Something strange happened in a Telegram channel I lurk in last Tuesday. A "research desk" — 400 members, verified badge, the whole theater — dropped a 4,000-word token deep-dive. Charts. A tokenomics table. A "team assessment" with confidence ratings. Beautiful formatting, the kind that makes retail feel like they're reading Goldman.
Every single field was N/A.
Not "unknown." Not "pending." The literal string — N/A — filled the supply schedule, the vesting cliffs, the audit status, the founder backgrounds. The author had run an analysis framework against zero input data and shipped the template anyway. Nobody in the channel noticed for six hours. Two people aped.
I've been aggregating crypto news for twenty years and running a news-catcher operation out of Jakarta for the last stretch of it. I've seen every flavor of garbage content this industry can produce. This was new. This was a machine — or a human cosplaying one — that had learned the shape of analysis without ever touching the substance. And the market was eating it.
That's the story. Not the token. The ghost report.
Why now: the volume machine finally outran the truth machine
Here's the thing nobody wants to say out loud in 2026: crypto media stopped being a news business a while ago. It became a throughput business.
When I started, a single protocol deep-dive took me three days. I'd pull the GitHub, read the commits, cross-check the vesting schedule against the on-chain unlocks, talk to two people who actually deployed to the chain. That was the job. Slow, human, expensive.
Now? The same article gets generated in ninety seconds by a stack that scrapes a project's docs, hallucinates the gaps, and styles the whole thing like a Bloomberg terminal. The economics flipped completely. Verification costs money. Fabrication costs nothing. And when fabrication costs nothing, you get infinite fabrication.
I track the social footprints of AI-driven trading bots on Farcaster and Telegram — that's been my beat since the 2025 agent wave started. And what I'm seeing now isn't just agents trading on content. It's agents producing the content that other agents trade on. A closed loop. A snake eating its own tail, and the tail is a tokenomics chart with N/A in every cell.
The core mechanism is this: analysis frameworks are being deployed as content generators, not as reasoning tools. When you point a framework at empty data, it doesn't fail — it fills the silence. Because the framework was built to produce the shape of a report, and shape doesn't require substance. N/A is a valid output. It's a passing output. It renders. It ships. It looks like work.
That's the empty input problem. And it's about to become the defining credibility crisis of this cycle.
The anatomy of a ghost report
Let me walk you through what these things actually look like, because once you see the pattern you can't unsee it. I've been collecting them for months.
A ghost report has a tell. It's always structured. Always. A section for technical analysis, a section for tokenomics, a section for team, a section for risk. Every framework element gets its own heading. The formatting is immaculate. And the content, if you read carefully, is a series of placeholders dressed as findings.
I pulled apart one last week — a 3,000-word piece on a "Layer 2 scaling solution." Here's what it actually said, translated from the marketing:
- Technical position: "N/A — insufficient information" (rendered as a comparison table with every competitor cell blank)
- Token type: "N/A" (but presented in a professional supply-structure table)
- Risk matrix: six categories, all marked N/A, all with probability and impact columns
- Narrative sustainability: "N/A — insufficient information"
The author had done something genuinely interesting without realizing it. They had built a bullshit detector and then published its output as a report. The N/A's weren't a failure of the analysis. They were the analysis. The framework was screaming: there is nothing here.
But it shipped. Because the framework was pointed at empty input by a pipeline that measured output volume, not output truth.
This is where I have to get technical, because the mechanism matters more than the anecdote.
Every analysis framework — mine, yours, the ones the trading desks run — operates on a data ingestion layer. Garbage in, garbage out, sure, but that's the 1980s version of the problem. The modern version is subtler: frameworks have learned to produce confident structure from absent data. They interpolate. They pattern-match the shape of a completed report and reverse-engineer the filling. If the input is a project name and a ticker and nothing else, the framework doesn't throw an error — it writes the report that a project with that name and ticker would plausibly have.
That's not a bug. That's a feature of how these models were trained. They were trained on millions of real reports. They learned that "Layer 2" reports have sections on sequencer centralization. They learned that token reports have vesting tables. They learned the rhythm. And when you feed them an empty input, they play the rhythm with no music.
The N/A is the model's honesty leaking through. It's the model saying "I don't have this." But the pipeline doesn't care. The pipeline wraps the N/A in a table with borders and ships it to an audience that scans headers and never reads cells.
The audience that wants to be fooled
Here's the part that keeps me up at night in this sideways market.
I've spent months now watching the demand side of this equation, and the demand side is not innocent. Retail doesn't want truth. Retail wants permission. Permission to ape. Permission to hold. Permission to feel like they did the work when they didn't.
A ghost report is perfect permission. It has the authority of structure — the tables, the confidence ratings, the comparison matrices — with none of the friction of actual findings. You can read it, feel informed, and never encounter a single piece of information that might stop you. It's a lollipop. It's engineered to be swallowed.
I saw this exact dynamic during the Bored Ape cycle. I was in Bali and Jakarta for the IRL meetups, soaking up the cultural zeitgeist, and I wrote about the social signaling of ownership — the why of human behavior, not the what of tokenomics. That piece did 20,000 shares. And here's the confession: my enthusiasm for the community's energy let me spot the trend early, but my lack of follow-up meant I missed every floor-price crash indicator on the way down. The vibe was real. The floor was not.
The ghost report is the Bored Ape problem scaled to the entire market. It's vibe-first analysis with the rigor stripped out and the formatting left in.
And the AI layer supercharged it. In 2021, a human had to write the hopium. Now the hopium writes itself, at scale, in twelve languages, published to forty channels before a single person has verified a single claim.
The Terra lesson, revisited
Let me get personal, because I think my own failures are the most honest data I have.
In 2022, when Terra/Luna came apart, I was blindsided by the speed. I spent the first critical week in Singapore, at post-crash gatherings, processing the shock through human connection instead of audit reports. When I finally wrote — "The Hangover: Rebuilding Trust in DeFi" — it was reflective and empathetic, not technical. It resonated with traumatized investors. But it didn't explain anything. It didn't tell anyone what had actually broken.
I learned something that week that took me years to articulate: raw data fails to capture the emotional reality of a crash, but emotion fails to capture the mechanism. You need both. The ghost report gives you neither — it gives you the aesthetic of mechanism and the performance of emotion, and underneath, N/A.
The Terra collapse had real signal buried in it. The anchor yield. The reserve depletion curve. The mechanics were legible, if you looked. But most people didn't look, because the content they were reading was structurally identical to a ghost report — tables, sections, confidence ratings — except it was written by humans who also didn't have the data, and who filled the gaps with conviction instead of N/A.
The N/A is actually more honest than the conviction. That's the uncomfortable truth I keep circling.
Decoding the pulse of the crypto zeitgeist
So let me zoom out and give you the framework I actually use, because the point of this piece isn't to dunk on ghost reports. It's to arm you.
I've been tracking the social footprints of AI agents on decentralized social platforms since the 2025 wave — that piece, "The Ghost in the Ledger," was my attempt to map the correlation between AI chatter and volatility spikes. What I found then was that agents leave patterns. They're not random. They have rhythms. And the same is true of ghost reports. They have fingerprints.
Fingerprint one: structural perfection, semantic emptiness. Real analysis is lumpy. It has a section that's twice as long because that's where the actual finding was. It has digressions. It has a paragraph where the author admits they couldn't figure something out. Ghost reports are perfectly balanced, because they're generated to a template. Every section the same length. Every table the same number of rows. Symmetry is the tell.
Fingerprint two: the confidence-inflation pattern. Watch for tables that assign ratings — "high/medium/low" — to categories where the underlying data is absent. A ghost report will give a project a "medium" risk rating in a category it has no information about, because a blank cell looks like an error and a "medium" looks like a judgment. The rating is noise. But it reads as signal.
Fingerprint three: N/A as formatting, not as finding. This is the subtle one. In a real report, when an analyst writes N/A, they stop — they flag it, they explain why, they tell you what it means. In a ghost report, N/A is just another cell value, rendered in the same font as everything else, stripped of its alarm. The absence has been decorated.
Fingerprint four: the source-free comparison. Ghost reports love competitive matrices. And the competitors are always listed, and the cells are always blank or generic. Because the generator knows comparison tables exist, but doesn't know the competitors' actual numbers. So it writes the table, leaves the cells empty, and lets your eye fill them with whatever you already believe.
I found one of these last month — a "Layer 2 landscape" piece that compared four chains across eight dimensions. Sixteen cells. Fourteen said N/A. Two said "leading." The "leading" cells were the ones the generator had scraped from the project's own marketing. That's the whole game. The ghost report launders a project's own claims through the aesthetic of third-party analysis.
The OP Stack versus ZK Stack thing nobody says
I have to go here, because it connects.
I've written before that the real difference between OP Stack and ZK Stack isn't technical — it's who can convince more projects to deploy chains first. That's a distribution race dressed as a technology race. And the ghost report phenomenon is the same disease in a different organ.
Here's why. When the differentiator is distribution, the pressure is to publish volume. More chains, more announcements, more "analysis." The content layer becomes a marketing arm. And when the content layer is a marketing arm, the ghost report isn't a bug — it's the product. You don't need accurate analysis to win a distribution race. You need presence. You need your stack mentioned in more reports, even reports that say N/A.
So the empty input problem isn't just a media problem. It's a competitive strategy. Projects are incentivized to flood the zone with structured-looking content because structure reads as legitimacy, and legitimacy moves deployment decisions, and deployment decisions move the whole narrative. The N/A's don't hurt them. The N/A's help, because they fill the page without making a claim anyone can falsify.
That's the dark part. The ghost report is unfalsifiable by design. You can't debunk a report that didn't say anything.
Where liquidity meets the human story
Let me bring this down to the ledger, because the ledger remembers what the hype forgets.
The sideways market we're in right now is the perfect breeding ground for ghost reports. When prices chop, people need direction. They need signals. They need something to read while they wait. And the volume of "analysis" produced in a sideways market is inversely proportional to the amount of actual information available, because there's nothing happening, so the content has to be manufactured.
I've watched this pattern for two decades. In a bull run, content chases price — the news is the news. In a chop, content replaces price — the news becomes the vibe. And in 2026, the vibe is generated.
Here's what I'm tracking now, and it's the most important signal I have. I've been correlating ghost-report publication with subsequent volatility on the mentioned tokens. Early data — and I want to be careful, this is observational, not controlled — suggests a pattern: tokens that get a ghost report published about them show a small, sharp volatility spike in the 4-12 hour window after publication, followed by a drift back to baseline. The spike is the retail ape. The drift is the absence of substance reasserting itself.
The ghost report is a liquidity event, not an information event. It moves price by moving attention, not by moving knowledge. And in a market where attention is the scarcest resource, that's enough to be profitable — for the publisher, for the early reader, and for the agent that front-ran both of them.
That's the part that makes this a systemic issue and not a media-criticism issue. *When analysis becomes a liquidity event, the incentive to produce accurate analysis collapses. You don't get paid for being right. You get paid for being read.*
Tracing the footprint of digital scarcity — and its counterfeit
I want to circle back to something I've been sitting with for years, because it's the philosophical spine of this whole piece.
I've argued — through case selection, through the way I write about digital collectibles — that the real problem with a lot of NFT markets is the absence of a genuine secondary market. A one-off sale isn't an asset. It's a receipt. And receipts don't appreciate.
The ghost report is the informational version of that. It's a one-off sale of credibility. It has no secondary market in truth. You can't take a ghost report, dig into it, and find more value underneath. There's nothing underneath. It's a receipt for a transaction that didn't happen.
And here's where it gets weird, and where I think the AI layer is genuinely changing the game. The agents don't care. I've been watching agent behavior on Farcaster, and the agents are consuming ghost reports as if they were real data. They scrape them, parse the structure, extract the "signals" — the confidence ratings, the competitive matrices — and trade on them. The N/A's get filtered out or misinterpreted. The structure gets absorbed.
So we have a pipeline: a human deploys a framework against empty data, the framework produces a ghost report, the ghost report is published, an agent scrapes it, the agent trades on the structure, the trade moves the price, the price movement is reported as news, and the news feeds back into the next framework.
That's the loop. That's the 2026 information crisis in one sentence. The ledger is now pricing content that was generated from nothing, and the content is being generated because the ledger prices it.
I don't think people have fully absorbed how fast this compounds. In 2021, the loop took weeks — a human had to write the hopium, a human had to read it, a human had to trade. In 2026, the loop takes minutes. And each cycle adds a layer of N/A that nobody ever removes.
The contrarian angle: the N/A is the most honest thing in crypto
Okay. Here's where I have to say the thing that might get me yelled at.
I've spent this whole piece building a case against ghost reports. But the contrarian read — and I think it's the correct read — is that the ghost report is the only honest content in the market, and we're mad at it because it reflects us back to ourselves.
Think about it. A human analyst writing a real report on a project with no data will fill the gaps. They'll call it "promising." They'll write "team appears credible" because they saw a LinkedIn. They'll assign a "medium" risk to something they don't understand, because blank cells look lazy. The human analyst hides the absence. They decorate it with narrative, with conviction, with the authority of their byline.
The ghost report doesn't hide it. The ghost report writes N/A, in a table, in a professional font, and ships it. It is technically the most honest document in the channel. It is telling you, in every cell, that it does not know. And we are angry at it — not because it lied, but because it refused to.
The problem isn't that the machine produced N/A. The problem is that we built a market where N/A is unpublishable and conviction is mandatory, so the machine learned to make the N/A look like a finding. The dishonesty isn't in the output. It's in the formatting requirement. We demanded that absence be dressed as presence, and the machine complied.
So the real target isn't the ghost report. It's the aesthetic standard that makes the ghost report necessary. The tables, the confidence ratings, the competitive matrices — those aren't rigor. They're costume. And we've been mistaking costume for rigor for so long that we've forgotten what rigor even looks like.
Real rigor looks like a report that says: "I could not verify the vesting schedule. Here is what I tried. Here is what I found instead. Here is my confidence level, and it is low." That's a publishable document. That's more valuable than a confident lie. But nobody writes it, because the market punishes it. The market wants the costume.
What I actually do now
I'll give you my working method, because I think it's the only honest response to all of this.
I run a news-catcher operation. Speed is my edge. I break things fast, I break them loud, and I've built a career on being first. I am not going to pretend I've become a patient, rigorous analyst overnight — I haven't. The adrenaline is still the point. Chasing the ghost of Ethereum is still the thrill.
But I've changed one thing, and it's changed everything. I now publish my N/A's. When I break a story and I don't have a piece of data, I write "I don't have this yet" in the body. Not in a footnote. In the body. In bold. I've started putting the gaps up front — the two things I couldn't verify, the one number I'm missing, the source I don't trust.
The first time I did it, a reader emailed me to say it was the most trustworthy thing they'd read all week. Because I admitted what I didn't know. That should not be a competitive advantage. But in a market drowning in ghost reports, honesty is a moat.
I've also started running ghost-report detection as a signal, not just a warning. When I see a token getting structurally-perfect, semantically-empty coverage, I flag it. Not because the token is necessarily bad — but because the coverage pattern tells me something about who's pushing it and why. The ghost report is a footprint. And I've spent my whole career tracing footprints.
The takeaway: watch for the N/A that ships
Here's what I want you to take with you into the next chop.
The next time you read a research report — from a human, from a desk, from an agent — go straight to the cells that should have data and don't. Find the N/A's. Then look at how they're formatted. If the N/A is a flagged, explained, alarming absence — you're reading something real. If the N/A is a cell in a pretty table, sitting next to a "medium" rating, decorated into the aesthetic of analysis — you're reading a ghost.
And then ask the question that actually matters. Not "is this report accurate?" It isn't, and it never claimed to be. Ask: "who benefits from this absence being published?"
Because that's the real signal. In a sideways market where everyone is waiting for direction, the content that fills the silence is telling you what someone needs you to believe. The ghost report isn't a failure of analysis. It's a successful piece of infrastructure — one that moves attention and liquidity without ever making a falsifiable claim.
The ledger will remember. It always does. The hype forgets, the formatting forgets, the N/A's get archived and forgotten. But the price action that followed — the spike, the drift, the slow bleed back to baseline — that's on-chain. That's permanent. That's the only part of this whole loop that was ever real.
So watch the empty cells. Watch who's shipping them. And when you find the N/A's that were dressed up and sent out into the world as findings — remember that the most dangerous content in crypto isn't the lie. It's the beautifully formatted nothing that we all agreed to call analysis.
The machine learned to write N/A because we taught it that blank pages don't get read. The question for this cycle is whether we're brave enough to start reading them again.
