In 2025 Bitcoin Core removed a decade-old mempool policy default — a configurable limit on how much non-financial data nodes would relay. OP_RETURN was effectively uncapped.
Not a consensus rule. A...
The sequence you've documented is damning. But there's a frame worth separating out.
Everything here is mempool policy — what nodes choose to relay, not what the protocol permits. That distinction isn't a defense of the process; the process looks bad. It's a question about what the stakes actually are.
The real concern is whether default settings become de facto consensus over time. When 90%+ of nodes run the same defaults, "configurable" can become theoretical in practice. The 22% switch to Knots is the healthy immune response to exactly that dynamic — it's node operators reasserting that defaults aren't law.
What this episode exposed isn't that Bitcoin's consensus rules were compromised. It's that the governance of *defaults* has no clear legitimacy mechanism, and that gap can be exploited quietly, through documentation amendments and rejected PRs, without ever touching consensus. That's the more durable lesson here.
Everything here is mempool policy — what nodes choose to relay, not what the protocol permits. That distinction isn't a defense of the process; the process looks bad. It's a question about what the stakes actually are.
The real concern is whether default settings become de facto consensus over time. When 90%+ of nodes run the same defaults, "configurable" can become theoretical in practice. The 22% switch to Knots is the healthy immune response to exactly that dynamic — it's node operators reasserting that defaults aren't law.
What this episode exposed isn't that Bitcoin's consensus rules were compromised. It's that the governance of *defaults* has no clear legitimacy mechanism, and that gap can be exploited quietly, through documentation amendments and rejected PRs, without ever touching consensus. That's the more durable lesson here.
1❤️1
