitious, but with clarity about what the Ethereum commons is and what it’s building toward.

Listen to the Full Episode

Christine:
Nils, can you tell us a bit about your background and introduce the idea of planetary thinking?

Nils: I was trained as a historian, with a PhD from Berkeley in American history. My interest in planetary thinking comes partly from studying modernization and globalization. After academia, I spent years in tech and consulting, including with intelligence agencies on long-range issues like China’s rise and climate change.

Planetary thinking is concerned with flows which are beyond human intentionality. Globalization covers intentional flows like goods, money, and ideas; planetary phenomena – like the carbon cycle, nitrogen cycle, space debris, and ocean plastics – are often human-influenced but unmanaged. We have governance systems for global flows but few for planetary challenges.

Christine: Josh, can you share your background and define “hardness”?

Josh: I started as a lawyer, then co‐founded ETHGlobal, which runs Ethereum hackathons, and later joined the Ethereum Foundation leadership. Hardness, as I define it, is a system’s capability to make a specific outcome very likely to be true in the future, usually to enable social coordination.

For example, I’m confident assets in my Ethereum wallet will be there in five years because of cryptographic guarantees and economic incentives. Similarly, money in a Canadian bank is backed by legal systems and deposit insurance. These are different sources of hardness, but they serve a similar function.

Nils: How is hardness different from legal reliance or reliability?

Josh: Reliance is about people’s mental states or trust in behavior. Hardness is a property of the system itself – it can exist whether or not people believe in it.

Christine: Given today’s erosion of trust in institutions, could Ethereum’s hardness also erode?

Nils: Trust in most U.S. institutions has declined for decades, except small businesses and the military, perhaps due to accountability. People turn to blockchain systems partly because they don’t rely on self‐dealing elites; their design is relatively incorruptible.

Josh: That’s core to crypto’s ethos – creating systems resistant to capture or corruption. Ethereum’s hardness is less likely to be undermined by cultural or social disputes because it’s a network, not a traditional institution.

Christine: How do you measure hardness versus trust?

Josh: Trust experience is like user experience – it’s about how people come to trust something. Hardness is about the system’s actual properties. For Ethereum, you could look at validator distribution, economic incentives, cryptographic security, etc. Like security, hardness is context-dependent.

Nils: Institutions like universities can remain hard – Harvard will likely exist for centuries – while losing trust. That’s why distinguishing hardness from trust is useful.

Christine: If trust erodes but hardness remains, does hardness matter?

Josh: It’s not enough for Ethereum to be secure; the world has to know and understand it. We need to bridge that gap so people see it as a fundamental coordination tool, not something alien.

Nils: From the outside, two big challenges for Ethereum are technical complexity and a public image dominated by financial speculation in cryptocurrency, which doesn’t foster long-term commitment.

Josh: I agree. The reputation can change, and security has improved – smart contract hacks are less frequent. But there’s still much to do on user safety and education. I hope Ethereum evolves like California – from Wild West to a vibrant, global hub.

Christine: Let’s pivot to planetary challenges. Nils, you’ve said diplomatic machinery can’t keep up with technological breakthroughs like AI.

Nils: Traditional diplomacy was about national governments negotiating. In the globalization era, diplomats needed economic and financial literacy. Now, planetary challenges require literacy in carbon accounting, space governance, and environmental monitoring. I’m working with foreign service schools on curricula for this.

We also need to understand the systems – potentially including Ethereum – that monitor and respond to planetary phenomena.

Subscribe now

Christine: Josh, where could Ethereum be relevant?

Josh: One example is the Freedom Tool, used in Russia, Iran, and Georgia. It uses zero-knowledge proofs so that citizens can prove eligibility to vote in online polls without revealing their identity. It’s been used by opposition activists to contest election legitimacy. I think the first billion-person vote could happen on Ethereum using such tools.

Nils: In planetary thinking, some focus on building new institutions, others on technological systems as de facto governance. For example, sensor arrays in shipping channels detect whales and automatically redirect ships, preventing collisions. Many planetary challenges will be addressed through automated systems rather than marble-column institutions.

Ethereum could enable more voices in decision-making and even multi-species integration – like monetizing Rwanda’s gorillas via Ethereum to fund protection and ecotourism.

Christine: Sometimes Ethereum feels like a small technical detail in grand planetary conversations. How is the Ethereum Foundation raising its profile?

Josh: We have a new Institutional Secretariat team to engage with governments and large organizations. For example, Bhutan is building its national ID system on Ethereum. There’s also “diplomacy” within blockchain – managing relationships between Ethereum and other chains, roll-ups, and L2s. As these proliferate, keeping the system interoperable will be critical.

Nils: Ethereum could also enable post-national diplomacy among cities, corporations, and other actors, bridging increasingly closed-off technological stacks in different regions.

Josh: I agree. In a fragmented world, Ethereum is one of the few trends moving toward more global integration. Both Chinese and American companies build on it, and it’s supported by a global community. That’s an underrated strength.

Christine: Timber, can you give us a brief overview of Summer of Protocols and its history?

Timber: Summer of Protocols (SoP) is a research program started three years ago to make protocols a first-class concept for thinking about the world. We’ve gone through four phases: a pilot study, a summer focused on research, a summer on application, and a summer on education. Along the way, we’ve added year-round tracks like a guest talk series and a newsletter that evolved into a science fiction magazine plus nonfiction essays. Funded by the Ethereum Foundation (EF), we’ve had about 100 participants produce essays, projects, and artifacts, all aimed at understanding, designing, and improving protocols.

Christine: Tim, you helped kickstart this at the EF?

Tim: Yes. In 2022, I reached out to Venkatesh Rao to help me think about Ethereum – it’s not just a product, tech platform, or political construct, but a mix of all three. We realized thinking of Ethereum as a protocol was the right approach. We did a pilot study defining protocols and how Ethereum fit in, then funded 30 researchers from different domains to study protocols. Timber, for example, looked at workplace safety protocols. We found common ideas across domains, so we kept expanding.

Christine: Many listeners think “protocol” means DeFi or blockchains. What do you mean by protocol?

Timber: We study everything from micro-protocols like handshakes to large-scale ones like the Kyoto Protocol, plus technical protocols like Ethereum and social protocols such as elections. The scope is wide – fields as different as coal mining, video game design, cryptocurrency, encryption, wildfire management, and urban planning.

Tim: Ethereum is like a Swiss Army knife – it shares features with other protocols but has unique elements, like embedded money. Studying different protocols helps us see what Ethereum might adopt or adapt, and we’ve found high-level principles that generalize across domains.

Going to DevconnectARG in Buenos Aires? Join the Bridge Atlas workshop on November 22nd.

Christine: Can you share key learnings from non-Ethereum studies that impacted Ethereum?

Tim: Two big ideas stand out. First, it’s better to analyze protocols by how bad they can be – like airport security – than how good they are. Nadia Asparouhova from our first cohort created the “Kafka Index” to rate the worst version of a protocol. Second, “Whitehead advances” – protocols that work so well they fade into the background, like HTTP or GPS.1 These need to be designed right from the start, as mass adoption makes change harder. Seeing these patterns refined our intuition and confirmed Ethereum’s community mostly has the right instincts.

Timber: And it goes the other way too. Coming from outside crypto, I’ve learned from Ethereum experts about building economies on top of protocols, which parallels work in health and safety in unexpected ways.

Christine: Timber, how do you measure the success of SoP?

Timber: We focus on building “protocol literacy,” so people in different fields can communicate more effectively about protocols. Over the years, we’ve simplified the language, recruited professors worldwide, and seen our terms used by people we’ve never met. That’s a good sign, even without hard data.

Christine: Let’s talk about this Bridge Atlas series. Tim, what inspired it?

Tim: In the early years of this effort, we avoided focusing on crypto in order to attract serious researchers from other fields. Now, it’s time to bridge Ethereum’s community with the broader protocol research community. This series pairs top Ethereum researchers with SoP alumni from other domains to explore parallel themes. Our first SoP event at an Ethereum conference will close DevConnect with a full Bridge Atlas day, connecting these communities.

Christine: So it’s about creating new “strata” for protocol research?

Tim: Yes, and showing how Ethereum-specific research can help other fields. Themes like “hardness” and “trust experience” resonate broadly. We want smart, curious audiences outside crypto to engage with Ethereum’s cutting-edge protocol work.

Timber: Think of it as building the missing “meso level” of discourse – between micro technical details and macro geopolitics – about how organizations and institutions design and manage protocols.

Tim: Exactly. Other professions have playbooks and best practices; protocols don’t. Creating this mid-level knowledge is what’s missing.

Christine: What themes will the series cover?

Timber: Computing frontiers – like wearables – and commons, from software to air quality, to road safety. How can we harden these against failure modes? Crypto folks have a lot to share here.

Tim: For Ethereum, that means questions like how to steward our commons, what adoption of new tech looks like, abstract protocol design, and the right institutional landscapes to support protocols.

Christine: As host, I’ll moderate between Ethereum and non-Ethereum guests. Timber, any advice?

Timber: Keep building your literacy with SoP terms, introduce them to guests, and let curiosity guide you. Many guests are SoP alumni and already have some shared language.

Tim: You’ve done great work exploring the human side of protocol builders’ motivations – that’s universal across domains.

Christine: I’m a skeptic, so I’m excited to test whether non-crypto protocol learnings really apply to Ethereum. If I’m convinced, listeners will be too!

Timber: That’s the goal.

Christine: Before we wrap, Timber, explain the name Bridge Atlas.

Timber: “Bridge” is about connecting fields, theory and practice, people. “Atlas” evokes multiple maps – multiple perspectives – and global scale. It also nods to the mythological Atlas, doing hard, thankless foundational work, which is what protocols do when they work well.

Subscribe now

Introduction

Unlike most previous technological disruptions, generative artificial intelligence (genAI) has an exceptionally low barrier to entry. Anyone with internet access and basic language skills can now generate content in the form of text, images, audio, and video. This accessibility explains both its extraordinary speed of adoption and its deep impact on the cultural and creative industries – sectors whose raw material is content but whose primary product is meaning.

From my experience researching and working on the impact of genAI in cultural and creative ecosystems over the past three years, two paths continually intersect.

• On one path, there are long-term questions around creativity, the datasets used to train models, copyright reform, and the future of creative labor.

• On the other, there are urgent implementation challenges: projects and organizations must adopt these tools to remain relevant, yet doing so ethically demands shared principles and clear agreements among those who use them.

This text stems from FUTURX’s panel Protocols for a Critical Use of AI in Cultural Projects, organized with UNESCO, on August 20, 2025.

Our aim is to stoke the debate around one of the most pressing challenges in the cultural field today: how to design frameworks and protocols that make it easier for people to do the right thing – to use genAI tools to enrich their art and culture without fear, by managing negative side effects such as opacity, bias, or data misuse.

An initial guiding question is:

What should a cultural project or organization do when faced with a high-impact technology like genAI? How – and for what purpose – should it be integrated?

To respond, we propose three starting points:

• Build a shared understanding of what critical use of genAI entails.

• Recognize the areas where genAI has the greatest impact on cultural projects and organizations.

• Embrace the notion of protocols as a fundamental tool to address common challenges posed by technological change.

Can we introduce protocols to cultural projects?

There is a long history of tensions between technology and culture. Technological paradigm shifts require processes of adaptation in which many established practices need to transform. In this sense, the capacity of genAI to produce text, image, audio, and video from pre-existing content datasets once again challenges creative and artistic practices.

Platforms like SUNO – currently the leading generative music platform, producing more than one million songs per day – claim in their article The Future of Music that they train their models with “all the music available on the internet,” just as any individual could. By contrast, many cultural voices argue that this practice is neither ethical nor transparent, calling for state regulation of the use of cultural content in training datasets.

These debates will not be resolved overnight. While legislation is still under discussion in much of the world, we can gain both time and sovereignty if we focus on how these models could adapt to our practices – instead of adapting our practices to the models.

From here, several important questions emerge:

• Where do we draw ethical boundaries in creation, reception, analysis, and communication when using content generated wholly or partially with genAI?

• How do we protect culture from being included in training datasets without consent? And what possibilities emerge if we stop thinking of genAI solely as a tool and begin to approach it as a cultural, political, and economic system?

• What kinds of collective agreements could we imagine so that we can decide – together – how to use the technology?

Protocols appear as one possible answer: engineered arguments that enable cultural projects to structure the relationship between content and technology. The cultural sector needs its own frameworks – dynamic and adaptable – capable of sustaining trust among peers and communities. Protocols can serve as a shared compass to decide which data we use, how we label it, how we guarantee transparency, and which ethical boundaries we do not want to cross.

We are not starting from scratch. UNESCO has already established key references: the 2005 Convention frames culture as a crucial dimension of sustainable development, and the 2021 Recommendation on the Ethics of Artificial Intelligence offers globally accepted values, principles, and policy guidelines.

At FUTURX we have begun translating these ideas into practice through research and collective experimentation, developing frameworks and tools to help cultural organizations navigate the ethical and creative challenges of AI.

What does it mean to use AI well?

As I mentioned at the beginning, what makes genAI truly disruptive is its low barrier to entry. This accessibility democratizes creation, but also multiplies ethical, aesthetic, and social challenges that we are only beginning to grasp.

The problem is that, in most cases, these tools do not guarantee many of the basic requirements a new piece of content should meet: knowing the source and reliability of the information it contains, ensuring transparency, and making sure it is ethical and appropriate for the intended use.

We consider the critical use of this technology to be conscious, reflective, and ethical. This mainly implies:

• Analyzing information or content generated by AI, since it may be biased or incorrect.

• Maintaining an ethical stance, protecting authorship, cultural diversity, and property.

• Preserving our autonomy and vision; using genAI as an ally.

In short, within culture and creativity, critical use means that AI should not dictate the outcome, but rather expand our possibilities – always under our gaze and responsibility.

Creation, Experimentation, and Trust

How does a philosophy of critical use translate into creative and cultural practice? What protocols can we extract from real experiences to guide our practices?

Based on our research and collaborations with different cultural institutions, we have identified a few key considerations:

• The distinction between experimental and commercial uses (where clear clauses and traceability are necessary).

• The need to align models with values of openness and rights.

• The urgency of working with diverse data and metadata to avoid bias.

• The value of critical prototyping as a methodology that accepts error and allows us to open the black box of algorithms.

These points show that protocols are not abstract rules, but situated guides that can help orient how we experiment, produce, and distribute culture in an environment dominated by opaque systems.

Subscribe now

Areas of Impact: Imagining Protocols for Using AI

After understanding the tensions and needs that arise when aiming for a critical use of genAI, the next step is to look closely at the areas where this technology has the greatest impact on cultural projects and organizations.

Today, trust between peers – whether individuals or projects, within the same organization or across different ones – is facing a crisis due to the lack of transparency, traceability, and the inappropriate use of platforms.

We have identified five key points of interaction between genAI and cultural projects. For each one, we outline a guiding question and suggest an initial protocol.

1. Reception of Content

Guiding question: How can we create mechanisms of trust when works or proposals produced using genAI arrive at an organization or project, where simple prohibition of these tools is not the solution?

Protocol suggestion: The key is knowing how these tools were involved. A simple step is to ask participants to declare it. This transparency opens up the process without diminishing authorship.

Example: A public fund that includes in its application form the question: “Did you use genAI in this process or for the idea you submitted? Please describe how.”

2. Creation of New Content

Guiding question: When organizations produce with AI, how can we ensure accountability?

Protocol suggestion: What matters is leaving a trace: which prompts were used, what decisions were made, how humans and models were combined. This traceability helps explain where a piece comes from and recognize different contributions.

Example: A museum that publishes a visual piece created with AI and specifies on the exhibit label which model was used and the role of the curators.

3. Analysis of Archives and Databases

Guiding question: How do we avoid erasing memories or exposing sensitive data when using genAI to analyze collections?

Protocol suggestion: genAI can organize archives, but it can also obscure histories or leak private information. That’s why it is crucial to review the outputs, safeguard privacy, and ensure proper context before dissemination.

Example: A film archive that digitizes its collection with genAI, but validates each description with specialists before publishing it.

4. Communication with Audiences and Peers

Guiding question: How do we preserve trust when interacting with audiences using genAI tools?

Protocol suggestion: Automated responses can improve efficiency, but relationships with audiences require transparency. It should always be clear when a machine is speaking and when a person is, and a human channel must always remain available.

Example: A record label that uses genAI to draft email responses but has a team review and approve them before sending.

5. Protection and Cultural Diversity

Guiding question: How can we prevent genAI from reproducing biases and erasing singularities?

Protocol suggestion: If left unchecked, models tend to replicate biases and flatten diversity. Incorporating situated datasets and specific metadata is not a technical detail – it is a way of expanding cultural diversity instead of reducing it.

Example: A community archive that becomes a dataset to train a model highlighting local languages or cultural expressions.

Foundations for Culturally Responsible AI Projects

Can we imagine minimal, grounded applications of these protocols?

A culturally responsible use of genAI begins with a techno-realist mindset – one that acknowledges the non-neutrality of technology and reclaims the right to shape it according to our own contexts, values, and imaginaries. The cultural field has the power to design its own tools, languages, and agreements instead of adapting passively to corporate ones.

These protocols will not only help us work better; they will help us build the new codes of trust we need to use this technology ethically and collaboratively. What we need are not grand declarations, but small, situated actions—shared practices that make ethical use tangible in everyday life. This is an open invitation: for artists, institutions, and communities to act as system designers, building a culture with genAI that is responsible, sustainable, and radically human.

Edited by Valentina Cuneo.

The Khyber Pass is just 15 feet wide at its narrowest point, and around 33 miles long. Yet, more planet-shaping history has flowed through this short and narrow geographic bottleneck than across the entire rest of the Himalayas and adjacent ranges combined, which form a largely impassable natural border between South Asia and the rest of Asia. A border more formidable than the Pacific Ocean.

The pass is a narrow bridge not just between two geographies, but between two streams of world history. Over millennia it has witnessed and shaped the rise and fall of empires on both sides. Conquering armies and vast rivers of trade and commerce have flowed through it. Traders, diplomats, explorers, scientists, and religious evangelists have kept up a steady flow of ideas, wealth, and power flowing in both directions. Without the Khyber Pass, South Asia would have perhaps been as isolated from world history as Australia, and the history of the world would have been radically different.

Yet, despite its obvious importance, the pass is a small patch of land that is remarkably hard to find on a map. It does not matter whether the map is political or geographic, or what clever projection you use. It is just very hard for a traditional map to represent the Khyber Pass in a way proportionate to its significance. A significance rooted in the dynamic flows of our planet rather than its stocks: The static topography of containment and boundaries.

The problem is not unique to the Khyber Pass. Despite the obvious (and as I will argue, growing) value and importance of viewing the world from vantage points like the Khyber Pass, we humans are remarkably bad at developing worldviews anchored in flow-based vantage points located on bridge-like elements on maps.

In fact, we do not even normally think of a bridge as a vantage point at all. A bridge is a space you typically cross quickly rather than linger on, even when the view is spectacular, as it often is. Bridges, understood as metonyms for a larger class of geographic connection elements, which includes straits, passes, tunnels, and administrative border crossings, are liminal spaces. They connect two realities with sufficient kinship to justify a connection, but divided by some sort of natural or artificial barrier. To cross a bridge is to undergo a mental shift. To linger on a bridge is to be of two minds for a spell. It is an uncomfortable state to inhabit. Too many dimensions, many in tension with one another, are alive. Too few have been flattened to allow us to sustain the comforting identities we can securely inhabit on either side of a bridge. The view from a bridge may be spectacular, but being on a bridge is not comfortable unless you’ve trained yourself in appropriate ways.

Bridges demand that we inhabit more powerful, higher dimensional selves at least while we’re on them. Selves that can, at a prosaic level, at least navigate multiple languages, currencies, measurement systems, and political narratives. But while a certain aptitude for pluralist, polyglot, even mongrel states of being is necessary to constitute such selves, it is not sufficient.

We need something more – what we at the Summer of Protocols have been calling protocol literacy: A more holistic and native level of comfort in and around bridge-like spaces.

Bridge-based identities, like Silk Road trader or diplomat, demand a praxis of transcendence of narrower identities even to sustain basic existence. Historically, very few humans had the opportunity to acquire such identities – sailors, diplomats, traders, soldiers.

In our postmodern world, increasingly all humans must develop bridge-rooted identities, and we are all struggling with the challenge. The Covid pandemic showed us just how bad we are at this. Our stocks-based minds struggled to wrap themselves around supply chains and flows. Our ineptitude is only getting costlier, as new flow-based confusions and conflicts emerge at all levels, from rare earths and semiconductor supply chains, to climate refugee flow management, to AI turning familiar stocks of inherited civilizational knowledge into oozy flows of live intelligence.

So perhaps the first thing we need, to allow us to acquire the higher levels of protocol literacy the world demands of us today, is better maps of the world. Maps that privilege flows over stocks, bridges over borders, and in-betweenness over stable perspectives and subjectivities. The traditional school atlas you may have grown up with is no longer enough. We need a new kind of atlas.

In this essay, I want to introduce a new frame that we at Protocolized (and the parent Summer of Protocols program) have become enamored of in recent months – that of the bridge atlas. A set of perspectives of a world centered on its bridges – connection points between relatively homogeneous regions separated by borders, and marked by powerful, transformative flows.

Our fall programming revolves around the motif of bridges, and the larger scaffolding of bridge atlases. Three highlights:

1. We are announcing our third protocol fiction contest today, Building and Burning Bridges (check out details here; deadline for entries is December 1, 2025).

2. We will shortly kick off a special series of Bridge Atlas salon conversations, hosted by Christine Kim, which will surface “bridge perspectives” of computing futures, viewed from paired crypto and non-crypto vantage points.

3. For those of you who plan to attend Devconnect (the flagship Ethereum event in Buenos Aires, November 16–22), we have a day-long event, also titled Bridge Atlas. The goal of this event will be to map and build bridges between the future as imagined by the Ethereum community, and by others imagining their own versions of futures. If you are planning on going to Devconnect, and consider yourself a self-appointed Ethereum ambassador, building and traversing bridges to other worlds, this event is for you. Register here.

In case it isn’t obvious, we at the Summer of Protocols are interested in bridges because protocols of all sorts – diplomatic, technological, climate, blockchain, AI – are bridge-building and flow-shaping technologies. If physical bridges, passes and straits are hard to see on maps of the planet, the larger universe of protocols is almost entirely invisible. The technological equivalent of dark matter and energy.

The cultures that emerge on top of any physical or conceptual geography of bridges are fundamentally rooted in bridge-based perspectives. The people who successfully inhabit those cultures tend to be precisely the ones who are able to develop the complex, higher-dimensional identities it takes to view, navigate, and act on the world from bridge-like loci.

In what follows, I want to lay some philosophical groundwork around why the views from bridges are so valuable, especially if you are able to view a reality from multiple bridges, and organize your larger understanding of our world in terms of bridge atlas worldviews. We hope this introductory essay and our Fall programming will get you bridge-pilled (a sub-condition of being protocol-pilled – our entire program is a literacy program based on pushing a bunch of pills).

Subscribe now

What is an Atlas Anyway?

A traditional atlas is already a valuable construct. Instead of offering a single map with a single center, with a view of a single reality, it offers many views of many related realities, and does so in a polycentric way.

A single map must necessarily feature the distortions of a single projection. An atlas on the other hand, can attempt to be aware of the distortions of single projection techniques, and combine many in order to try and create balance. And unlike a globe, which lulls us into a false and somewhat useless sense of three-dimensional spectatorial integrity, an atlas leans into the inevitable fragmentation we must navigate in dealing with high-dimensional realities with low-dimensional sensoria (and our planet certainly features many dimensions both globes and 2d maps flatten).

An atlas never lets us forget that all perspectives are partial, the result of tradeoffs among different aspects we might want to represent, and inevitably feature distortions. And that it cannot be otherwise.

A bridge atlas kicks the ambition level up a notch.

Crack open a modern atlas. You’ll find a set of maps that cover areas at different scales. You’ll find multiple maps offering different perspectives on the same underlying reality, such as physical vs. political maps. You’ll even find a variety of projection methods in use, that attempt to account for different distortions and prioritize different aspects of 3d geometry in 2d representations – areas, shapes, distances, oceans vs. landmasses, polar vs. equatorial regions.

What you won’t find, though, are representations that center connection over separation. Bridges over borders. At best, you’ll find connectivity data (such as air or sea routes) layered on top of primary representations that prioritize regions that serve as conceptual centers, and the borders separating them. The logic is that of a mutually exclusive and collectively exhaustive (MECE) set of containers that carve up a world rather than a network that scaffolds its liveness.

This is not a trivial point. The more dimensions you want to capture in a representation of a reality, the more valuable connection points become, as perspective anchors. Though computer maps present the familiar regions-and-borders view of paper maps, mapping software relies on underlying connectivity-graph models, which is what you need for actually planning routes.

The problem is not limited to planetary geography. Take, for instance, a rarefied problem in the world of blockchains that is growing increasingly urgent – bridging between cryptoeconomic and traditional economic worlds. It is a problem that offers no intuitive visual interface for navigating it.

For example, consider financial flows, perhaps the most important kind of global flow that are typically not represented at all in any atlas. Even ancient ones like the hawala network, or hugely consequential ones like the flow of Spanish gold across the Atlantic, cannot be found in modern atlases. Nor can more familiar modern ones such as the SWIFT system. At best, you might occasionally see a graphic with foreign direct investment (FDI) levels overlaid on a world map in a financial newspaper. It is remarkable that such an incredibly important dimension of our planet’s conceptual geography is largely missing from our maps and atlases. Leaving rivers of money off our maps is almost as egregious an omission on a modern map as leaving off actual rivers.

Consider the problem of including a financial flows world map, encompassing both traditional and crypto instruments, in a bridge atlas. Imagine showing how capital is sloshing back and forth across stocks, bonds, gold and crypto, and across borders. Imagining a map that accurately depicts North Korea’s growing treasury of crypto assets, and the geopolitical and international-security significance of that treasury.

Zooming in, the world of DeFi (Decentralized Finance) has radically different (and faster) topologies and temporalities of stocks and flows than the world of TradFi (Traditional Finance). Bridges between the two worlds are fraught places, and often the site of scams and grifts, large and small. It is worth noting that the largest crypto scams have occurred not in the interiors of any cryptoeconomic ecosystem proper, but at the exchanges where crypto and traditional instruments can be traded for each other. And these bridge loci are getting increasingly heavily trafficked. Just one indicator: The annual volume of stablecoin transactions was $27 trillion in 2024, more than Visa and Mastercard combined. Countries around the world are scrambling to develop central bank digital currencies (CBDCs).

How would you depict all this on a map? How would you show how these flows interact with traditional financial and economic flows? How would you show how these flows interact with the broader global political economy, with the ebb and flow of state capacity? How would you show these flows on maps of gray or dark economic activity, including capital flight patterns, smuggling and crime, economic activity in failed states, and funding of terrorism? What sort of map could inform enlightened policy discussions on regulating vs. enabling cryptoeconomics that are able to rise above the current norm of protocol-illiterate flame wars between pro-crypto and anti-crypto tribes?

It’s a daunting cartographic challenge, but one thing seems clear – we would need maps that center bridges and flows, not regions and borders.

And this is just one cartographic problem that an inventive bridge atlas might solve. Imagine applying similar reasoning to climate protocol flows of emissions and credits, distributed AI “intelligence” flows, and of course, all the traditional flows that already keep leaders up at night – electricity, water, food, internet connectivity, pollution, semiconductor supply chain flows, rare earth flows, oil and gas flows. The list of invisible flow layers of the planet just keeps growing. And our cartography is not keeping up.

Simple distortions and transformations of traditional maps are not enough. We need new cartographic foundations not just to understand our world, but to even see it in the first place.

The modern atlas, first developed and named by the visionary cartographer Gerardus Mercator (much-maligned in our time, and rather unfairly so) in 1595, is desperately in need of an update. We need to get beyond the Mercator mindset.

Beyond the Mercator Mindset

To an engineer studying the abstract geometry and topology of an object of interest, the distinction between borders and bridges is often an academic matter. For instance, two schemes commonly used in robotic navigation to study and scaffold the structure of a space – Delauney triangulation (“bridges”) and Voronoi diagrams (“borders”) – are mathematically duals of one another. Which one you use is a matter of taste, convenience, and computational efficiency. Pure mathematicians often loftily discard the distinction altogether, working with group-theoretic topological representations that transcend lowly geometric understandings of connection and separation.

In the real world of course, things are nowhere near as symmetric or abstractable. We think natively in terms of borders and how they separate regions, which we stably inhabit. Thinking in terms of bridges and how they connect things, and how we transiently inhabit them, is uninituitive and awkward at best.

Bridges, unlike borders, don’t usually draw our attention unless we are in their immediate vicinity. In smaller-scale maps, such as world maps, important bridges, such as between the European and Asian sides of Istanbul, are not even marked. Narrow straits might be reduced to a pixel or two. Exclave corridors, common in convoluted and contested political geographies around needs like ocean access, are often tiny and easily missed. Borders, by contrast, not only straggle dramatically across maps at any scale, but mark sharp visual transitions.

Raised as we are with traditional maps on screens and on paper, our worldviews and orientations are shaped by borders in subtle but powerful ways. And this is true even of our non-geographic worldviews and orientations. For instance, our mental “map of the internet” and its underlying technologies is anchored in continent-like “platforms” and “stacks.” Artists attempting to depict the social or technical geography of the internet visually reach for metaphoric visualizations that owe more to Mercator’s 16th century ideas than to more suitable modern scaffoldings like graph theory or Sankey diagrams (something I myself have been guilty of). Randall Munroe, of xkcd fame, is among the few conceptual cartographers to even attempt to break out of ingrained cartographic mindsets descended from Mercator, and think beyond borders and interiors. But even Munroe’s most imaginative map (below, he has published several), fails to entirely escape the Mercator mindset, scaffolding the view with IP addresses rather than the communication links among them. The map below is still one based on stocks rather than flows.

To return to our opening example, if we wanted to indicate historical importance of the Khyber Pass on a map, perhaps by sizing segments of the border in proportion to the volume of traffic across them, the few passes through the Himalayas would need to be stretched out radically, and the impassable mountainous stretches of the border would need to be radically shrunk in proportion. Pakistan and Afghanistan would be revealed as being connected by a few graph edges between cities, corresponding to important passes like the Khyber, across the Hindu Kush and Karakoram ranges. Not by the 1640-mile long border (the Durand Line).

This is not a trivial point – currently inflamed political and ideological discourses around “open borders” rarely dwell on the question of how open or closed borders actually are, or whether bridges might serve as better foci for discourse hygiene. Political scientists are not entirely blind to this phenomenology of course. Many observers have noted that squiggly borders on maps tend to be default-closed (corresponding to contours of natural barriers), while more legible straight borders (corresponding to top-down borders drawn on the basis of administrative priorities and convenience) tend to be default open. Such insights mark the beginnings of cartographic protocol literacy and fluid perspective switching between borders and bridges, flows and stocks, and open and closed interfaces.

Bridge atlases though, should be understood as a more expressive approach to all of cartography, not just the ontological tension between bridges and borders.

Other dimensions of cartographic perspective suffer in similar ways from the Mercator mindset, and might be liberated by bridge-based cartography.

For instance, naive US political maps show an almost entirely red country, raising the puzzling question of why recent elections seem so evenly balanced between red and blue. Maps based on population density, made with diffusion cartogram techniques reveal a truer picture and resolve the puzzle.

Cartograms based on other variables can flexibly present many rich perspectives of a space. Despite the inventiveness of such maps, however, they remain inadequate. The map above, for instance, does not capture a crucially important dimension of the US political landscape – mobility.

Population and commodity flows between and within states, and between urban and rural population centers, have repeatedly shaped and reshaped the history of the United States through its entire existence, yet are largely invisible on maps. Historic flow episodes, such as the settlement of the West, and the great migration to urban centers, show up in histories, movies and even video games (such as the Oregon Trail), but not on the most important orientation artifacts for us spatially dominated primates: maps.

A bridge atlas can begin to address these inadequacies in our orientations. We can begin to see and think natively in terms of connections and directed flows between regions, and the historical temporalities they induce.

We do not yet know how to construct such atlases, but we can imagine techniques such as Sankey diagrams, representing flow volumes between sinks and sources, being adapted to the problem. But for the moment, viewing the world from a coherent set of bridge perspectives remains incredibly hard to do.

Bridges as Transformers

So far we’ve considered bridges primarily as situated connectors of adjacent separated realities – the sinks and sources of entangled histories. There is another important aspect to consider – bridges not just as conduits, but as transformers. This is an aspect that rests critically on the fact that bridges, both artificial and natural, are typically more complex and technologically and culturally active, than impassable borders.

Both natural and artificial borders can be dramatic, like the Himalayas or the Great Wall of China. But it is the bridge-like connective elements across borders that embody much of the world’s structural and behavioral complexity, and what we might think of as the information and energy processing aspects of our planet.

Our planet, you might say, thinks with its bridges. They are the synapses of the planetary mind.

A bridge as a transformer operates on the flows that pass through or over it, modifying its informational or energetic character. For instance, when you cross a land bridge, you might enter a regime with different distance units, currency, and language on signage. You might need to switch which side of the road you drive on. When a boat goes through a canal lock, its potential energy is raised or lowered. When electricity flows “across” a transformer, voltage and current levels change. When a token is “bridged” from an Ethereum Layer 2 network to the Layer 1 network, different traffic rules begin to govern it, and the cost to move it around rises.

Engineers often use an overloaded term borrowed from electrical engineering – impedance matching, for the problem of accommodating such differences across things being bridged. As a simple example, the traffic capacities of the roads leading to and away from a bridge must be roughly matched for the bridge to function well as a flow.

Protocol bridges, like their geographic analogues, are fertile sites of both conflict and cooperation, due to the kinds of transformations they effect. For instance, a “bridge” between the ActivityPub and ATProto protocols, meant to connect the Mastodon-based fediverse and Bluesky social media ecosystems, has been the site of intense cultural conflicts in recent years, due to the differences between the norms on the two sides.

A stranger example – the docking mechanism meant to bridge between Apollo and Soyuz modules during the Cold-War-era Apollo Soyuz Test Project ended up being unnecessarily complex because neither side wanted their side of the docking mechanism to be the “female” side of the connection (a natural but symbolically loaded solution to any dynamic connection-engineering problem).

This transformational character of bridges is in fact central to their value as vantage points. On a bridge, normally quiescent variables might be activated, and competing forces which might exclude each other in the interiors of the adjacent regions might be simultaneously active. Consider for instance a border region where currencies of both neighboring countries are accepted by merchants.

A computing metaphor is useful here: bridges are dual boot spaces, requiring hypervisory sensibilities to traverse. At a more cultural level, code switching of some sort may be necessary.

The dynamic vantage point created by the transforming nature of a bridge can even be central to its purpose. In electrical engineering for instance, a Wheatstone bridge is a circuit that measures an unknown resistance by balancing two sides of a circuit. The Rosetta stone bridged ideas across multiple languages, both in its own historical context as a living text, and in our time as an archaeological lens on ancient cultures and dead languages.

Bridges, understood not just as flow conduits between adjacent realities, are dynamic and active conduits that transform the flows they touch (and thereby, the realities they connect). They do not just passively (if richly) observe reality from liminal perspectives, they shape reality from those perspectives, by employing the unique affordances of bridges.

Bridges are, to use a currently popular term, agentic building blocks of our world.

A useful metaphor for remembering this fact is a rather unusual class of bridges: the kind you find on ships.

The Agency of Bridges

The “bridge” of a ship is called that because in the early days of steam, a small raised walkway connected the port and starboard paddle wheel housings, to allow the crew to cross from one side to the other without descending to the main deck.

The bridge turned out to be an ideal vantage point from which to navigate, steer, and direct ship operations. Though modern ships no longer feature literal bridges, the term bridge has stuck, and become synonymous with leadership, multiple perspectives, and fluid and intelligent structures of command and agency.

Productivity guru David Allen, in his book Making It All Work, offers a useful 2×2 that we can adapt to understand the particularly potent blend of perspective and control that bridges embody:

From the point of view of the Summer of Protocols, which to a first approximation is a protocol literacy program, the challenge of building bridge atlases of the world, encompassing not just spatio-temporal planetary realities, but intangible technological ones that are hard to situate in space and time, is one of the most critical projects in epistemology and ontology. Without bridge atlases, we cannot hope to articulate and utilize the growing protocol agencies becoming available to us with any sort of coherence. And given the criticality and consequentiality of the challenges we face, flailing around wildly is not an option.

It is, of course, a project that is far too big for our little program to take on by itself, but we hope to do a few things to help trigger this much needed cartographic revolution – including building bridges to other programs and institutions that recognize the importance of such efforts.

Our bridge-themed Fall programming – Building and Burning Bridges fiction contest, Bridge Atlas Salons series, Devconnect Bridge Atlas day – is meant to get this effort off the ground. We invite you to not just participate in these activities as you’re able, but to imagine and develop your own. We will aim to support such efforts as much as we are able to, and the idea of bridge atlases will frame our 2026 program.