Same Money, Different Soul: why I build for Bitcoin Blake2b

• The money is identical, and that is the point
• The hash is not the reason, the header is the real work
• The users came, the miners did not
• Who builds the block decides what gets in
• Money, not JPEGs, at least for one year
• Capture, and the question of who sits at the table
• What happens to the duplicate coins of the big holders?
• The honest weak points
• The people who build it
• My take, and why I build for Bitcoin Blake2b

In August 2026 Bitcoin split in two. A small group of users forked the chain at block 961,632, changed the proof of work from SHA-256 to BLAKE2b, and activated a data limit that the majority had refused. Since then, every bitcoin that existed before the fork exists twice, once on each chain, under the same keys. But only the people who hold their own keys can actually reach both. If you hold real bitcoin on chain in self custody, you own the new coins too. If your bitcoin sits at an exchange, in an ETF or with a custodian, the key holder decides, and so far the key holders have decided nothing for you.

Most of what I read about this fork asks the wrong question. People ask which chain is the “better money”. But the money is the same. So I spent the last weeks comparing both chains property by property, with sources and dates. Because this does not fit into an article, I built an HTML companion page for it: Same Money, Different Soul. It has more than 150 rows of comparison, a filter that shows only the points where the two chains really differ, and a search over the whole page. Every section of this article links to the matching part of the companion page, so you can go deeper wherever you want. Here I share the main findings and, at the end, my own view and why I build software for the new chain.

§The money is identical, and that is the point

Bitcoin Blake2b inherits everything from Bitcoin up to the fork: the full history, the 21 million cap, the halving schedule, the addresses and the keys. There was no premine and no new issuance. If you run through the classic list of what makes good money, verifiable, scarce, portable, divisible and so on, both chains score the same. You can see this in the traits of money table.

Only three things really move. Bitcoin has sixteen years of history and enormous hashrate behind it, and time and hashrate cannot be forked. The new chain has a different idea about who should control the protocol. So this is not a monetary debate. It is a governance debate, and the hash function is mostly a costume. I summarised the seven real differences at the top of the site.

§The hash is not the reason, the header is the real work

BLAKE2b is a modern and clean hash function, a bit faster in software than SHA-256. But this does not make the chain more secure, and it does nothing against quantum computers. The quantum risk sits in the signatures, and both chains use exactly the same signature scheme. Anyone who sells BLAKE2b as “quantum safe” or “ASIC proof” is wrong. BLAKE2b miners from the Siacoin world have existed for years. The ash comparison shows this clearly.

The more interesting change is one that almost nobody talks about. The fork introduced a new 164-byte block header. The mining machine now only hashes a small standard work header and has no logic for building blocks. At the same time the new header fixes some old weaknesses of Bitcoin, for example the transaction count ambiguity behind CVE-2017-12842, and it removes ASICBoost. This is real engineering work.

But I also want to be fair here. The same release switches off the timewarp protection on mainnet, while it is active on the test network. A pull request to fix it is open, but it is not merged yet. For me this is the strongest technical criticism of the fork, and I think a project earns trust by naming its own weak points. Details are in the header section.

§The users came, the miners did not

The number that explains the whole fork is this one. When the data limit was proposed on Bitcoin, around 18 percent of the public nodes were running the client that supported it. But only about 2.53 percent of the miners signalled. The mining industry did not vote no. It simply did not show up.

A change of the proof of work is exactly the tool for this situation. SHA-256 machines cannot mine BLAKE2b, and the other way round. So the users changed the lock, and the old mining fleet stayed outside. The node and miner numbers are on the site.

§Who builds the block decides what gets in

Censorship resistance is not decided by the mining hardware. It is decided by whoever builds the block template, because that person chooses the transactions. On Bitcoin today about four pools build around 70 percent of all blocks. If a government or a big company wants to block certain transactions, it only needs to put pressure on a few nameable operators.

The answer is self templating: every miner builds the block on their own node. Tools like DATUM do this, and on Bitcoin Blake2b they were available from the first day. When I started with this chain, that was one of the main reasons I run my own node and my own template instead of pointing a machine at somebody else’s pool.

Here I have to correct my own first draft. Three weeks ago the biggest pool on the new chain had around 18 percent and I wrote “no dominant pool”. Today the three largest pools mine about 72 percent of the blocks, and one of them has a third. That is about as concentrated as Bitcoin. The difference is that leaving a pool and building your own template is cheap and normal on this chain, and some pools, like CONVOY, are built on that model. But decentralisation here is an intention and a toolset, not a result. You can follow it in the decentralisation section and in who picks the transactions.

§Money, not JPEGs, at least for one year

Since 2023 Bitcoin blocks carry a lot of data that has nothing to do with payments: inscriptions, Runes, BRC-20 tokens. I noticed this myself when I synced a node. Validation becomes really slow when it reaches the blocks from 2023 to 2025.

The new chain limits this kind of data at consensus level and reduces the block weight to 800,000 units, which is typically about 300 kilobytes. But all of this is temporary. The rules expire on 1 September 2027 unless they are extended. So the honest sentence is “a minority chose lean blockspace for one year”, not “Bitcoin fixed spam”. The rules are listed in the RDTS section.

§Capture, and the question of who sits at the table

In July 2026 nine institutions, including BlackRock and Strategy, launched the Bitcoin Security Consortium and pledged 15 million dollars to fund Bitcoin developers. Strategy and BlackRock’s ETF hold together about 1.58 million bitcoin. The consortium says it has no role in governance, and I believe the quantum threat is real and needs paid work. Still, influence follows money. The two largest holders now sit closest to the people who write the code. There is no conspiracy needed to see that incentives can drift.

This becomes very concrete with quantum computing. One proposal for Bitcoin, BIP-361, would after some years freeze all coins that did not move to a new quantum safe address type. That affects about 6.7 million bitcoin, including the coins believed to belong to Satoshi. The intention is protection. The effect is that coins you own could become unspendable by consensus. For me this is where “your keys, your coins” really gets tested. Bitcoin Blake2b has the same quantum problem and no roadmap yet, so it will face the same question one day. The difference is who sits at the table when it is decided. See who funds the developers and the quantum fault line.

§What happens to the duplicate coins of the big holders?

This was the question I found most interesting. Every bitcoin that existed at the fork also exists as a Blake2b coin, and that includes the coins of the ETFs and of Strategy. At the current price of around 311 dollars that is roughly half a billion dollars on paper. So who gets it?

I read the filings. BlackRock’s IBIT, Fidelity and Grayscale have written into their documents that they irrevocably abandon forked coins. Their shareholders get nothing. Bitwise sends such coins to its sponsor. Strategy has no fork policy at all, but its custodians have no duty to support a fork. And the market could not absorb these coins anyway: on the exchanges there are only a few dozen dollars of orders close to the current price. Paper value is not real value.

There is one company in the middle of all this. Coinbase is the custodian, or one of the custodians, for almost every US spot bitcoin ETF, including BlackRock’s IBIT. Only Fidelity keeps its coins elsewhere. In 2024 Coinbase said it held about 90 percent of all ETF bitcoin. It is also one of Strategy’s custodians, and it holds the coins of millions of normal customers. Whoever holds the keys is the only one who can technically reach the Blake2b coins, and for all these coins that is Coinbase. So what is Coinbase doing with them? The honest answer is: we do not know. Coinbase has not said a single word about the fork. Its terms leave forks to its own discretion, the ETFs have written that their custodian holds nothing for them, and Coinbase’s own fork policy from 2018 excludes assets with low liquidity, which rules this coin out. In the past Coinbase credited Bitcoin Cash to its customers, but it never supported Bitcoin Gold, and a court confirmed that customers had no right to it. So the Blake2b coins behind the ETFs are most likely just sitting under Coinbase’s keys, owned by nobody who claims them. Bitcoin University made a video about exactly this question, Will Coinbase Steal Customers’ Bitcoin?, and I added the details to the Coinbase part of the companion page.

For me this is the clearest lesson of the whole fork. Only people who hold their own keys received the new coins in a way they control. Everybody else depends on a decision that somebody else has not made. “Not your keys, not your coins” was never shown more clearly.

But there is a detail that most people miss, and it is the reason I made replay protection the centre of this section. A normal Bitcoin signature is still valid on the new chain. Only signatures that use the new “unified sighash” are protected. So if somebody moves pre-fork coins on Bitcoin with a normal signature, anybody can broadcast the same transaction on the Blake2b chain, and the Blake2b coins go to the same receiver. This is not theory. It is documented by the project and shown step by step in public video guides.

For every one of us this means: if you sell or move old bitcoin without splitting your coins first, you may give away your Blake2b coins without knowing it. And if you move Blake2b coins with an old wallet, you may move your bitcoin too. Split first, with a wallet that signs with replay protection, and test with a small amount. The full story is in the duplicate coins section and the practical steps are in replay and coin splitting.

§The honest weak points

I do not want to write an advertisement. The new chain is small and fragile. Its hashrate grew about twelve times in three weeks, but it is still around four orders of magnitude below Bitcoin, so a 51 percent attack is much cheaper. It trades on only two no-KYC exchanges, the order books are very thin, and the reported volumes do not match them. It is not on CoinGecko or CoinMarketCap. The ticker fight around “XBT” did not help either. You find the numbers under security and scale and where it trades.

The software around the chain grew faster than I expected. There are two desktop wallets based on Sparrow, early Android wallets, community firmware for Trezor, Coldcard, SeedSigner and Krux, Electrum servers, and first Lightning builds. But almost all of it is unofficial and not audited. That is normal for a chain that is one month old, and it is also a real risk. The ecosystem table lists everything.

§The people who build it

What impressed me most is how much work a small group of volunteers did in a few weeks. Bitcoin Knots carries the consensus code. Teams and individuals like privkeyio, paulscode, Retropex and Kilombino built wallets, firmware, servers, explorers and home server packages. Independent developers built mining gateways, so that no single implementation decides how blocks are made. And people like Matthew Kratter with Bitcoin University explain the fork to a wide audience and ask the hard questions. I linked all of them in the builders section.

§My take, and why I build for Bitcoin Blake2b

I do not think Bitcoin on SHA-256 is going away, and this is not a call to sell anything. But I believe that sound money needs an exit. For years the answer to every problem in Bitcoin was “use your voice and convince the others”. This worked as long as the voices were users and developers. Today the loudest voices are ETFs, treasury companies and a few mining pools, and a normal user cannot out-lobby them. When voice stops working, exit is the old cypherpunk answer. Bitcoin Blake2b is such an exit, and even people who never use it benefit when it exists, because an alternative keeps the majority honest.

I also do not believe in an alternative that only exists on paper. An alternative is only real if normal people can run it at home, on their own hardware, without asking anybody for permission and without trusting a company. This is why I build with TwentyOne.Life. We maintain Shulcrum, an Electrum server for the new chain, so that wallets can ask your own node instead of a stranger’s server. And we package Shulcrum and the Shrike wallet for StartOS, so that a home server can index the chain and serve your wallet over Tor. It is small work compared to what others did, and I say openly where it stands: Shulcrum is verified on the test network, and the verification on mainnet is still running.

My personal summary is simple. The money of both chains is the same. The difference is who controls the protocol, who builds the blocks, and whether “your keys, your coins” is still true in twenty years. Bitcoin on SHA-256 wins on scale, liquidity and history, and this will not change soon. Bitcoin Blake2b wins on independence, and it still has to prove that it can stay independent when it grows. I see it as insurance against capture, not as an investment. And I prefer to help build that insurance while it is small, instead of wishing for it later.

The whole comparison is on the companion page. The numbers are from 26 September 2026 and on a chain this young they change fast, so I will update them. If you find a mistake or have a better source, please tell me. I am happy to correct it.