CAPTURE

An investigation across four articles into how informal power over Bitcoin Core was assembled, exercised, and defended.

Article Two: The Lever

https://www.citadel21.com/the-lever

CAPTURE

An investigation across four articles into how informal power over Bitcoin Core was assembled, exercised, and defended.

Article One: The Network

https://www.citadel21.com/the-network

How Bitcoin Core's developer selection, developer funding, and developer moderation were consolidated under a single informal architecture
naddr1qqv8...

Human nature insures there will never be a time when Bitcoin does not need to be defended and fought for.

Complacency is an attack vector.

Eternal vigilance is the price of liberty.

By repeatedly adjusting default mempool policy to match what miners will accept anyway (large OP_RETURN uncapped because “they’ll just mine it via bypasses like Libre Relay, or direct APIs”), we are implicitly conceding that miner greed + economic incentives are the ultimate rule-setter, not node-enforced principles.

Meaning a (cleverly hidden) capitulation of Bitcoin as a decentralized project.

You could say that the “CENSORSHIP!” argument from Core and their supporters on the concept of nodes filtering is a roundabout way of critizing decentralization itself.

Nodes were always supposed to be the sovereign check, they decide what they accapt and relay.

When we keep loosening policy to align with whatever is the current grift “use case”, and by extension what is short term profitable for miners, it trains the entire ecosystem to treat restrictive node behavior as pointless theater.

Over time this hollows out node sovereignty: running a full node becomes more about passively observing the chain that miners + L2s + data-spammers have already decided on, rather than actively enforcing a monetary-first standard.

As a cuck bonus it also leads to higher resource costs for every honest node (bandwidth, RAM, storage) à fewer independent verifiers in practice

Decentralization starts looking like a performance act. Miners produce the blocks, a handful of relays and L2 sequencers steer the flow, and nodes just… validate after the fact.

It’s not a hard-fork capitulation (consensus rules haven’t changed), but it is a cultural, philosophical and operational one. The most profound capitulation in practice.

The philosophy flips from “Bitcoin should resist non-monetary garbage even if it costs us some short-term fee revenue” to “whatever pays miners gets standardized because resistance is futile.”

Once you accept “miners will do it anyway” as the justification for policy, you’ve already handed the character of Bitcoin over to the highest bidder. Nodes stop being the immune system and start becoming just a polite audience.

The OP_RETURN uncap looks a lot like another quiet step toward a two-tier network (miners + insiders set the tone, everyone else just watches. Keep doing this and running a node risks becoming a branding exercise instead of the actual source and guarantee of Bitcoin’s decentralization.

Spent some time researching this
naddr1qq0x...

In July 2023 and September 2023, two pull requests hit the Bitcoin Core repository.

One tried to tighten data limits. One tried to remove them entirely.

They were moving in opposite directions. The people behind them knew what they were doing.

***

PR #28130, July 2023: Peter Todd files to remove the OP_RETURN mempool limit entirely.

PR #28408, September 2023: Luke Dashjr files to extend -datacarriersize to cover the SegWit/Taproot inscription loophole.

Luke tightening. Todd eliminating. Both active at the same time.

***

Todd NACKs Luke's PR, calling it censorship.

PR #28408 is killed. PR #28130 is closed without adoption — but it has done its job: established the position publicly and seeded the argument for next time.

Kill the defence. Advance the attack. The pincer has two arms.

***

What Todd does not disclose when NACKing Luke's patch: he operates Libre Relay.

Libre Relay is a direct-to-miner relay service routing non-standard transactions — including inscription-heavy ones — to miners, bypassing mempool policy entirely.

***

Todd's argument: filters are ineffective because miners include non-standard transactions anyway.

Libre Relay is part of the infrastructure that makes this self-fulfilling.

He built the bypass. Then cited the bypass as proof limits don't work.

***

April 2025: Todd files PR #32359 to remove the limit.

He later admits on Stacker News:

"This pull-req wasn't my idea. I was asked to open it by an active Core dev because entities like Citrea are using unprunable outputs instead of OP_Return."

***

Citrea: a VC-funded ZK-rollup whose business model needed more on-chain data storage.

The PR was not an organic expression of Todd's technical views.

It was a brief, handed to him by an unnamed active Core developer, to solve a corporate client's problem.

***

Samson Mow calls it "PR laundering" — routing through Todd to produce the appearance of independent initiative.

Antoine Poinsot (Chaincode Labs) connected to early discussions.

The same Poinsot who disputed Luke's CVE in October 2024. Both ends of the sequence.

***

Jameson Lopp publicly advocates for the PR.

He does not disclose he is an investor in Citrea — the same company whose data needs triggered the PR.

***

The PR draws 423 thumbs-down against 105 thumbs-up.

Ava Chow had stated publicly in December 2023: "If it is controversial, then we don't touch it."

June 9, 2025: Gloria Zhao merges it anyway.

***

PR #32406 does something beyond uncapping OP_RETURN.

It removes -datacarrier and marks -datacarriersize as deprecated — the switches letting node operators filter data-carrying transactions from their mempools.

Luke Dashjr authored those options in 2014.

***

The justification: the flag is "obsolete" and a "footgun."

Gloria's announcement tweet confirmed the deprecation explicitly.

The escape hatch was being closed behind the change.

***

Under sustained pressure, Core maintainer Ava Chow reversed the deprecation via PR #33453 — merged hours before the v30 release window in October 2025.

User configurability preserved — for now.

The 100KB default was not reversed.

***

Thirty-one Bitcoin Core contributors sign an open letter supporting the merge.

The letter uses the word "censorship" to describe any opposition.

***

Gloria's public statement on X:

"Demanding that Bitcoin Core prevent certain transactions from being mined reflects a misunderstanding of the relationship between open source software users and developers."

She deletes the account on May 15, 2025.

***

The community response is the largest organised opposition to a Core change since the 2017 block size wars.

Bitcoin Knots — Luke's alternative implementation — surges from ~2% of the network to over 21%.

5,114 Knots nodes at time of v30 release. BitRef data confirmed.

***

Dennis Porter, who had raised over $200,000 for Core developers:

"My faith in their work is now broken. I will no longer be financially supporting Core development."

***

Nick Szabo — pioneer cryptographer, silent on social media for five years — returns for v30.

"I strongly recommend not upgrading to Core v30."

He also flags criminal liability for node operators storing illegal content they can no longer remove.

***

Bitcoin Core v30 ships October 11, 2025.

Within weeks, downloads are pulled. The release contains a bug capable of deleting Satoshi-era wallet.dat files during migration — potentially destroying funds held by early holders who had not backed up separately.

***

A release delayed for weeks due to governance controversy shipped with the most serious wallet safety bug in years.

Knots supporters noted the contrast immediately.

***

One supporter's statement is worth preserving.

Ark Labs Ecosystem Lead Alex Bergeron stated publicly he intends "to use all of the additional OP_Return space and WILL use it to make Bitcoin more like Ethereum, except better."

A proponent confirming what critics warned.

***

2023: reject Luke's patch. 2025: merge the uncap using the open loophole as excuse.
2021: try to remove Dashjr as BIP editor. 2025: mute him on the OP_RETURN PR.
2014: Luke builds the configuration option. 2025: try to deprecate it in the same merge.

Not a series of independent events.

VC money is a helluva drug.

Symptoms include:

- sudden inability to define the word "spam", previously straightforward concept, now philosophically complex

- the word "censorship" begins to apply to things that were previously called "node policy"

- previously held view that "arbitrary data does not belong on a monetary ledger" becomes "who are we to say what belongs?"

- strongly held technical positions develop a mysterious flexibility