Philosophy on Arem

Transversal Watchfulness

Thu, 23 Apr 2026 00:00:00 +0000

A crisis reveals what an economy is willing to pay for—and what it leaves out of the frame. DeFi’s April 2026 sequence—the USR depeg and the Kelp exploit—brought into view a structural asymmetry that few analyses have taken the time to name.

Days before Kelp DAO, a depeg hit the stablecoin USR. Users who were alerted early enough got out. That early signal came from Pharos Watch, a stablecoin observatory launched two months earlier by an independent developer, self-funded, who recently opened a public donation page simply to keep the project running.

That observation raises a question Kelp only made sharper: who pays for DeFi to keep making sense of itself?

What protocols fund—and what they don’t

The picture is more nuanced than it first appears.

Over the past two or three years, DeFi has learned to fund part of its own security layer. LlamaRisk has been an Aave service provider since 2024, with an annual budget of around one million dollars in AAVE, vested against KPIs. Chaos Labs operates under a similar model. These are professional risk providers, paid by the very DAOs they analyze.

Since February 2026, the Ethereum Foundation has structured a partnership with the Security Alliance (SEAL) to fund a security engineer dedicated to combating drainers. The ETH Rangers program, which closed in April 2026, funded six months of stipends for seventeen security researchers working on public goods.

Serious startups exist as well. Blockaid, which published Kelp’s technical post-mortem and an open-source DVN audit script, raised 83 million dollars from Ribbit Capital, Google Ventures, Sequoia, and Greylock. Its clients include Coinbase, MetaMask, and Uniswap.

Some core contributors, banteg among them—whose reading of LayerZero’s deployment code became a reference during the crisis—draw salaries from major protocols.

So the ecosystem has learned to pay some of its watchers. But it pays them conditionally—and that condition draws a sharp line.

Servile watchfulness, transversal watchfulness

Every paid structure mentioned above shares one defining property: it serves an identified payer.

LlamaRisk looks at Aave because Aave is paying. Chaos Labs follows the same logic. Blockaid protects Coinbase and MetaMask under commercial contracts. SEAL operates with targeted sponsorship from the Ethereum Foundation. The work these actors do is serious and valuable. That does not make them independent observers. Their mandate is defined by whoever signs the check.

Alongside this servile watchfulness runs another function, rarer and more exposed: transversal watchfulness. The kind that observes several competing protocols without belonging to any of them. That maps the system as a whole rather than acting as a vendor to a single actor. That can call out a problem affecting a potential payer without jeopardizing its own economic survival.

The distinction is easy to state: servile watchfulness serves an identified payer. Transversal watchfulness observes the whole without belonging to any. The first is funded. The second is not.

Two structures carry this function almost alone in 2026.

Pharos Watch, mentioned above, tracks 205 stablecoins across every major chain. Its founder, TokenBrice, recently stated publicly that he funds the tool out of pocket and is aiming to transition to community funding via Giveth by the end of 2026.

DeFiScan, maintained by the DeFi Collective, produces an independent evaluation of DeFi protocol decentralization based on a formalized framework. It operates as a non-profit, funded through Giveth and open contributions.

To these, add a handful of researcher-practitioners who publish analysis alongside their primary engagements, and a scattered layer of independent editorial observers.

This transversal layer is structurally fragile. It rests on the personal persistence of a few founders. By definition, it cannot sell services to what it observes without becoming servile itself. And the ecosystem has yet to build mechanisms capable of funding it without capturing it.

What Weick called sensemaking

In the 1970s and 1980s, organizational sociologist Karl Weick studied collective disasters—the Mann Gulch fire of 1949, the Bhopal catastrophe of 1984—to understand why groups facing ambiguous situations lose their ability to act coherently. His answer fits in one word: sensemaking. Organizational survival under stress depends less on computational power than on the capacity to build a plausible shared understanding of what is happening.

Weick showed that this function is not automatic. It requires structures capable of producing shared meaning independently of the particular interests of the actors involved. These are the structures that make collective sensemaking possible.

Pharos, DeFiScan, SEAL, banteg, Tay, ZachXBT, LlamaRisk, editorial observers—each contributes to that function. Their coexistence, their ability to read one another, and the fact that their outputs are publicly verifiable make convergence toward a plausible shared account possible.

When these structures are missing, collectives move through crises blindly—not because information is absent, but because no one is responsible for making it collectively intelligible.

DeFi is, by design, a high-ambiguity environment: opaque composability, hidden dependencies, a proliferation of actors, no central authority. It requires sensemaking more than most technical ecosystems. And yet the economic structure it has built rewards the production of complexity handsomely, and the production of shared understanding barely at all.

The tension has been articulated recently by others. Gitcoin, which has carried the question of public goods funding in the Ethereum ecosystem for years, integrated sensemaking into its grants architecture starting in 2025. Their formulation, in recent research, goes further: when AI extends mediation from what we see to how we reason, the risk becomes existential. They pose the question at a civilizational scale. It arises at DeFi scale with the same force—only closer to home.

What both traditions converge on, each in its own way, is the same reality: a collective capacity to understand does not emerge on its own. It requires dedicated actors, maintained capabilities, stable funding—and, crucially, structural independence from the interests being observed. Without these, there is no shared understanding. There are only competing narratives.

Once the frame is set, the problem facing post-Kelp DeFi comes into focus. It is not that DeFi lacks watchfulness. It has developed plenty. It is that no mechanism exists to fund its transversal layer.

When Aave pays LlamaRisk a million dollars a year, it pays for risk analysis on Aave. That is legitimate and useful. But no one, in that equation, is paying LlamaRisk to publicly say something that might embarrass Aave when the moment calls for it. No one is paying for an independent structure to observe Aave and its competitors with the same lens. No one is paying to map dependencies that cut across protocols without belonging to any of them.

The standard objection to publicly funding this kind of function is that it would create political dependency. It would not. L2Beat, for years, has served as a reference for evaluating Ethereum rollups without being captured by any of them—precisely because it built structural independence: diversified funding, transparent methodology, and a stable team. That model exists. It is simply not generalized.

The standard libertarian objection—users who need this information should pay for it—runs into a well-known economic constraint: information goods behave like public goods (non-rivalrous, partially non-excludable), which prevents markets from producing them efficiently on their own. No one pays for watchfulness while it exists. Everyone regrets it once it disappears.

It must also be acknowledged that the ecosystem can coordinate when incentives align. Fluid’s response—together with Lido, ether.fi, 1inch, and KyberNetwork—to unblock aWETH positions stuck on Aave post-Kelp, processing more than 400 million dollars within hours, is a recent example. That operational coordination is real and valuable. It does not solve the problem raised here. Fluid coordinated with Lido because their users overlapped and their commercial interests converged. No one, in that configuration, was tasked with holding a perspective that might have challenged any of the parties. Coordination under aligned incentives works. Watchfulness without an economic mandate remains orphaned.

What this concretely asks for

A sustainable DeFi—one that lasts a decade rather than a cycle—will have to consolidate mechanisms that are only just beginning to emerge.

For mature DAOs. Aave, Uniswap, Lido, Sky sit on treasuries in the hundreds of millions of dollars. Setting aside a fixed fraction—well below one percent—as an endowment for public intelligibility infrastructure, with no service counterpart and no contractual capture, would be a technically trivial move. It has not yet been seriously put on the table.

For the commercial actors who depend on them. The emerging intelligence layers selling DeFi attention—dashboards, institutional analytics, paid risk platforms—all rely on public sources (DeFiLlama, DeFiScan, Pharos, independent analysis blogs). In other ecosystems, it is standard practice to reinvest a fraction of revenue into upstream public goods. In DeFi, this norm remains informal at best.

For public goods funding programs. Gitcoin Grants, Optimism RetroPGF, Octant exist—but are underutilized for transversal intelligibility infrastructure. Coordinating one or more dedicated rounds around this category would have outsized impact.

A concrete example is emerging right now, covering part of that terrain. TheDAO Security Fund, endowed with 170 million dollars, launched its first round via Giveth on April 21, 2026, with 500 ETH in matching funds for Ethereum and L2 security projects. Its announced scope (incident response, security research, on-chain investigation, security tooling, threat intelligence) covers the defensive layer of transversal watchfulness. It does not cover the analytical and editorial layer that produces meaning outside of incidents: independent mapping, non-technical observatories, public critical analysis. Still, this marks a first: a public goods funding mechanism explicitly targeting part of this space. Pharos and DeFiScan are eligible. Editorial observers likely are not. The signal matters: the ecosystem is beginning to acknowledge that it must fund what it does not directly consume.

For users. Pharos has a Giveth page. DeFiScan as well. SEAL accepts donations. Those who rely on these resources daily can support them directly—not as charity, but as informed self-interest. If these structures disappear, users will be left navigating only protocol-aligned narratives.

For the ecosystem as a whole. Cite your sources. Tag contributors. Acknowledge dependencies. Publicly recommend the free tools you rely on. Less material than the other levers, but not trivial. Recognition sustains volunteer motivation and helps lay the groundwork for a more formalized economy of attribution.

A sixth lever deserves to be made explicit. The coming temptation will be to outsource to artificial intelligence what humans have so far provided for free. Why fund Pharos when an LLM can scrape and summarize peg data? That logic accelerates the devaluation of functions whose value lies precisely in their human, situated, independent nature. An LLM can process data. It cannot hold a public position against the interests of a protocol it does not understand. Confusing the two will degrade DeFi’s epistemic layer more reliably than any exploit.

To close

A sustainable DeFi is not the one with the most sophisticated protocols. It is the one in which the collective capacity to understand what is happening is funded as seriously as the capacity to generate yield. That is not the case today. This gap is a systemic risk no dashboard currently measures.

The structures carrying this function—Pharos, DeFiScan, independent researcher-practitioners, and editorial observers such as the one behind this piece—will not hold indefinitely without material support. Every founder who burns out, every researcher who shifts to paid work, every independent observer who shuts down their blog creates another blind spot.

The question raised by the April 2026 events is not whether DeFi will survive Kelp or USR.

It is whether DeFi will equip itself with the material means to keep making sense of itself.

Without those means, the next crises will be navigated blindly—not for lack of tools, but for lack of people still holding them.

2. Delegate Without Handing Yourself Over

Mon, 13 Apr 2026 12:00:00 +0000

Series opening: Self-Custody, But How Far?

The problem is no longer whether to use AI. The problem is how far to let it in.

Too much of the AI debate hardens into a lazy binary. On one side: acceleration without brakes — wire everything up, automate everything, move faster. On the other: principled distance — keep away, distrust everything, and congratulate yourself for your supposedly clearer austerity.

Neither posture survives ordinary life for very long.

The serious question is harder: what kinds of delegation can we accept without losing command of our own practice?

Web3 readers already have one useful instinct. They learned in other domains that performance alone does not make a system legitimate. You still have to ask who controls what, within which limits, and under what conditions of reversibility.

Applied to AI, that instinct now has to become a discipline.

Assistance, delegation, abandonment

The first task is to separate what the moment keeps trying to blur.

Assistance is the simplest case. The tool helps, suggests, reformulates, accelerates. It expands our capacity to act without displacing the center of decision.

Delegation begins when the tool no longer just helps. It preserves context, reorders, filters, prepares, monitors, preselects, or triggers certain operations within a defined perimeter.

Abandonment begins when delegation becomes too broad, too opaque, or too comfortable to remain genuinely governed. The system still appears to serve us, but it becomes increasingly hard to explain what exactly it is doing, why we trust it, and what it would take to take back control.

These are not the same thing. And yet many current uses slide from the first into the second, and sometimes into the third, without ever naming the shift.

The real test is ease

The problem of delegation does not show up first in extreme cases. It shows up in easy ones.

The decisive moment is not usually when a system suggests some absurd transaction. It is when, day after day, it becomes normal for it to read for us, sort for us, reconnect for us, monitor for us, and prepare work before we even arrive.

Here again Morin helps. The danger is not only spectacular error; it is dependence quietly formed in the continuity of service. We then begin to mistake an increase in power for a thinning of presence.

What threatens autonomy is not always hostility. Often it is softness.

Where individual responsibility begins

This needs to be said plainly.

Protocols, builders, and tool providers have duties of clarity, boundary-setting, and prudent design. But they cannot carry the whole burden of educating users or operators determined to delegate everything without discernment.

There is, therefore, an irreducible form of individual responsibility that has to be recovered.

Vitalik Buterin offers one of the clearest practical expressions of this concern when he argues for a local-first, compartmentalized, tightly bounded LLM setup. The value of that setup is not that it provides a universal template. Its value is that it makes one thing unmistakable: as AI becomes an ordinary environment for action, the decisive question is no longer just model capability, but the limits we are willing — or unwilling — to place around delegation.

That does not mean becoming paranoid. It does not mean retreating into some heroic austerity. It means asking, with a little firmness: what am I actually trying to gain, and what am I unwilling to lose in the process?

The question is harder than it sounds. Most of us want the same things: more speed, more clarity, less friction, less fatigue, more continuity. The trouble is that these gains have a cost — and the cost is not always paid in data. Sometimes it is paid in habits of dependence.

A minimal discipline of delegation

There is no perfect stack. There is, however, a minimal discipline simple enough to use without turning into a purity program.

1. Keep the core of judgment in your own hands

Exploration, mapping, summarizing, and comparison can all be assisted. But the moment a reading turns into an actual judgment should remain visible. If you no longer know where help ends and pre-decision begins, you have already delegated too much.

2. Bound action before you expand assistance

The danger is not that a system is intelligent. It is that it is intelligent and deeply connected. A text assistant does not carry the same risk profile as a tool wired into messaging, sensitive documents, automation, a wallet, or execution rights. The wider the access, the sharper the boundaries need to be.

3. Prefer legible layers to magical ones

A good tool is not just one that works. It is one whose role can be described simply. If it takes five explanations to understand what it sees, what it keeps, what it prepares, what it can trigger, and what stops it, the delegation is already too opaque.

4. Treat comfort as a warning sign

Comfort is a real benefit. It is also often the moment when vigilance starts to drop. A mature AI practice does not reject ease; it refuses to become hypnotized by it.

5. Build the way back before you need it

Healthy delegation has to remain reversible. If you cannot unplug a tool without losing the intelligibility of your practice, access to essential information, or the ability to act without it, then the tool is already taking up too much ground.

A workflow more useful than an ideal stack

For many everyday uses, a clean discipline of delegation already changes a great deal.

Keep in your own hands

Decisive interpretation, capital allocation, reputationally sensitive choices, and any access involving keys, secrets, or wide permission scopes.

Safe to assist

Exploratory research, topic mapping, summarization, reformulation, comparison, and note structuring.

Delegate only inside a bounded perimeter

Monitoring, draft preparation, classification, low-consequence automations, and narrow coordination tasks.

Never delegate without explicit friction

Signatures, fund movements, external messages, irreversible actions, and any modification of critical permissions or system parameters.

What this second piece was trying to make practical

We are not going to get through this moment by pure refusal. Nor will we get through it by delighted surrender.

What we need is something else: a way of living with AI without handing over, in a single motion, the very faculties it makes harder to notice.

That calls for less purity than poise. Less grand moralizing than an art of limits.

The third piece moves the question onto its sharpest testing ground: DeFi, where every delegation eventually takes the form of a permission, a bound, a risk, or an architecture.

1. The Next Frontier of Self-Custody

Mon, 13 Apr 2026 11:00:00 +0000

Series opening: Self-Custody, But How Far?

We learned not to leave our assets with custodians. We have not yet learned not to hand our judgment too quickly to new intermediaries.

Crypto began with an intuition that looked technical at first and turned out to be much larger: wherever an intermediary can be removed from the equation, it is usually better to learn how to live without it.

From that intuition came wallets, habits, slogans, and forms of caution. Not your keys, not your coins did more than teach a generation of users to distinguish convenience from control. It changed the way they moved through the digital world.

What has not happened yet is the extension of that discipline to our own faculty for judgment.

We know how to ask where the funds are, who controls permissions, where the points of capture lie. We are far less used to asking where the model runs, what data it sees, how much of our work it absorbs, what it prepares on our behalf, what it saves us from doing, and what kind of dependence takes shape as it becomes hard to do without.

The center of gravity has moved. For years we talked about empires of data. We now have to talk about empires of delegation.

What has actually changed

Web2 mostly captured traces: our clicks, our preferences, our social graphs, our attention.

This moment adds a more intimate layer. AI enters our notes, archives, queries, reading habits, summaries, work routines, information triage, draft formulations, preparatory judgments, and in some cases already parts of action itself.

It is no longer just a tool. It becomes an environment — sometimes a work environment, sometimes a mental one.

And once a tool becomes an environment, the question changes. The point is no longer just whether it is useful. The point is what it reorganizes in us: how we read, rank, remember, anticipate, and decide.

For a web3 reader, that shift is decisive. Trust minimization no longer concerns only assets, contracts, or execution rails. It reaches something harder to preserve: discernment.

A deeply crypto contradiction

This is not really about hypocrisy. The problem is more uncomfortable than that — and more revealing.

Crypto has always been drawn to tools that produce an edge: more speed, more clarity, more informational asymmetry, more capacity to explore, compare, and execute. It makes perfect sense that the space would be among the first to embrace persistent copilots, augmented workflows, agentic layers, and systems that preserve context while compressing time.

So the contradiction is not moral. It is structural.

Two legitimate demands are colliding. On the one hand: do not depend on a trusted intermediary. On the other: gain speed, clarity, and processing power through systems that keep getting stronger.

The first built a culture. The second is beginning to rearrange its habits.

No slogan will resolve that tension for us.

The age of setups

That is why the current fascination with AI setups deserves to be taken seriously.

One builder runs Claude through Telegram. Another has agents pass task state through .md files. A third organizes memory in Obsidian, plugs in search tools, ties the whole thing into coding or research workflows, and starts the day with the sense that the machine has already been working through the night.

None of this is scandalous in itself. But all of it says something precise about the present.

We are no longer looking only for answers. We are looking for continuity.

We want systems that preserve context, reconnect scattered fragments, surface what must not be forgotten, reorder notes, watch what needs watching, prepare the ground, and make us feel that part of the work has already been done before we even sit down.

That promise is powerful. It is also exactly where sovereignty has to be rethought.

The question is no longer only whether a tool is spying on us. We also have to ask what it is taking over inside our relation to the world: memory, monitoring, pre-sorting, classification, recombination, the first movement of discernment. These operations look secondary — until we realize they already shaped a decisive part of the way we act.

Morin, or the useful bad news

This is where Edgar Morin becomes newly useful.

Morin forces us to think together what we would rather keep apart: autonomy and dependence, local power and systemic effects, service rendered and gradual dispossession. Applied to AI, the lesson is severe but clarifying: a tool of augmentation can become an infrastructure of relinquishment without ever ceasing, locally, to be useful.

The danger does not necessarily arrive as a visible catastrophe. More often it settles in under the appearance of comfort.

You notice quickly when you left your funds with a dubious custodian. You notice later when you have outsourced too much of your reading, your memory, your sense-making, or the preparation of your decisions.

Cognitive dependence rarely looks like coercion. More often it looks like help that has become second nature.

The issue is not only the models. It is also the surrounding software.

The problem cannot be reduced to frontier models, major labs, or the opposition between local and cloud.

It also concerns the environment AI is helping to proliferate around protocols: dashboards, wrappers, research tools, specialized assistants, monitoring layers, coordination services, micro-products built quickly — sometimes carefully reviewed and robust, sometimes not.

We need to be precise here. Using AI to build software is not, in itself, a problem. Not every AI-assisted build is vibe coding in the strong sense.

The issue begins elsewhere: when the cost of producing a service falls much faster than the cost of reviewing it, understanding it, and qualifying it.

At that point we do not just have more tools. We have more surfaces of trust to evaluate.

And in web3, where a dashboard, an interface, or a copilot can shape a risk reading, an allocation, a signature, or an execution path, that shift matters a great deal.

What this first piece wanted to make visible

The question, then, is not whether AI is compatible with web3 culture. It already is — almost too well.

The real question is more awkward: can crypto extend to cognition, organization, and action the same demands for sovereignty it learned to formulate for assets?

We now know where the fault line runs. A culture of self-custody that never asks what it is handing over to cognitive intermediaries may end up missing the most important stage of its own development.

The next piece starts from there: how to recover a discipline of delegation firm enough to matter and simple enough to use.

Intro: Self-Custody, But How Far?

Mon, 13 Apr 2026 10:00:00 +0000

Crypto taught itself not to leave assets in someone else’s hands. The next question is whether it can bring that same discipline to judgment, organization, and eventually to action itself.

Why this series

Crypto proved something essential long before most of the digital world was ready to hear it: when you can remove an opaque intermediary, you usually should.

That instinct gave us wallets, habits, slogans, and architectures. Over time it gave us something larger than tooling: a culture. Self-custody. Trust minimization. Resistance to capture. An insistence on clear permissions and visible points of control.

And yet something is shifting as AI becomes ordinary.

The same circles that learned to distrust intermediaries around assets are now increasingly willing to hand over their notes, reading habits, research workflows, contextual memory, preparatory judgments — and soon parts of their ability to act — to systems they do not fully understand and do not really control.

So the problem is no longer just data. It is delegation.

That is the real turn. Crypto already knows how to ask the right questions when money is involved: who controls what, who holds the keys, where the capture points sit, which trust assumptions are being accepted.

What remains uncertain is whether that rigor can be extended to thought, to organization, and to the conduct of action.

What these essays are after

We learned self-custody for assets.

We still have to learn it for judgment.

That is the thread running through these three pieces.

The first names the contradiction: we are not only living in the aftermath of Web2 and its empires of data. We are moving into a world where empires of delegation are beginning to take shape.

The second tries to recover a workable discipline. AI is no longer just an answer engine. It is becoming a work environment, a contextual memory, an organizational layer, and in some cases already a preparation for action.

The third explains why DeFi is one of the first places where this question can no longer be postponed. Sooner or later, every delegation shows up as a permission, a limit, a risk, or an architecture.

Edgar Morin is the quiet philosophical thread behind the series. Not as borrowed authority, but as a way of reading the moment: hold together autonomy and dependence, local power and systemic effects, genuine service and gradual dispossession. In other words, learn how to think in a complex environment without pretending that a local fix will settle the whole matter.

What this series does not promise

Not a universal stack. Not a new form of digital purity. Not a lazy indictment of AI.

It aims at something more useful: a way of telling apart what can be delegated, what should remain under direct control, and what we should now demand from the tools we let into our practice.

The three pieces

1. The Next Frontier of Self-Custody

The diagnosis: what AI is actually changing inside a culture that thought it already knew what digital sovereignty meant.

2. Delegate Without Handing Yourself Over

The practical and philosophical core of the series: how to distinguish assistance from delegation, and delegation from abandonment — then recover a discipline strict enough to matter and simple enough to live with.

3. DeFi as a Laboratory of Delegation

The proving ground: dashboards, wrappers, agents, vibe coding — and then two contrasting cases, Money League and Polaris, that make the problem impossible to miss.

More Dashboards, Less Wisdom? The DeFi Paradox.

Tue, 07 Apr 2026 09:00:00 +0200

In a world saturated with dashboards, alerts, and metrics, the real issue is no longer how to see more. It is how to choose an exposure, what powers to tolerate, and how far we are willing to preserve freedom of action without sacrificing it to total visibility.

Omniscience Without Mastery

Our age combines an odd mixture of omniscience and impotence.

Signals proliferate, clues pile up, traces multiply, and yet our ability to turn them into something legible, ordered, and politically usable keeps slipping away. In DeFi, that condition becomes almost experimental. Everything seems observable. Flows are public, metrics are abundant, and interfaces for reading the system have multiplied at great speed.

But this apparent triumph of visibility leaves a harder question untouched: what do we do with a world we can inspect ever more closely without actually learning how to inhabit it better?

The Observability Stack — and Its Limits

The ecosystem has already built an impressive observational infrastructure. DeFiLlama, for example, does more than aggregate numbers: it defines its own metrics, distinguishes TVL from borrowed funds, and reminds users that net flows often say more than a badly interpreted stock variable. L2BEAT, meanwhile, has forced a rollup conversation centered on trust assumptions, decentralization stages, and residual powers. DeFiScan does something similar for DeFi protocols, explicitly acknowledging that a decentralization framework does not measure smart contract risk, nor economic risk in its entirety. This is real progress: DeFi is no longer short on ways of reading itself.

But that rise in visibility has exposed a newer difficulty. The problem is no longer simply that risk is poorly seen; it is that it is seen through a plurality of heterogeneous cuts, rarely commensurable and often in tension with one another. Each actor produces its own surface of intelligibility, its own method, its own way of ordering uncertainty. One tracks flows, another maps powers, a third watches code vulnerabilities, a fourth models credit dynamics, a fifth focuses on the fragility of a monetary subsystem. None of this is useless. But a pile of specialized readings still falls short of a public intelligence of risk.

A Reading Map

Reading DeFi now requires an order of operations. Not in order to collect tools, but to discipline judgment before action.

From Signals to Judgment

Read from top to bottom: start with power, then context, then failure, then risk design, then mediation, then money.

1Powers / Trust Assumptions

L2Beat DeFiScan

2Terrain / Market Context

DeFiLlama Dune GrowThePie

3Breaks / Failure Modes

ChainSecurity Forta Hypernative

4Risk Parametrization

Chaos Labs Gauntlet

5Credit / Vaults / Mediation

Block Analitica Credora CuratorWatch vaults.fyi Morpho

6Monetary Layer

Pharos Bluechip Stablewatch

The method is straightforward: power → context → failure → parametrization → mediation → money. The most common mistake is to reverse that order — to begin with yield, a polished dashboard, or the most visible metric, when the decisive questions are still questions of power and dependence.

The Missing Mediation: Toward a Public Grammar of Risk

At this point, a new requirement emerges. A mature ecosystem cannot rest on a scattered landscape of dashboards, frameworks, scorecards, and expert tools alone. It also needs a more legible mediation layer: a public synthesis capable of offering a common point of entry, making broad risk profiles comparable, and orienting judgment without pretending to exhaust reality. Not a magical score that crushes every difference into an opaque verdict, but an aggregation clear enough to guide and decomposable enough to remain honest.

In other words, DeFi probably needs more than a collection of specialized tools. It needs a public grammar of risk.

The Privacy Constraint

And this is precisely where a decisive limit has to be introduced. From an Ethereum point of view, the answer cannot be to celebrate ever more observation, ever more monitoring, ever more traceability, as though full transparency were the natural endpoint of a healthy system. In Why I support privacy, Vitalik argued in April 2025 that privacy is not a luxury but a safeguard of decentralization itself: whoever controls information already holds a form of power. The Ethereum Foundation framed the same idea in more institutional terms: privacy is the freedom to choose what you share, when you share it, and with whom.

So the right kind of legibility is not the kind that makes everything visible to everyone at all times. It is the kind that makes structures, dependencies, and powers more intelligible without abolishing users’ room for withdrawal, discretion, and autonomy.

That is why DeFi’s problem is not merely a tooling problem. It is an orientation problem.

A Doctrine of Exposure

A good craftsperson does not begin by choosing tools. They begin by clarifying their intention. In DeFi, that means choosing a doctrine of exposure.

You can look at ten dashboards, two decentralization frameworks, three audits, a handful of runtime alerts, and a synthetic rating layer, and still fail to clarify anything essential. Because these instruments do not observe the same layer of reality. One maps the terrain, another maps powers, a third watches for software failure, a fourth models credit or liquidation dynamics, a fifth tracks the stability of a monetary subsystem. A map is not a compass. And a stack of screens is not yet a doctrine of exposure.

So the real question is not: which tools should I use?

The real question is: what kind of actor do I want to be in this environment?

That sounds abstract, but it is deeply practical. Are you primarily seeking yield? Capital preservation? Strong exit liquidity? Maximum proximity to self-custody? Strictly reduced trust assumptions? Limited experimental exposure? Why? On what time horizon? Until that orientation is clarified, tools mostly serve to compensate for the absence of strategy. And they do it badly.

Simplicity, Trust Minimization, and Order of Reading

A second correction has to be added here, and it comes directly from Vitalik. Not every interpretive problem should be solved by adding more interpretive layers. Some should be solved by greater structural simplicity. In Simplifying the L1, published in May 2025, Vitalik argues that Ethereum should move toward an architecture that is easier to understand, audit, and maintain.

That argument is decisive for DeFi: a system that can only be inhabited through a permanent caste of analysts, curators, scorecards, and alerts is not yet a broadly intelligible system. Observability is useful, but it should not become a substitute for architectural sobriety.

This is also what restores real meaning to trust minimization. Stages, for Vitalik as for L2BEAT, are not just a technical taxonomy. They force a political question into the open: who can still override the code, under what conditions, with what legitimacy, and with how much reaction time for users?

Once the problem is framed that way, practice changes. A cautious depositor should not begin with yield. They should begin with power. Who can upgrade? Who can freeze? Who can redirect risk? Is there a real exit window? A user should not be looking merely for a “good product,” but for a regime of exposure compatible with their tolerance for dependence.

Where Practice Becomes Philosophy

The practical dimension becomes clearer at that point. Maturity does not mean watching everything. Maturity means being able to connect a signal to a possible action. A trust assumption that is too heavy should be enough to keep you out. Fragile exit liquidity should lead you to size down. Excessive complexity should push you toward a simpler, perhaps less profitable, but more legible system. A strong dependency on an interface, a multisig, a curator, or a freeze-enabled stablecoin should be enough to make you give up some yield in order to preserve some freedom.

And this is where practice finally becomes philosophy.

Because behind every doctrine of exposure lies an ethic of action. One can inhabit DeFi according to a speculative logic, where every signal is treated as tactical advantage. One can inhabit it according to a prudential logic, where the central task is to reduce avoidable blindness. One can inhabit it according to an ethic of autonomy, where the priority is to remain on the side of self-custody, privacy, simplicity, and legible power. And one can inhabit it according to a more ambitious institutional logic: helping build stronger forms of public judgment, better standards, more honest risk taxonomies, and more robust mediations between code and action.

That, it seems to me, is where the right reading of DeFi now begins. It does not merely need better tools. It needs a discipline of judgment capable of ordering signals without sacrificing Ethereum’s political ends: privacy, self-custody, censorship resistance, simplicity, open source, and minimized trust assumptions. The Ethereum Foundation now says this explicitly in its February 2026 DeFi position. Vitalik, for his part, provides the conceptual vocabulary for understanding why it matters: more legibility, yes — but not at the price of a more sophisticated form of dependence.

So the answer is neither fatalistic withdrawal nor dashboard superstition.

It is a doctrine of exposure.

Then a discipline of interpretation.

And behind both, a certain idea of freedom.

How Can an Open Collective Hold Together?

Sun, 22 Mar 2026 00:00:00 +0000

Why voting and incentives are not enough

A sound system of governance does not rest on voting mechanisms, incentives, or technical excellence alone. It also requires a legible direction, credible norms, forms of appropriation, and figures capable of embodying what they defend. The real question, then, is not only how decisions are made. It is this: how does a common order emerge that free participants can regard as legitimate, useful, and worth sustaining?

Beyond command: the problem of sustaining participation

We still tend to associate power with the ability to impose. To govern, on this view, is first to command, arbitrate, discipline. That picture is not always wrong. But it quickly becomes inadequate when what must be sustained is an open collective, composed of free, mobile, heterogeneous individuals who can withdraw more easily than they can be made to obey.

In such a setting, the central problem is not simply how to secure a decision. It is how to sustain participation over time. A collective does not hold together for long through constraint alone, or even through the mere convergence of interests. It holds when a shared direction becomes clear enough, just enough, and credible enough to be taken up by those who participate in it.

Gentle power: orient rather than compel

That is what the idea of gentle power makes it possible to think.

We can understand it as a form of government that seeks less to compel than to orient. It abolishes neither authority nor rules, but places them within a more demanding frame: a line must appear not merely acceptable, but worthy of being followed. Gentle power works through vision, example, shared norms, and the quality of institutions and relationships. It does not produce mere compliance; it seeks assent.

Gentle power vs. shallow influence

This is also what distinguishes it from influence in the shallower sense of the term. Marketing and storytelling can capture attention, sometimes even enthusiasm. They are not enough to found a common order. Gentle power does not simply aim to persuade an audience. It aims to make conscious cooperation possible: a form of cooperation in which members of a collective can recognize the reasons for acting as reasons they can genuinely make their own.

The distinction matters. Where influence often seeks effect, gentle power demands coherence. It rests less on what is proclaimed than on what is made visible, what is rewarded, what is tolerated, and the way rules are actually lived.

Implications (1): leadership

This changes, first of all, the meaning of leadership. From this perspective, leadership is not primarily a matter of concentrating authority or deciding faster than everyone else. It consists in giving an intelligible direction, setting criteria, and making the collective more capable of governing itself. Example becomes central here. A group places less trust in principles that are merely stated than in conduct it sees embodied.

Implications (2): legitimacy

It also changes the way legitimacy is understood. Legitimacy does not flow from procedure alone, even when procedure is flawless; nor does it arise from effectiveness alone. It emerges from a subtler alignment: between clear rules and intelligible decisions, between words and conduct, between a real possibility of participation and a just way of handling disagreement. A difficult decision is accepted more readily when it appears to proceed from a common world that can be understood, rather than from a simple balance of power.

Implications (3): endurance

And it changes the way a collective endures. Constraint may secure immediate execution; it rarely produces deep fidelity. For an organization to last, something else must take hold: the sense that it deserves time, attention, perhaps even a measure of renunciation. That is the point at which norms and values cease to be decorative. They become conditions of stability.

Why this matters for protocols and DAOs

This is especially important for protocols and DAOs. In such environments, coercion is structurally limited. One can leave a community, stop contributing, sell tokens, fork the code, or direct one’s attention elsewhere. A protocol therefore cannot rely, over time, on obedience. It must generate trust, legibility, a sense of fairness, and a form of attachment that is not purely opportunistic.

And yet many organizations are better at aggregating preferences than at articulating a common direction. They excel at managing technical variables, but struggle with deeper questions: what matters here? What kind of conduct do we want to encourage? What limits do we want to set? What deserves protection beyond immediate utility?

Where gentle power begins

Gentle power begins precisely there: in the capacity to form a shared sense rather than merely juxtapose interests.

Norms as operative architecture

That requires serious work on norms. Not abstract values filed away in a manifesto with no practical force, but operative standards: what is valued in exchanges, what is expected of central figures, what is built into rituals, tools, contribution mechanisms, and forms of recognition. Culture is not an optional layer. It is an essential part of the architecture of governance.

Narrative and play

This is why narrative and play matter here.

Narratives, when they are not reduced to exercises in communication, give shape to a shared experience. They connect the present to memory, ambition, and a certain idea of what the collective is trying to become. They help transmit landmarks, name tensions, and make choices more intelligible.

Play serves a related purpose. It creates a space in which rules can be learned, situations tested, roles explored, and the effects of an institutional design grasped without immediately hardening into rigid norms. In both cases, the point is not to entertain the collective, but to form it.

Risks and conditions of legitimacy

None of this should be romanticized. Gentle power can slide into a more palatable form of manipulation. The language of vision, values, or community can easily conceal quite conventional relations of power. That is why this form of power is legitimate only on one strict condition: those who orient others must themselves accept demanding standards of coherence, transparency, and contestability.

Structure still required

Nor can any organization rest on diffuse assent alone. It also needs rules, procedures, responsibilities, and, at times, sanctions. Gentle power does not abolish structure; it makes it finer and more inhabitable. It does not replace institutions. It raises the standard they must meet.

Time and political work

And this approach takes time. Forming a collective capable of sustaining a shared horizon, living through disagreement, and integrating difference without dissolving into fragmentation requires real political work. It requires pedagogy, spaces of translation, suitable formats, and a certain patience. Without that, the appeal to assent remains verbal.

What gentle power is (and is not)

Gentle power is therefore neither weakness nor vague moralism. It is a demanding hypothesis about how free beings may be guided. It reminds us that a collective becomes more solid when it learns to elicit forms of conduct rather than extract behavior. And in the end, the quality of an organization may be measured less by its capacity to constrain than by its capacity to make people willing.

Note

Note — This reflection stands at the intersection of several lines of thought: Spinoza on affects and the power to act; Tarde on imitation and the social circulation of conduct; Foucault on power as a productive relation; Charles Taylor on the social imaginary and on forms of recognition that make a common order livable. One distinction should nevertheless be kept clear: soft power, in the strict sense, belongs first to Joseph Nye, who defined it as a power of attraction rather than coercion.

Philosophy on Arem

Transversal Watchfulness

What protocols fund—and what they don’t

Servile watchfulness, transversal watchfulness

What Weick called sensemaking

The funding blind spot

What this concretely asks for

To close

2. Delegate Without Handing Yourself Over

The problem is no longer whether to use AI. The problem is how far to let it in.

Assistance, delegation, abandonment

The real test is ease

Where individual responsibility begins

What this second piece was trying to make practical

1. The Next Frontier of Self-Custody

We learned not to leave our assets with custodians. We have not yet learned not to hand our judgment too quickly to new intermediaries.

What has actually changed

A deeply crypto contradiction

The age of setups

Morin, or the useful bad news

The issue is not only the models. It is also the surrounding software.

What this first piece wanted to make visible

Intro: Self-Custody, But How Far?

Why this series

What these essays are after

What this series does not promise

The three pieces

1. The Next Frontier of Self-Custody

2. Delegate Without Handing Yourself Over

3. DeFi as a Laboratory of Delegation

More Dashboards, Less Wisdom? The DeFi Paradox.

Omniscience Without Mastery

The Observability Stack — and Its Limits

A Reading Map

The Missing Mediation: Toward a Public Grammar of Risk

The Privacy Constraint

A Doctrine of Exposure

Simplicity, Trust Minimization, and Order of Reading

Where Practice Becomes Philosophy

How Can an Open Collective Hold Together?

Why voting and incentives are not enough

Beyond command: the problem of sustaining participation

Gentle power: orient rather than compel

Gentle power vs. shallow influence

Implications (1): leadership

Implications (2): legitimacy

Implications (3): endurance

Why this matters for protocols and DAOs

Where gentle power begins

Norms as operative architecture

Narrative and play

Risks and conditions of legitimacy

Structure still required

Time and political work

What gentle power is (and is not)

Note