LearnBitcoin

The Bitcoin School

For people who want to understand Bitcoin. Not flip it.

Start the Journey

Plain language. Sourced chapters. Live on-chain data from a real node. Huge glossary, Rabbit Holes, and one opinionated Journey from "what is money?" to running a full node. No altcoins, no affiliate links, and no signup needed to read any of it.

From the glossary

Transaction Index (txindex)

An optional node setting creating a full index of all transactions by TXID for direct lookup.

Read full entry

New on the site.

Bitcoin only.

No altcoins. No tokens. No "crypto." Other coins have other sites. We have one.

No bullshit.

No price predictions. No affiliate-bait. No fear marketing. Every chapter cites its sources, and the live on-chain numbers come straight from a full Bitcoin node.

Have fun.

Interactive, opinionated, well-crafted. Rabbit-hole vibes. Built by Bitcoiners for people who want to actually understand.

Down the rabbit hole.

All rabbit holes

Corrected in public.

All corrections
  • BIP 360 (Pay-to-Merkle-Root)

    BIP-360 says a P2MR output becomes vulnerable to long-exposure attacks whenever one of its scripts reveals a public key, so the protection depends on not reusing keys. P2MR is a destination for coins that move to it and does nothing for keys that stay exposed. Neither BIP mentions the other, and BIP-361 lists a post-quantum signature BIP as its requirement, which P2MR is not. The entry now says all three.

    What was wrong
  • BIP 361 (Post Quantum Migration)

    The draft puts Phase A about three years (160,000 blocks) after BIP-361 activates and Phase B two years after that, about five years in all. Its authors say no rescue protocol is believed possible for P2PK and point to a separate Hourglass proposal instead, list abandoned coins staying unspendable as a benefit ('supply discipline'), and cite 2027 to 2030 as an 'as early as' estimate sourced to McKinsey. The entry now says all of this and adds Taproot outputs as the third way a public key ends up on chain.

    What was wrong
  • CRQC (Cryptographically-Relevant Quantum Computer)

    Demonstrated logical qubit counts are in the dozens, not single digits: a neutral-atom processor ran circuits with up to 48 logical qubits in December 2023, on small codes with error detection rather than deep error correction. Neutral-atom processors are gate-based too, so the hardware section now groups platforms by the kind of physical qubit they use and cites the Google Willow and Harvard atom-array demonstrations.

    What was wrong
Edit this page on GitHub