TechFlowPost has published an essay by Conall O'Reilly that asks a foundational question: who defines Ethereum. The article treats Ethereum as both a philosophical revolution and a practical system. In that framing, it is a cultural consensus built on general-purpose, shared, neutral infrastructure, but it is also a set of software tools and an ongoing R&D effort. Specifications, engineering implementations, and continuous research together make up what people call the Ethereum protocol.
When philosophy becomes protocol
The essay starts from the idea that language carries reason and philosophy, yet any attempt to describe a concept also reduces it. Ethereum's core ideas, it argues, can be expressed more formally through mathematics, logical architectures, and code. In that sense, the Ethereum protocol is the product of an attempt to make those ideas real.
If Ethereum is understood as a shared computational reality, the piece says it rests on three ontological pillars:
- cultural convergence, where participants coordinate socially and culturally around a common definition of the Ethereum protocol;
- cryptographic truth, which defines what counts as truth inside the system;
- majority representation, a mechanism that relies on a social credential outside the system, certified through cryptography, to determine the majority view behind that shared reality.
Under that structure, cultural agreement around the protocol comes first, then a truth standard that can be checked cryptographically. A majority-representation mechanism lets participants converge on Ethereum's instantaneous state. That state defines what Ethereum is at a given moment, and observers can identify it only because they hold some representative credential that comes from outside the system.
The article places cultural consensus around the protocol's definition above everything else. That, it says, is why truth is strongly subjective. Weak subjectivity still matters at a higher level: observers can build subjective axioms on top of cryptographic truth and use them as a basis for further knowledge.
In practice, Ethereum is a network of many independent users running the same version of Ethereum client software and exchanging information over communication layers such as the internet. Protocol definitions combined with information flow drive the network toward a converged and updated protocol state. Each participant, using the initial protocol definition, cryptographic verification, and majority representation, converges on the network's instantaneous and determinate state.
The essay says majority representation is necessary because the current state has to be derived deterministically from a sequence of fixed operations in network history, while observers cannot be assumed to receive events in the same order. They therefore need to vote based on what they observe. That is how a network of thousands of computers running the same software can agree on Ethereum's present state and update that shared definition in real time as users act.
From technical definition to a human coordination problem
The article argues that Ethereum evolved from a static technical definition into a continuously changing, coordinated network, but its foundation is still cultural consensus at the protocol-definition layer. That culture, in turn, comes from the combined beliefs of the people who care about Ethereum.
It introduces what it calls a cultural paradox. Cryptography can tell right from wrong at the technical level, but culture has no formal correctness test because culture is its own standard. Arguments for or against cultural change carry an internal paradox and struggle to justify themselves. Calling for change is already an act within the culture one is trying to reshape.
That leads into a governance paradox: if governance shapes culture, what shapes governance. Does cultural consensus come first and then governance rules, or does governance arrive first and direct culture in a systematic or ideological way. The essay does not resolve the question neatly.
Applied to Ethereum and its roadmap, governance becomes a human coordination problem. Participants have to agree both on what the protocol is and on how it should evolve. Without a company or a state to serve as the source of cultural truth, the task becomes unusually difficult. If people cannot agree on the object of consensus itself, reaching agreement becomes harder.
The essay contrasts this with governance in the physical world, where power often comes from direct geopolitical force, such as who controls the most nuclear capability. Those with hard power can influence or control subjective definitions, including law and order. Ethereum, by contrast, runs on soft power. Decisions are often shaped by people who carry cultural influence rooted in precedent, are deeply aligned with Ethereum's ideas, and have standing inside the ecosystem.
The Ethereum Foundation is one such actor. The article describes it as the de facto originating institution for Ethereum's ideas and says it carries significant cultural weight. It also gives a snapshot of the client landscape: as of September 2025, 32 teams across 15 organizations maintained 13 independent Ethereum protocol clients.
Those contributors coordinate through All Core Developers, or ACD. According to the article, Ethereum's governance rules are established there. Related documents define how Ethereum Improvement Proposals, or EIPs, are designed and reviewed, and how network upgrades proceed. Historically, ACD coordination itself was loose and not clearly codified. Even now, the essay says, people and roles are organized more through practice than through rigid written rules.
At a practical level, ACD's job is to coordinate the ecosystem around software forks that change Ethereum's protocol definition at a specific point in time. The process, as described in the essay, begins with researchers and developers proposing changes, then finalizing them, testing them, publishing the final specification, implementing the code, and finally activating the changes across the network on a set date. A fork cycle can take months and, in some cases, up to a year.
The question Ethereum is trying to answer
The article frames Ethereum's central proposition this way: what does it mean to achieve human coordination through computation. The truth basis behind that proposition is philosophical and cultural, but the answer has to appear in technical and architectural form. In other words, the philosophy has to become a usable software protocol before Ethereum's question can be answered.
The essay then asks what the actual question is if Ethereum itself is the question. Its answer: Ethereum explores the meaning of existence under computational coordination. The goal is to turn the abstract idea of a Turing machine into an interactive domain with computational existence.
At the local level, self-sovereignty holds. A user can observe a computational world that exists only on a local machine and has full control over that strongly subjective environment. Once there is more than one observer, a minimum level of strongly subjective consensus is needed so that people can observe and interact in a broader interpersonal space. In that sense, every observer subjectively defines truth in a local environment.
Information and data do not carry meaning on their own, the article says. Definitions and perspective give them meaning. Truth also depends on scarcity. If any observer can repeat a statement without limit, that statement cannot count as true. Coordination therefore depends on agreeing on definitions and preserving the scarcity of true statements inside a correctness framework.
To make that more concrete, the essay uses electronic cash. Cash is presented as a simple coordination game in which each participant has a balance, and only the controller of that balance can authorize a transaction that reduces it and increases someone else's by the same amount. If N users participate, two things are required:
- agreement on the game itself, including programming rules and the cryptographic tools involved;
- the ability to verify the correctness of the game's instantaneous state and keep the game moving over time.
Cryptography builds the spatial dimension by creating the scarcity needed for truth. In technical terms, balances are tracked by account, each account has a cryptographic identity such as a public key, and users hold private keys so that only the owner can authorize valid transactions.
There is also a time dimension. To know everyone's current balance, the full ordering of historical transactions has to be clear; otherwise it is impossible to tell whether past events are feasible. The article describes this coordination game as something embedded in the logic of a Turing machine because it includes both storage, the balances, and execution, the transactions.
Its larger point is that putting coordination on top of a general-purpose Turing machine has a decisive advantage: it can run every computable game. From there, the piece reduces coordination to two meta-dimensions, a cryptographic spatial dimension that creates truth scarcity and a temporal dimension that gives a coordination game order. Together they define instantaneous state and future state transitions. That is what the essay calls consensus.
Consensus, in this view, is the irreducible truth basis that supports the existence of a computational domain and makes coordination possible. The essay then distills Ethereum into a bottom-layer architecture: a Turing-complete computation model, a cryptographic spatial dimension that preserves scarcity of machine actions, addressable computational storage, operations that act on that storage, and a time dimension that turns machine state into structured truth at different moments. In the article's plain-language version: things exist inside addresses, and those things change over time according to actions initiated by users associated with those addresses.
Ethereum as a real system
The essay then moves from theory to engineering. For Ethereum to exist in practice, it has to be deployed at distributed scale. There is no central coordinator. The network is made up of many computers working together in real time. Since consensus cannot be broken apart, the article treats it as an architectural axiom: a set of computers jointly confirms the present state of something and coordinates the changes that happen to it over time.
In Ethereum, that changing state lives in the Ethereum Virtual Machine, or EVM, and users inside the network generate state changes through general-purpose actions. The article begins its description of the EVM with storage. Storage consists of code and arbitrary data, addressed by location and linked to cryptographic identity. External operations on storage are broadcast to the network as messages and coordinated through the preset consensus process.
Execution is where the article slows down. The result of a Turing machine's execution is not predictable in advance, it says, so if EVM computation is to remain deterministic, computation has to happen somewhere. Yet verifying whether a result is correct is algorithmically different from performing the original computation. Verification complexity does not rise in lockstep with the complexity of the original task.
For that reason, the bottleneck in proving correct execution is data rather than compute. Compute is needed to produce the result, while verification does not require large amounts of compute. The implication is that consensus can, in principle, cover general-purpose execution at any scale, with the real limit coming from the network's ability to store and transmit data. Since data is finite, the network needs rate-limiting mechanisms that turn data capacity into a measurable resource.
The article then ties this back to Turing completeness. Because the set of possible operations is unbounded, hard-coding capabilities into the architecture, or embedding precompiles inside the EVM, would limit the kinds of coordination Ethereum can support. The EVM's end goal, the piece argues, is extreme generalization and abstraction so that any implementable function can be written in code.
That abstraction can extend even to cryptographic identity. EVM code attached to an address can call arbitrary custom verification functions rather than being locked into a fixed public-key scheme. The article says not only computation but also the forms of execution and storage on Ethereum should be generalized, so that Ethereum can support a full spectrum of possible modes of existence.
The physical universe does not have a closed boundary of existence wrapped in consensus, but Ethereum does. In Ethereum, knowledge is global and execution and storage eventually converge on the same root of truth confirmed by consensus. Even so, only verification has to be global. Ideally, computation itself should be as local as possible.
The essay then turns to data and execution more specifically. Storage and data cannot be eliminated and must exist globally in some form, but they can be compressed into cryptographic proofs and included in consensus without raw data being fully replicated across the network. Data availability can also be generalized through economic metering, temporary data models, and network sharding. In that setup, the network can reach global consensus on data availability while still keeping access to the data itself private.
The same principle applies to execution. Only operations whose execution paths intersect at the same address need to be computed inside the same local view. Disjoint paths and subpaths can run in parallel within their own execution scopes, coordinated by network observers as needed. The article describes this as a way to distribute aggregated execution and data throughput more evenly and to generalize execution and data availability at the architecture level. Under that model, throughput inclusion, ordering, and privacy protection are handled by multi-threaded network coordination rather than every node replaying one single-threaded execution path.
The limits of cryptographic truth
The article's most forceful argument comes next. The source of correctness inside Ethereum, it says, is cultural consensus around the protocol's definition, a definition of what the network can be. But that is still not enough to define what the network actually is in reality.
The real state of the network since genesis has to be derived from information outside the system. That information records the historical actions of observers and how the majority understood the order of events. Given an instantaneous state and a set of operations, one can tell whether those operations produce a new state that fits the protocol definition. What cannot be known from that alone is whether those operations, and their ordering, would be recognized by the broader set of Ethereum observers. That is the consensus problem in the article's formulation.
Ethereum's spatial dimension, the essay says, rests on cryptography and protocol components that let state-machine actions be checked statically. But whether those actions happened, and when, depends on the subjective perception of network observers. Ethereum therefore has to define who its observers are in order to measure majority opinion and interpret the events that make up instantaneous state.
What makes Ethereum unusual, in the article's telling, is that the definition of the observer is culturally blurry, and a large amount of information was lost into the strong subjectivity introduced at genesis.
The essay contrasts this with Bitcoin. At genesis, Bitcoin had only a static protocol definition, and its genesis block reward could not even be spent. The protocol did not hard-code external conditions into the system at inception. Even if Satoshi Nakamoto pre-mined coins, the definition of the observer under proof of work remained culturally homogeneous because it came from energy spent outside the system.
Ethereum's pre-mine, the article argues, effectively introduced a rule: the real definition and instantaneous state of the modern Beacon Chain are shaped in part by observers embedded at genesis, namely those staking ETH.
The logic is laid out plainly. Strong subjectivity and the definition of the observer must come from outside the system. Consortium chains and centralized systems hard-code operator roles into the protocol and subjectively predefine observers. Systems such as Bitcoin only predefine the protocol and then rely on an external, more general culture to define observers and the majority view. In the article's example, Bitcoin defines the network's real state as the chain confirmed by the most expended hashpower, making the observer more culturally homogeneous and more abstract.
Ethereum, the article notes, also used proof of work before 2022 and then switched to proof of stake, using the scarcity of ETH inside Ethereum to build a general culture around the observer. But the essay argues that proof of stake uses the system's own token as the observer's representative credential, which means the token and the observer collapse into one another. On that basis, it says the subjectivity problem is not actually resolved.
For networks that pre-mined stakeable tokens at genesis, the article sees a similar issue to the one found in centralized systems: the observer is defined before genesis and written into the protocol, preventing a truly universal culture from emerging.
It gives a numerical example for Ethereum today. Out of 122 million ETH, 72 million came from the pre-mine and 50 million came from proof-of-work mining. From that, the essay concludes that majority perception under proof of stake is skewed because some actors were granted observer status from the outset. The 50 million ETH that originated through proof-of-work mining can, in this framework, be seen as carrying forward an ontology built from external energy expenditure.
How the Beacon Chain defines canonical state
The essay says instantaneous protocol state comes from an observer ontology that defines majority perception. Inside Ethereum, the Beacon Chain serves as the component that grounds the derivation of the network's immediate reality. The article breaks that process into four steps:
- Genesis: begin from a genesis state with preconfigured stake distribution and initial network and virtual-machine state. The article says Ethereum's 2022 proof-of-stake genesis inherited the state of the original proof-of-work network from 2016 while also setting up a validator registry, with validators gaining voting power by staking tokens.
- History: trace all votes cast by staked observers on network state, reconstruct how observer perception formed over time, and then accept user actions in the order recognized by the network and agreed by the majority.
- Current state: arrive at the latest state accepted by the majority and share a canonical view of the virtual machine and validator registry with other observers.
- State transition: from the tip of canonical history, vote with other observers to determine the next canonical state based on live user actions in the virtual machine.
The article adds that a subjective genesis can also be replaced by a recent checkpoint that an observer personally accepts as valid. The definition of the network, and even of genesis itself, is tied to strong subjectivity. Accepting a recent statement such as "I consider this historical checkpoint valid" under the assumptions built into the protocol is what the essay identifies as weak subjectivity.
Trust assumptions in proof of stake
The trust assumption for the protocol is stated clearly in the article: a supermajority of observers by stake, 2/3, follows the protocol's definition of honesty and remains honest, including in cases of unintentional dishonesty. If that condition holds at any moment, the network can converge on a canonical current state.
If it does not hold, an outside observer cannot distinguish between a current state that came from a legitimate canonical history and one that came from a malicious stake supermajority rewriting history. In other words, whatever state an observer sees, they can only verify that it is cryptographically possible. They cannot prove that the history leading to it reflects the view of an honest majority.
The article goes further and says the problem of cryptographic validity in history is even more subtle. The assumption that a staking majority is honest applies not just to chain structure but to every event in history. Whether historical events are correct is meaningful only within the subjective frame of an honest majority. If a staking majority accepts a historically invalid event at the cryptographic level, then the majority is no longer honest, the assumption fails, and validity itself becomes blurry.
That is why, in the article's view, everyone is always assuming the existence of an honest majority. Since that assumption underpins every other claim of validity, there is little practical value in cryptographically checking history while refusing to trust the majority.
The DAO fork is used as the concrete example. If canonical history includes an event that your software definition treats as invalid, the article says you cannot tell whether you are looking at a malicious state rewrite or a coordinated hard fork that your old software simply does not understand. The DAO fork fits that pattern. A majority accepted an event that was cryptographically invalid under the old rules and wrote it into canonical history. Under the old protocol definition it was invalid, but at the subjective level the majority recognized it as part of main-chain history.
The essay's conclusion from this is that cryptographic verification has inherent limits. Consensus can help an honest majority coordinate under the protocol's own definitions of validity and honesty, but honesty itself remains subjective and is, in a paradoxical way, defined by the truth recognized by the majority. That is why the definition of Ethereum contains an internal subjectivity. Truth still has to be interpreted by the observer. Even if one person rejects the network majority's view, no one can prove that person's definition of Ethereum is less correct than someone else's.
Back to engineering and participation
In closing, the article condenses the whole argument into an engineering picture. Culture reaches agreement at the protocol-definition layer and becomes the initial subjective axiom of truth, defining what counts as correct inside the protocol's rules. On top of that foundation, a general-purpose virtual-machine architecture can be built. Then a majority-representation mechanism grounded in external observers generates consensus and makes Ethereum a globally shared computational reality.
At the level of network operation, the summary is straightforward: users broadcast general-purpose operations to a shared virtual machine, the network's computers order those operations and vote on their sequencing, and the system arrives at Ethereum's instantaneous state.
The article also says culture is messy and the revolution is imperfect, while the path to advancing Ethereum's vision is not singular. The current arrangement maintained through ACD is only one narrow way of defining Ethereum. People can propose other views and even change how the broader public understands the world computer.
The essay ends with an invitation. The door, it says, remains open to anyone who wants to join. If you believe in freedom, then welcome to the future built on Ethereum.