Resolves YES if, before January 1, 2028, Bitcoin's core security breaks and AI was substantially responsible. I resolve this based on my own judgment, guided mainly by expert consensus and news.

"Broken" means one of Bitcoin's basic guarantees has failed in practice. Either coins can be spent by people who don't hold the keys, or confirmed transactions can be reversed. Clear YES cases:

• A practical way to derive private keys from public keys (breaking secp256k1 ECDSA/Schnorr).

• A practical attack on SHA-256 that undermines mining or address security.

• Successful deep reorgs or double-spends on mainnet.

• A protocol or Bitcoin Core bug exploited on mainnet with serious consequences, such as inflating supply, splitting the chain, or mass theft.

• Quantum computers breaking secp256k1 keys in practice. Given the timeline, I’ll presume AI was decisive in getting there unless there’s decent evidence otherwise.

"Substantially responsible" means an AI system did the key work: finding the math, finding the bug, or running the attack. A human using AI as a better search engine doesn't count.

Doesn't count:

• Hacks of exchanges, custodians, or wallet software, and phishing, malware, or weak/brainwallet keys. These are failures around Bitcoin, not of Bitcoin.

• Vulnerabilities found and patched before serious exploitation.

• Relatively small mining efficiency gains that don't let anyone attack the network. E.g. a 10% better mining method just makes a better miner. A SHA-256 shortcut big enough that one party can cheaply outpace everyone else's hashrate does count.

(Description mostly generated by Opus 5.5. I read through it, iterated, and edited a bit.)