Delusions Don't
Make Consensus

It would be convenient if BIP-110's support base broke cleanly along technical lines — people who understood the code opposing it, people who didn't understand it supporting it. That's not what happened. Plenty of BIP-110's supporters could read the diff, evaluate the script-depth restrictions, and argue the node-storage grievance on its technical merits. Some of them still hold that BIP-110 was the right call, or at least a defensible one, even after the stall. Technical fluency didn't inoculate anyone against the pattern this piece is actually about, and pretending otherwise would be a more comfortable story than the true one.

The true one is this: conviction can outrun its backing regardless of how much someone knows. Bitcoin was being treated, in this telling, as a dumping ground for arbitrary data; node operators were footing the storage bill forever for a fee miners collected once; a temporary rule change looked like a reasonable way to push back. That argument had real technical substance behind at least part of it. What it didn't have was 55% miner signaling, and somewhere between the argument's substance and its declared confidence, a gap opened up that no amount of code review closes. That gap — between what a position has earned and how certain its holders act — is the actual subject here, and Bitcoin coders are not exempt from it any more than anyone else is.

01 — A pattern with a two-hundred-year-old precedentJefferson, Monroe, and how conviction gets ahead of itself

Jefferson's Yeoman Farmer ideal ran on faith that an agrarian republic could out-argue the economic machinery Hamilton was actually building. Jefferson wasn't uninformed — he was one of the most well-read men of his era, fluent in finance, law, and political theory. He still spent decades certain that a nation of independent farmers could out-compete a nation built on banks and manufacturing, and the republic ended up running on the financial system he distrusted anyway. Monroe's Doctrine declared the entire Western Hemisphere off-limits to European powers with a navy too small to enforce a fraction of that claim. Monroe had been a soldier, a diplomat, a two-term president — someone who understood power about as well as anyone in the country. He still declared a policy the United States had no way to back up, and it only held because it happened to line up with what the British Royal Navy wanted anyway.

Neither man was unsophisticated. Both were about as expert in their domains as it was possible to be in 1800s America. And both still ended up declaring a conviction as if the backing for it already existed, when the backing was thin or borrowed from somewhere else entirely. That's the actual mechanism worth naming: expertise raises the quality of the argument someone can make, but it does nothing to guarantee that their certainty will track how much support the argument actually has. BIP-110's 55% threshold, met with 2.53% signaling from what turned out to be effectively one pool, is the same shape. Knowing Bitcoin's codebase in detail didn't stop some of its most technically fluent supporters from acting as though the numbers would eventually catch up to the conviction, rather than the other way around.

02 — What actually happenedThe numbers, as plainly as they can be put

The technical debate about script restrictions and data storage is genuinely contested and worth taking seriously on its own terms — that's not what this piece is relitigating. What follows is what's actually documented and verifiable about how the fork played out, because that part isn't in dispute by anyone, including BIP-110's own supporters.

The gap didn't stay static, either. In the hours after the split, the minority chain's next difficulty adjustment was estimated at roughly 350 days out, against 14 days for the main chain. By the time the dust settled a few days later, that estimate had stretched to 6.3 years — a number recalculated continuously from actual block times, meaning every additional idle hour pushed it further away rather than closer. A chain that can't lower its own mining difficulty without first surviving the wait it doesn't have the hashpower to survive isn't in a slow recovery. It's in a mechanism that gets worse the longer it goes unaddressed.

Worth sitting with too: the 2.53% wasn't spread across the mining industry. Every signaling block in the final voluntary period came from a single pool, Ocean, which had switched to signaling by default on July 15 — and reverted back to non-signaling as its own default just a day after the fork stalled. AntPool and F2Pool each mined a non-signaling block during the mandatory window specifically to demonstrate their position. Foundry USA ran an internal hashrate-weighted vote and never published a result or signaled either way. Once the mandatory window opened, signaling on the dominant chain dropped to 0.00% — not low, zero. What briefly looked like a contested 2.53% was, structurally, one pool's policy default being read as an industry position.

Why 2017 worked and this didn't.

BIP-110's authors explicitly modeled it on BIP 148, the 2017 User Activated Soft Fork that forced miners to activate SegWit. It's worth being specific about why that one actually worked, because BIP-110 skipped the part that mattered: by the time BIP 148 threatened to go live, major exchanges, wallet providers, and Bitcoin-accepting businesses had already credibly aligned behind SegWit, so miners faced a real threat of mining blocks the rest of the economy simply wouldn't price. That's what forced the capitulation — not the node software, the economic backing behind it. BIP-110 never built that alignment. Inscription-type transactions have been generating competitive fee revenue for pools since 2023, giving miners a financial reason to stay put, and no exchange or custodian ever signaled it would treat a BIP-110 chain as anything other than noise.

That showed up concretely: Coinbase and Kraken's public status feeds reported normal Bitcoin operations straight through the split, with no special handling announced. BlackRock's IBIT prospectus — like those of other major spot-ETF issuers — states the trust permanently and irrevocably abandons any claim to forked or airdropped assets, meaning ETF holders were never going to receive BIP-110 coins even if the chain had survived. Compare that to 2017, when exchanges published detailed splitting instructions and prepared new trading pairs well ahead of the Bitcoin Cash fork. The silence this time wasn't an oversight. It was the market declining to treat the event as one that needed a response.

The restrictions could reportedly be routed around anyway.

Independent of the signaling math, developer Martin Habovštiak demonstrated that arbitrary data could still be embedded on-chain in a way that avoided both OP_RETURN and Taproot — the very methods BIP-110 targeted — and that anyone could verify the bytes with standard node commands. If the restrictions can be bypassed by a single developer demonstrating the technique once, the case for accepting the governance risk of a contentious rule change gets considerably harder to make, regardless of how legitimate the underlying storage grievance was.

Even sympathetic voices thought the tactics went too far.

Samson Mow, who has said directly that he shares the concern about blockchain spam, opposed BIP-110 specifically over the lack of broad consensus — and modeled a "1% attack" scenario in which renting enough hashpower and running a few thousand cheap nodes could simulate a level of support the movement never organically had. He later wrote a viral piece of satire, framed as a survivor's diary from a future where the BIP-110 chain still exists with only 83 remaining participants, anchoring every fictional detail to a real data point already on the record — Lopp's Sybil-node chart, the pseudonymous author's identity, the mechanics of how a single extra byte could break the whole scheme. It's a useful reminder that some of the sharpest critiques of the movement's tactics came from people who agreed with its stated goal.

None of that is technically complicated once it's laid out. It's just a difficulty adjustment doing what a difficulty adjustment does to a chain that doesn't have the hashpower to sustain it. That part is trustworthy because it's arithmetic, independently reported by outlets that otherwise disagreed with each other about everything else. What's harder to trust is any confident early read of what the humans involved were thinking — that part only became clear by watching the community, not the block explorer.

03 — What actually settles itIt wasn't the code. It was the behavior.

To be precise about this, because it matters: this isn't a technical verdict. Whether BIP-110's design was genuinely bad engineering, versus a reasonable fix its critics were too quick to dismiss, is a real and contested question that deserves people who can actually read the script-depth argument and Jameson Lopp's warning about breaking BitVM-style contracts — and plenty of technically capable people land on different sides of it. What's easier to evaluate from outside the code is how the movement behaved once it started losing, using the same Jefferson-and-Monroe pattern already laid out: conviction that outran its backing, held onto anyway. That's the part that actually earns the word "delusional," and it doesn't require picking a side in the technical debate to see it.

1. ✕

   The CSAM argument was baked into the proposal's own logic before it ever became a talking point

   Months before activation, Jameson Lopp predicted almost exactly what would happen: that some "last-ditch" supporter would put child sexual abuse material on-chain specifically to coerce miners and node operators through legal pressure, since an actual criminal would have no reason to broadcast evidence of a crime onto a permanent, public, radically unprivate ledger. What makes this worse than a prediction coming true is that the incentive structure was reportedly written into BIP-110's own early draft language: a version of the proposal's reasoning argued that a bad actor could put CSAM on-chain to force a reorg, and that doing so created "an economic incentive for onchain CSAM" under the rule as proposed — meaning the argument being used to justify the fork also described the exact attack the fork would make more attractive. Once signaling stalled out, proponents used the CSAM framing as a pressure tactic anyway. Ocean's own Jason Hughes publicly rejected it and pointed out BIP-110 didn't even remove the technical ability to store arbitrary data on-chain. That the tactic got deployed after someone predicted it, after someone inside the movement debunked it, and after it turns out to have been part of the proposal's own reasoning, is where this stops being a technical disagreement.

2. ✕

   Some of the loudest support wasn't people at all

   Jameson Lopp has said directly that a meaningful share of the online movement behind BIP-110 was manufactured — that he personally confronted and unmasked AI bots on both X and Nostr that were repeating BIP-110 talking points and running circular arguments to waste people's time. Separately, in March 2026, he flagged what looked like a Sybil-inflated surge in BIP-110-signaling node counts — a chart captioned "Spot the Sybil Attack," showing one actor potentially running thousands of cheap nodes to simulate broader support than actually existed. A movement that needs bots and inflated node counts to look bigger than it is has already told you something about how much real conviction is underneath it.

3. ✕

   The Flat Earth talk in the Knots Discord

   This is the detail that actually clinches it. A community organized around a fairly narrow technical dispute — how Bitcoin should handle non-payment data — drifted, in some of its channels, into entertaining Flat Earth arguments. That's not a subtle signal. It's a sign of a group that has stopped checking its beliefs against anything outside itself, on any subject, not just the one it originally formed around. Understanding BIP-110's code isn't required to recognize that pattern, because it isn't really about Bitcoin at all.

4. ✕

   The refusal to accept 2.53% as an answer

   A healthy technical community loses arguments sometimes and updates. This one, when it lost, didn't treat 2.53% signaling as information — it treated it as an obstacle to route around, first by discussing a proof-of-work change that would let it abandon the miners entirely, then by continuing to defend the proposal's legitimacy after its own governance process had moved it to Closed status. That's the pattern that actually reads as delusional: not holding an unpopular technical opinion, which is fine, but being structurally unable to let evidence change the plan.

04 — In their own wordsWhat the people actually involved said

Re-reading the primary sources, rather than relying on general impressions of the community's mood, sharpens the picture. Here's a representative sample, on both sides of the argument.

Reading these together, the disagreement was substantive and, on the technical merits, fairly one-sided against BIP-110. But it was an argument between people who mostly agreed on the problem — node storage costs are real, arbitrary data on-chain is a real design tension — and disagreed about the fix. That's a normal, healthy fight for an open protocol to have. It isn't a story about one side being sane and the other delusional. That story only gets earned once you look at what happened after the fight was lost.

05 — What kept happeningThe fallout wasn't just technical either

Two more things happened after the stall that reinforced the same pattern — that the real story here was institutional and behavioral, not a clean math problem.

Luke Dashjr was removed as a BIP editor.

On August 10, Bitcoin's remaining BIP editors revoked Luke Dashjr's repository permissions, deleting his name from BIP 3, the document that lists who is allowed to manage the proposal process. Fellow editor Olaoluwa Osuntokun wrote that Dashjr had used his editor position to fast-track BIP-110 by assigning it a number outside the normal process, calling it a conflict of interest. Bryan Bishop confirmed the change on the Bitcoin Development Mailing List; Bitcoin Core contributors Matt Corallo, Antoine Poinsot, and Anthony Towns, along with libsecp256k1 maintainer Jonas Nick, backed the removal. Dashjr rejected the characterization and called it an unauthorized abuse of power. Separately, he had argued in July that abandoning BIP-110 at that point would itself be "a contentious hardfork attempt," since in his view there was no consensus to reject it either.

Whether removing an editor without a written process was the right call is a governance question this piece isn't equipped to referee. What's clearer is simpler: even Bitcoin's own protocol-governance layer, widely assumed to be settled and procedural, turned out to have an undefined edge case that let one person push a controversial proposal through, with no formal way to undo it cleanly once it became a problem. Jameson Lopp pointed that gap out himself. It's a reminder that "Bitcoin's process" being solid was itself an assumption worth checking, not a given.

No replay protection meant the risk didn't end when the chain stalled.

BIP-110 shipped without replay protection, meaning a transaction signed on one chain could be valid on the other. Developer Kevin Loaec warned ahead of the split that holders who sold or moved "airdropped" fork coins risked inadvertently authorizing the same spend on the real Bitcoin chain — attackers could specifically target large holders, since a successful replay against them would pay off more. This is the part of the story that has nothing to do with belief or ideology at all: ordinary people who held no opinion whatsoever about BIP-110 were exposed to a real, concrete risk of losing real bitcoin, purely because the fork existed and shipped without a basic safety mechanism. Whatever else is true about the arguments on either side, that risk was avoidable and it fell on people who weren't part of the argument.

06 — What still holds upNot everything about the original argument was baseless

None of this is an argument for treating the original grievances as naive. A few things still seem right, and are worth standing behind:

The underlying grievance about node storage was real.

Even critics who thought BIP-110 was a bad and dangerous fix, like Jameson Lopp, didn't dispute that the storage asymmetry — miners paid once, node operators storing the data forever — is a genuine, long-standing tension in Bitcoin's design. That part of the argument was sound. What didn't follow from it was that the proposed fix was sound.

The neutrality argument is a real and serious one.

Saylor's and Back's core worry — that once a protocol's maintainers start ruling on which fee-paying transactions are "legitimate," there's no principled place to stop — is a genuine structural concern about censorship resistance, not just a rhetorical flourish deployed to win an argument.

Echo chambers are a real mechanism, not just a slur.

The psychological account — cognitive dissonance driving groups to double down rather than update when reality contradicts their narrative — is well-documented, and it's the specific, falsifiable thing that actually played out: not a general vibe, but the CSAM rhetoric, the Flat Earth talk, and the refusal to accept 2.53% as an answer, all in the same channels, in the same week.

07 — The actual takeawayThe code isn't the story here. The behavior is.

This isn't a claim to have suddenly acquired the technical grounding to say BIP-110 was bad engineering — that would just be trading one kind of overconfidence for another. The narrower, honest claim is this: a lot of belief in BIP-110 formed by trusting the people around it rather than the specification itself, and the clearest signal that the trust was misplaced came from watching that same community turn on outsiders, entertain conspiracy theories unrelated to Bitcoin entirely, and refuse to treat its own failed signaling numbers as real information. That's not a verdict on the code. It's a verdict on what happens to a group of people once conviction stops being checked against anything outside itself.

The wilted rose above isn't decoration. It's the image that fits: something that looked, from a distance, like it was still standing — right up until you looked closely enough to see it wasn't going to recover. That's the Knots community by the time the Flat Earth talk showed up. It's also a fair picture of any belief, in any community, that's never been checked against anything outside itself.