Last Tuesday a research report landed in my inbox. Copenhagen was doing its late-autumn thing โ gray light, wet cobblestones, the kind of cold that makes you grateful for a radiator and a reason to stay inside. The report ran fourteen pages. It carried eight analytical sections: technology, tokenomics, market structure, ecosystem position, regulatory compliance, team and governance, risk profile, narrative. It had a risk matrix with severity, probability, impact, and mitigation columns. It had a five-star scoring system. It closed with the standard disclaimer about the extreme volatility of crypto assets and the importance of doing your own research.
Every field said N/A.
Not 'unknown, pending data.' Not 'insufficient evidence, flagged for follow-up.' Just N/A, repeated with the serene regularity of a metronome, page after page, a cathedral of scaffolding raised around an emptiness. I read it twice and poured a second coffee. By the time the streetlights came on I had arrived at a conclusion I did not expect: this blank report was more honest than roughly ninety percent of the crypto research I read last year.
I want to be careful, because that sentence is easy to turn into a cheap provocation. It is not. It is a technical observation about where the crypto industry actually spends its intelligence, and where it has quietly stopped looking.
I have been watching this industry for nineteen years and running an education platform inside it for most of the last decade. I have audited liquidity mechanisms, interviewed policy drafters in Brussels, sat in a Nรธrrebro coffee shop with first-time investors trying to understand why their savings had evaporated. I have never once seen a portfolio destroyed by an N/A. I have seen hundreds destroyed by confident, well-formatted, perfectly paginated nonsense. Behind every hash, a heartbeat โ and behind too many research reports, nothing at all.
So let me use this empty document as a lens. Not to mock it. To understand what it is telling us.
The scaffolding economy
Somewhere in the last research cycle, crypto discovered that frameworks are cheaper than insight. A scorecard can be templated once and filled in forever. A risk matrix looks like rigor from a distance. A five-star rating imports the emotional authority of a consumer review into a domain where the stakes are entire life savings.
This is not unique to crypto, but crypto has a particular vulnerability to it, because so much of the underlying reality is genuinely hard to observe โ and because the audience has been trained to accept the aesthetic of analysis as a substitute for analysis.
I noticed it first in the layer-2 space, where I spend most of my technical time. Through 2023 and 2024 the rollup narrative became increasingly polished: dashboards, throughput charts, ecosystem health scores. Underneath all of it sat questions the dashboards could not answer โ who actually controls the sequencer, what happens during a reorg, how cost behaves when the shared data layer saturates. A beautiful interface is not a decentralized system. A completed framework is not knowledge. Both are optimized for the moment of presentation, not the moment of truth.
Where information is actually born
Any credible claim about a crypto network passes through three layers, and almost all the industry's investment goes into only one of them.
The first is capture: raw data pulled from nodes, indexers, RPC endpoints, explorers, governance forums, and โ increasingly โ from the human beings who use the thing. The second is parsing and verification: reconciling that raw data against the actual contracts, the actual token contracts, the actual unlock schedules, the actual proposal text. The third is judgment: what it means, what it portends, what a rational actor should do.
The crypto research complex has industrialized layer three. It has built thousands of templates, rating systems, and conviction frameworks for the act of judgment. It has barely invested in layers one and two, and it treats their output as given.
Consider a single number: circulating supply. It appears in every report ever written. It is almost never a fact. It is a claim made by a team, scraped by an aggregator, and repeated by a researcher who never opened the vesting contract. The contract itself has an admin key. The key may be held by a three-of-five multisig whose signers are anonymous. The report says 'circulating supply: 340 million.' The truth is a graph of claims with a custody chain hanging off it. N/A is what you get when someone finally tries to trace that chain and fails.
On-chain data has its own ways of lying. Reorganizations rewrite recent history. Indexers lag by minutes to hours, so a live dashboard is often a memory of a moment that no longer exists. Bridge accounting double-counts the same asset on both sides, and the same liquidity can be summed onto three chains at once by anyone naive enough to add it up โ and in my experience, plenty of people are. Total value locked is not a number. It is a procedure.
So when a report arrives with every field marked N/A, what has happened is that the pipeline โ famously the least examined part of this industry โ has failed visibly for once. The N/A is not an absence of analysis. It is the honest output of a broken chain of custody. That is, in its own strange way, a gift.
What my own failures taught me
I learned this the hard way, and not from a report. In 2020, during DeFi Summer, I worked with three independent developers to pull apart Uniswap V2's liquidity mechanics. We were not writing a scorecard. We were reading raw transaction data, and what we found was that gas fluctuations were hitting low-income users disproportionately hard โ a pattern no dashboard displayed, because dashboards show averages. The average user was fine. The median user was being quietly taxed. We published fifteen interactive pieces about it; fifty thousand people read them. That insight existed only because we went beneath the layer where the templates live.
Earlier, in 2017, I ran a grassroots education project in Copenhagen called Ethos Ledger, funded by forty-five thousand euros of community micro-donations. I personally interviewed one hundred and twenty first-time investors who had lost money to rug pulls. The finding that has stayed with me longest is not technical. Almost none of them were undone by a lack of technical literacy. They were undone by a lack of emotional resilience โ and by reports, threads, and posts that gave them the feeling of understanding without the substance. A filled-in document is a powerful thing to hand a frightened person. A blank one is nearly impossible to sell.
That is the incentive structure. It explains why this industry produces so many filled-in documents and so few honest blanks.
Three kinds of theater, one denominator
I keep meeting the same pattern across the sectors I follow, and once you see it you cannot unsee it.
Proof of reserves. Most exercises attest to a subset of liabilities at a single point in time, with no continuous auditing and no consolidated balance sheet. It is a photograph presented as a livestream. The format is convincing. The coverage is partial.
Real-world assets. For years the pitch has been that tokenization will bring trillions of dollars of traditional finance onto public chains. I have sat in those rooms; I have run workshops for Nordic bank employees. The institutions are genuinely interested in the technology. They are not, in most cases, interested in your public chain. They want permissioned environments, familiar legal wrappers, and audit trails their supervisors already recognize. The story is told on layer three. The actual institutional demand is happening somewhere else entirely.
Data availability and layer-2 economics. The entire research apparatus above rollups is calibrated to a cost structure that is about to move. Blob space is a shared, finite resource. When it saturates โ my working assumption is within roughly two years โ the fee environment that every ecosystem-health metric was built against will shift, and gas costs will climb again in ways no template modeled. The frameworks will not warn you. The fee will.
Three sectors, one denominator. In every case, the confidence is generated by the packaging, not by the evidence. Code is law, but empathy is truth โ and neither of them lives in a scorecard.
The regulatory layer has the same blind spot
In 2022, when my portfolio was down seventy percent and I was rebuilding, I co-founded a non-profit focused on regulatory education. I spent six months reading the EU's MiCA draft and interviewing forty policymakers and developers. What I found is that regulation has a layer problem too. The rules are written at layer three โ disclosure, fair marketing, suitability โ while the failures happen in the plumbing at layers one and two, where nobody drafts paragraphs. A well-written disclosure requirement applied to a broken data pipeline produces beautifully compliant reports about nothing. Our N/A document would have sailed through a MiCA-style review. That is the point, and it is not a comfortable one.
The contrarian read: stop romanticizing the blank
Here is where I push against my own argument, because the comfortable version of this essay ends with 'and that is why we must demand transparency.' Too easy.
The real problem is not that some reports are empty. The real problem is that the industry is drowning in full ones. A blank report is rare. A confident report built on a broken pipeline is ubiquitous, and far more dangerous, because it converts a data failure into a belief. The blank tells you nothing and you know it tells you nothing. The full one tells you nothing and you act on it anyway. Between the two, the blank is not the scandal. It is the immune response.
But the romanticizers miss something. The blank report is still not good. It still consumed fourteen pages and someone's salaried hours. It still entered a decision process. An honest failure is only valuable if it triggers a repair; otherwise it is failure with better manners. Surviving the winter to plant the spring means planting something โ not just describing the frozen ground beautifully.
And there is a harder truth underneath. A blank report is honest, but it is also safe. N/A protects the analyst from being wrong. In a market that rewards conviction, the person who writes 'insufficient data' is never quoted, never promoted, never invited back to the conference stage. The incentive gradient points away from honesty at every turn.
The agent complication
Now add the thing I actually spend my days on in 2026.
I am leading a pilot in which AI agents run micro-education campaigns for new adopters, coordinated through a DAO. I have been writing about the convergence of AI and crypto under the banner of what I call the Cognitive Commons. And I can tell you with some confidence that the N/A problem is about to get much stranger.
When humans fill in a template, the failure mode is laziness or dishonesty. When agents fill in a template, the failure mode is fluency. An agent does not feel the friction of not knowing. It experiences a prompt. If the pipeline behind it is broken, it will still produce fourteen polished pages, because producing polished pages is what it is for. The blank report we received was a human artifact โ a pipeline with a human at the end of it, failing visibly. Automate the layer above the failure and the failure becomes invisible.
This is not an argument against agents. It is an argument for pushing verification into the layer beneath them, and for treating provenance โ which block, which contract, which oracle, which timestamp โ as the primary product of research rather than an appendix. Trust no one, verify everyone, feel everyone. The feeling part matters more than ever, because an agent cannot feel the wrongness of a number. It can only format it.
What I am watching instead
I do not think the answer is more frameworks. I think it is fewer, deeper, slower โ and I think the market is closer to rewarding that than it looks. The last two years have been a long sideways grind, and sideways markets are where positioning happens. They are also where the gap between a real signal and a pretty chart becomes expensive to ignore, because there is no rising tide left to hide the mistakes.
So here is what I watch, and what I would ask you to watch with me. Not the star ratings. The provenance. Not the narrative section. The raw capture layer. Not the confident ten-thousand-word thesis. The report that says: we could not verify this, and here is exactly why, and here is the pipeline failure we found when we went looking.
If you want one forward-looking question to carry out of this, take this one. When the agents start writing the research, who will be left honest enough to write N/A?
Philosophy before protocol, people before profit. The ledger remembers, but the heart forgives. The pipeline, unfortunately, does neither.