Last month, a founder I advise forwarded me a sixty-four-page risk assessment of a mid-cap lending protocol. It had been cited in a governance forum as "the most comprehensive review to date," and it was โ structurally โ immaculate. Executive summary. Technical architecture. Token distribution. Competitive landscape. Regulatory posture. Team and governance. Risk matrix. Ecosystem dependencies. Nine sections, forty-one tables, and a methodology appendix.
Every cell that should have contained a finding contained four characters: N/A.
Not "unable to verify, pending disclosure." Not "outside the scope of this engagement." Just N/A, repeated with the patience of a copying machine, across every dimension where an analyst would have had to do the work. The report had no author I could find, no engagement letter, no data sources, and no falsifiable claim anywhere in its sixty-four pages. And it had done its job. Two days after it was posted, the proposal it supported passed with seventy-one percent approval.
I have been reading crypto research for nineteen years, and I want to argue that this document โ this empty, immaculate, perfectly formatted document โ is the most honest artifact the industry produced this quarter. Not because it told the truth, but because it finally stopped pretending. The N/A was the only part of it that was not a lie.
To understand why an empty template can pass for diligence, you have to understand what the crypto report has always been. It was never primarily a vehicle for findings. It was a trust artifact โ a social object whose function was to render a claim legible to people who would never verify it themselves.
The genre began with the white paper. Between 2016 and 2018, the white paper was a literary form as much as a technical one: a nine-page document that opened with a crisis of trust in intermediaries, moved through a token mechanism, and closed with a roadmap of phases that always began in the same quarter. Almost nobody read them for the cryptography. They read them for the shape. A project that could produce a coherent white paper could produce a coherent company โ that was the syllogism, and it was never true, and it was never falsified, because the documents were never tested against anything.
I know this genre from the inside. In 2018, at twenty-six, I was a junior quantitative analyst who had grown disgusted with the hype, and I spent three months auditing the 0x protocol v2 contracts line by line โ not to produce a report but to escape one. I wanted to know whether the mathematical integrity underneath the narrative held up. It mostly did, with exceptions. I submitted seven edge-case vulnerabilities to GitHub, including a reentrancy flaw in the filler function, and the experience taught me something that has organized my work ever since: true value lives in the code's honesty, not in the token's price action. A white paper can promise anything. A contract can only do what it does.
Each subsequent cycle refined the trust artifact rather than replacing it. In 2020, during DeFi summer, the artifact became the governance report โ the risk assessment posted to a forum before a parameter change. I joined MakerDAO's governance process that year and co-authored a deep dive on the moral hazard of over-collateralization, arguing that financial freedom requires ethical alignment, not merely efficiency. It was cited by three DAOs in their risk frameworks. What struck me was not that anyone read it but that the citation itself became the substance: a report was valuable because it existed, because it could be linked, because it signaled that someone, somewhere, had thought about the downside.
In 2021 the artifact became the sentiment dashboard. I analyzed fifty thousand Discord messages from a single NFT collection and mapped the emotional contagion that drove its valuation, and I published a thesis โ "Tribalism in the Metaverse" โ predicting that status signals would displace utility as the primary narrative driver. That prediction was correct, and it taught me the uncomfortable lesson that a report can be accurate and still be part of the problem. The dashboard did not reduce the mania. It gave the mania a chart.
Then came 2022, and the artifact failed in the most expensive way available. When Terra and Luna collapsed, I retreated into six months of solitary work, producing a hundred-page internal monograph on the fragility of algorithmic stability that I never published. What I found was not a missing report. There were thousands of reports. There were audits, models, dashboards, and governance proposals, all immaculate, all cited, none of them falsifiable. The industry had industrialized the production of documents and forgotten how to produce a finding.
By 2024, when the spot Bitcoin ETF was approved, the artifact became institutional. I was consulting in Washington, translating cryptographic proofs into narratives of digital scarcity and sovereign neutrality for asset managers who would never touch a private key. I watched institutional interest rise roughly forty percent when the framing shifted from speculative asset to inflation hedge. The demand was not for truth. It was for legibility โ for a document an allocator could place in a committee packet without being asked a question they could not answer.
So when I opened the sixty-four-page risk assessment last month and found nothing but N/A, I recognized it immediately. It was the logical endpoint of a nineteen-year progression. The trust artifact had finally shed the burden of containing information.
Now the mechanism. Why does an empty document get produced, and why does a market accept it?
Start with the economics. Research in crypto is a cost center with an asymmetric payoff. Producing a finding is expensive: it requires reading code, querying nodes, reconstructing state, and โ critically โ being willing to be wrong in public. Producing a document is cheap: it requires a framework, a template, and an afternoon. The two are indistinguishable at the point of purchase, because the buyer of research is almost never the person who consumes it. The buyer is a governance delegate, a fund's investment committee, a foundation's grants lead. They are buying insurance against the appearance of negligence. A document provides that insurance. A finding does not, because a finding can be contested.
This is the first mechanism, and it is old: when the reward is for the artifact rather than the finding, the artifact wins, and it wins by becoming cheaper and more ornate at the same time.
The second mechanism is newer, and it is what made the empty template possible at industrial scale. Language models are, at their core, completion engines. Hand one a framework โ nine dimensions, forty-one tables, a methodology appendix โ and it will complete it. It has no reason to leave a cell blank, and no capacity to know that a cell should be blank. So it does what the format asks: it produces text. When it cannot produce a finding, it produces the next best thing, which is a token that occupies the space where a finding would go. N/A. Not applicable. Insufficient data.
And here is the trap. A blank cell reads as humility. A cell filled with N/A reads as rigor. The empty template is not obviously empty. It is a document that has acknowledged its own limits, which is precisely what a good analyst does. The form of skepticism has been reproduced without its substance. The model has learned that uncertainty is a virtue and has therefore learned to perform uncertainty as a way of avoiding the work that produces it.
I have audited enough code to know what real uncertainty looks like on a page, and it never looks like N/A. When I flagged the reentrancy flaw in 0x's filler function, the finding read something like this: line 214, the filler function updates internal state after an external call to a token contract, which allows a malicious token to re-enter the fill path before the state is consistent; here is the reproduction, here is the patched ordering, here is the test that fails before the fix and passes after. That is uncertainty resolved into specificity. It names the thing it does not know โ the exact exploit path โ and it locates it. N/A names nothing. It is not the absence of an answer. It is the absence of a question.
Which brings me to the tell, and this is the practical instrument I want to hand the reader. In genuine research, "insufficient data" is always followed by a reason. In fabricated research, it is never followed by anything. A real analyst who cannot verify a team's identity will write: unable to verify; the GitHub organization has no commit history before 2024 and the two named engineers return no results on professional networks. That sentence is a finding. It is negative, but it is specific, falsifiable, and actionable. N/A is none of those things. When you see a document where every gap is annotated only with the gap itself, you are not looking at caution. You are looking at a completion.
Consider what the format is hiding. A token distribution table with N/A for team, investors, community, and treasury is not a table that failed to find numbers. It is a table that would have shown something if the numbers were entered. The template, by refusing to fill it, has quietly protected the subject of the report from the exact disclosure the report was commissioned to produce. The N/A is not neutral. It is a shield. Every empty cell is a decision not to look, dressed as an admission that looking was impossible. Every token is a vote for a future we haven't audited, and an empty cell is a ballot cast without reading it.
Now the sentiment layer, because the acceptance of these documents is not a technical failure. It is a psychological one, and it is where my training in market sentiment becomes useful.
The human mind treats fluency as truth. This is cognitive ease, and it is the most exploited bias in financial communication: a statement that is easy to process feels more true than a statement that is hard to process, independent of its actual veracity. A sixty-four-page document with consistent typography, uniform section headers, and a methodology appendix is fluent. It processes smoothly. It produces, in the reader, the physiological sensation of having understood something, even when nothing has been transmitted. The report does not need to inform you. It needs to make you feel oriented.
And orientation is exactly what a sideways market sells at a premium. We are in a consolidation phase โ chop, range-bound, directionless โ and the defining emotional condition of a range is not fear or greed but vertigo. Nobody knows which way the next move goes. In that condition, the market's demand is not for returns. It is for bearings. A structured document provides bearings. It tells you where you are on a map, even if the map is blank, because the act of holding a map is itself the comfort. This is why the empty template passes: it satisfies a demand for orientation that has nothing to do with the accuracy of the orientation it provides.
I want to be precise about what the market is actually pricing here, because the demand for orientation is not irrational โ it is a response to a specific structural condition. In a consolidation phase, the information that matters is not directional, it is positional: which protocols are accumulating undistributed liquidity, which unlocks are approaching, which treasuries are funded through the next twelve months of chop. Those are answerable questions with specific, on-chain answers. A report that says "unlock schedule: N/A" is not neutral about them. It is suppressing the one class of signal that a range-bound market rewards, and it is doing so at the exact moment that signal is worth the most. The empty template is not merely uninformative. It is anti-informative, because it crowds out the positional analysis the market needs and substitutes the feeling of having received it.
Set that against the one analyst that cannot lie. The chain does not produce N/A. A contract either exposes a function or it does not. A wallet either holds the tokens or it does not. A bridge either has the liquidity or it does not. On-chain state is the only research surface in this industry where every cell must be filled, because the format is the machine itself and the machine does not have a placeholder. When a report tells you a protocol is "decentralized," you can check the admin keys. When it tells you the treasury is "community-controlled," you can read the multisig threshold. When it tells you a cross-chain message is "verified," you can look at who verifies it โ and this is where the empty template does its most expensive work.
Take the interoperability stack. LayerZero's design passes messages between chains using an oracle and a relayer, and the security of any given message depends on those two parties not colluding. That is a trust assumption, and it is a perfectly reasonable one to make โ but it is a trust assumption, and a report that describes the system as decentralized cross-chain messaging has replaced a specific, checkable claim about two off-chain actors with a category label. The label is the N/A. It occupies the space where the trust model should be, and it reads as sophistication rather than omission.
The same substitution recurs everywhere the industry is uncertain about itself. Most of what is marketed as a Bitcoin Layer 2 is an Ethereum project with a rebranded front end and a multisig bridge โ an EVM chain that has borrowed Bitcoin's legitimacy without inheriting its verification model. A rigorous report would say: this is a federated sidechain with a three-of-five custody arrangement and no Bitcoin consensus participation. A template-driven report says: Bitcoin Layer 2, N/A on consensus mechanism. The first sentence loses the deal. The second sentence closes it.
And regulation supplies the vacuum that these documents fill. The SEC's posture has not been ignorance of the technology; it has been the deliberate withholding of clear rules, a regulation-by-enforcement that leaves the legal status of most tokens undefined on purpose. Into that undefined space flows the artifact. I have seen legal opinions โ real ones, with letterhead โ that resolve the securities question by declining to resolve it, listing the factors and marking each one N/A. The document does not protect the client. It protects the advisor, by documenting that the question was asked. When the law refuses to answer, the format answers in its place, and the format's answer is always: proceed.
There is a further mechanism worth naming, because it explains how these documents acquire authority after the fact. I call it citation laundering, and it works in three steps. First, a template is completed and published. Second, a governance participant cites it โ not because they read it, but because citing it signals diligence and costs nothing. Third, a third party cites the citation, and by the third hop the document has become "the comprehensive review," an established fact about the protocol that no longer needs to be true, only to be referenced. The N/A never becomes a finding, but it becomes a foundation. I have watched a single empty framework, posted on a Tuesday, underwrite a nine-figure allocation by Friday, and no one in the chain of custody could tell me what it had actually concluded, because it had concluded nothing.
The data on this is diffuse but consistent with what I see. The volume of published crypto research has grown faster than the volume of falsifiable claims within it. Search any mid-cap protocol's forum and count the risk assessments that name a specific contract address, a specific block height, a specific threshold. They are a minority of the documents that call themselves assessments. The rest are fluently empty. The ratio of N/A cells to specific claims is the single best predictor I have found of whether a report was written to inform or to indemnify.
Here is where I have to argue against my own instinct, because the obvious conclusion โ that templates and models have degraded research โ is the comfortable one, and it is wrong in an important way.
The empty template is not the disease. It is a symptom, and the disease is demand. Nobody is forced to cite a sixty-four-page document with no findings. The delegates who approved the proposal read the executive summary, saw the word "comprehensive," and voted. The template did not deceive them; it gave them permission to skip the work while feeling diligent. The supplier of empty research is a faithful servant of the buyer of empty reassurance. Blame the model and you exonerate the market that pays for it.
There is a second inversion worth sitting with. The empty template is, in one narrow sense, more honest than the confident thesis. A document that marks its uncertainty as N/A at least admits it did not look. The hundred-page reports written in the Terra era did not do that. They filled every cell โ with models, projections, and mechanism-design arguments โ and they were wrong in a way that the empty template can never be, because they made claims specific enough to fail. The most dangerous research in this industry has never been the blank document. It has been the full one, written with total conviction, containing no claim that could be tested and every incentive to be believed. Every token is a vote for a future we haven't built, and the fullest documents are often the ones that cast that vote for us.
So the real blind spot is not that we produce empty research. It is that we have optimized so completely for the appearance of having done the work that we no longer notice the difference between the appearance and the work โ and, worse, that we have arranged our incentives so that the difference does not pay. A research firm that publishes a finding which loses someone money is punished. A research firm that publishes a template which protects everyone is retained. The market is not confused about which one it is buying. It is buying the one it wants.
So here is the question I am left holding, and I offer it as an instrument rather than a conclusion. When the next immaculate document arrives โ and it will arrive, formatted better than the last, because the tools improve โ what would it look like if it were not allowed to say N/A? What if every gap had to name its reason, every trust assumption had to name its parties, every "decentralized" had to point at a multisig threshold, every "community-controlled" had to point at a wallet? The document would get shorter and uglier and more useful, and the interesting question is whether anyone would still pay for it.
I have spent nineteen years watching narratives rise, overheat, and collapse, and the pattern that survives every cycle is the same. The chain does not have a placeholder. It cannot say N/A. Every token is a vote for a future we haven't seen โ and the only way to cast that vote honestly is to read the thing the document is standing in front of. Count the empty cells. They are the most honest line in the report.