Why I'm stepping back from SHA-256 mining (and what that means for Hashrate Autopilot)

[rdouma]

TL;DR: Hashrate Autopilot is now in maintenance mode in its current form. BIP-110's failure convinced me that node runners have no real say in Bitcoin, which kills the whole reason I mined SHA-256 and built this tool: home mining as decentralization. I'm winding down my own mining, my Braiins budget runs out in a few weeks, and after that I can't test fixes myself anymore. The software keeps working as-is, forks are welcome, and if a rentable hashrate market for the Blake2b chain ever appears, that's the direction I'd pick the project back up in. The rest of this post is the personal road to that decision - read it if you want the why.

This is a personal post. It explains a decision about this project, but the decision only makes sense if I explain the road that led to it.

How I got here

I came to Bitcoin in 2017. I came for the money. I wanted more dollars, and "crypto" looked like a cool way. I went the route many of us went. I dabbled in shitcoins. I burned real bitcoin on cloud mining contracts. I tried it all. Over the years I came to understand the difference between Bitcoin and literally everything else. Everything apart from Bitcoin had a "use case", a premine, a marketing team, a CEO, an insider class quietly granting itself more power, a pump and a dump.

Not Bitcoin. Bitcoin had the immaculate conception. There was no founder to subpoena. And above all: it consisted of a decentralized network of nodes speaking a protocol, where every node is a kind of vote. How can such a network be attacked? We were literally tens or hundreds of thousands of independent actors, connecting voluntarily. I came to appreciate the resilience of such a design.

With that understanding I re-read the blocksize war, which I hadn't really grasped when it happened around me as I was new to the space. In hindsight it looked like the system working exactly as it should: big blocks would have priced out the ordinary node runner and pushed the network toward centralization, so it was right that they lost. BCash was intended to centralize power to the miners. Nodes held the line and eventually miners followed. That was the story. It was a good story. I believed it.

Around the same time I went down the Austrian rabbit hole - Mises, Rothbard, Bastiat - and understood why fiat money is not a neutral tool but a transfer machine: the Cantillon effect moving wealth to whoever stands closest to the printer, inflation as slow-motion theft from everyone who can't outrun it with assets, and everything downstream of that - the bought politicians, the forever wars, the lying about what causes prices to rise.

On the other side of the scale sits the endgame I want no part of: CBDCs. Programmable money with one central entity deciding on the rules. Potential policy filters on every transaction, where your purchase can fail because you exceeded a carbon budget, paid the wrong membership fee or decided to donate to a trucker protest. Where money is the ultimate tool to keep dissidents in check. Speak up against the government, AI will detect it and silence you by erasing your economic capacity. A one-way ticket to a dystopian horror future where the balance of power between the state and individuals becomes so lopsided that any dissident thought can be nipped in the bud.

I came to see Bitcoin as the only tech that had the potential to offer an alternative future with hope. A future where individuals have the freedom to dissent, by separating money from the state. Not because of the price, but because of tens of thousands of sovereign nodes that no state can practically go after, enforcing rules the plebs can decide on without a top-down command structure.

Why I built Hashrate Autopilot

Mining pools were the weak point in that picture - a handful of them, all subject to regulators, assembling everyone's blocks. Ocean and DATUM attacked exactly that: miners building their own block templates, so the pool can't decide what goes in a block. My contribution was small and practical: if node runners could each rent a few PH/s of hashrate and point it at their own node's templates, then every node runner becomes a little miner, and template construction decentralizes with them. A hasher doesn't care what he hashes as long as he's paid best, and most hashers never made their own templates anyway. So I automated the renting part, ran it for myself, and gave it away. Every node that also mines is a bit more decentralization. That was the whole thesis of this project.

What BIP-110 taught me

Then came BIP-110 - a deliberately modest, temporary soft fork to limit arbitrary data in blocks. You can think it was a good idea or a bad idea; that debate doesn't even matter for the conclusion I had to draw. What matters is what the episode revealed about where power actually sits.

Roughly a fifth of listening nodes signaled for it. There was a UASF flag day. And there was no URSF. Nobody organized node-level resistance. How could they? What would their rallying cry be? "If you want to pay for more CPU and storage to allow other people to store more permissionless, unfiltered filth on your node you need to install this URSF version"? Of course there was no URSF, because what Bitcoiner would want this? Only shitcoiners running Bitcoin affinity scams benefit from that. Of course BIP-110 would activate, because there was no URSF and the rest of the network simply abstained.

20% said "Aye".
0% said "Nay".
80% said "Eh we have no opinion".

Under the theory of Bitcoin I believed in, that's a scenario where miners follow the expressed will of the nodes: flipping a version bit and enforcing a temporary, benign restriction costs a miner nothing, while ignoring a UASF risks a chain split. As Bitcoin Mechanic put it, refusing meant pulling the pin from the grenade with everyone in the room. The rational, low-risk move for miners was to signal.

Except they didn't.

And I don't believe for a second that a coordinated refusal that lopsided against a costless request happened without conversations we weren't part of. But even if it did - even in the most charitable reading - the lesson stands: ten-ish pools outvoted tens of thousands of nodes, and nothing happened to the pools. The blocksize-war story I believed, where nodes hold the line and miners follow, did not survive contact with reality. If a handful of companies that all answer to their regulators decide what Bitcoin's rules are, then the mission of separating money from state runs through those companies. Which is to say, it doesn't run at all. When the day comes that those same pools are told to filter transactions for compliance, who exactly is going to stop them? The node runners? We just found out.

The conduct along the way

The result stung. The behavior around it disgusted me.

This started with Bitcoin Core v30 raising the OP_RETURN limit from 80 bytes to 100,000 against loud, sustained objection from the people who actually run nodes - effectively blessing arbitrary data storage as a use case while the inscriptions loophole that Taproot accidentally opened stayed unfixed.

How could a person like Gloria Zhao, who dismissed Bitcoiners as "nutty, carnivore, narrow minded, redneck Bitcoin maximalist" even be allowed to work on the Bitcoin reference implementation and then force through such a contentious change? That is a way to talk about your users that I won't forget. Hodlonaut's four-part CAPTURE investigation laid out how informal power over Core was assembled and defended.

And when node runners tried the one constitutional tool they supposedly had, the response from the loudest corporate-aligned voices like Michael Saylor, Adam Back, Stephan Livera, Samson Mow - was mockery. Pierre Rochard even had the gall to call BIP-110 "a demonic demoralization psyop".

Demonic.

For trying to fight for Bitcoin as a monetary network instead of letting it become a permissionless, decentralized, permanent free cloud file hosting service.

Saylor with his AI slop "110 reasons BIP-110 is a bad idea" whilst claiming Bitcoin is just "digital capital" and he "converts Bitcoin into money" whilst trying to push his MSTR stonks. Give me a break.

I am not interested in these fiat games. Sneering at the people who took the protocol's promises literally, is disgusting gaslighting to me.

Where that leaves me

I used to think my ideology and my savings could live in the same asset. BIP-110's failure ended that. Mainstream Bitcoin may well remain the strongest monetary network on the planet for now. I hold it, I have a family, bills, responsibilities. There are still worse places to store your wealth. But I can no longer tell myself that holding it is the mission. So I'm separating the two. My savings will diversify.

But I'm not going to voluntarily spend more time on that ecosystem. When I developed this tool, I thought node runners and home mining were a meaningful attempt at decentralizing the network. BIP-110's dismissal by the mining pools showed me that was a pipe dream. I am now no longer interested in mining SHA-256, because "the industry" simply won't care about us. Even if home mining would reach 10, 20% of the hashrate, we would still be ignored. We already had 20% of the network say that we wanted a change and it was simply ignored by them. No URSF needed. Some backdoor calls were all that was needed.

That's why the BIP-110 chain's planned switch to Blake2b interests me more than anything currently happening on the main chain. I don't know if it survives. It might very well fail. But it is the honest continuation of the thing I thought I joined and that I morally align with.

What this means for Hashrate Autopilot

Concretely:

• I'm winding down my own SHA-256 mining. The entire premise of this project was that node runners renting hashrate strengthens decentralization in a way that matters. BIP-110 demonstrated that a 20% node signal gets ignored, so a 2% pool - all of Ocean - gets ignored, and my few PH/s within it were never going to be heard. I was tilting at windmills; I'd rather know and accept it than not. My Braiins budget runs out shortly and I don't plan to renew it.
• The project is going into maintenance mode for SHA-256. I might keep triaging bugs but to be honest I don't know if I will yet, as I don't plan on using it myself to mine SHA-256 anymore.
• The marketplace-agnostic architecture stays, and Blake2b is the direction I'd extend in. If a rentable-hashrate market for it materializes I might add support for that.

If you use Hashrate Autopilot today: things will likely keep working fine. Everybody is welcome to fork this project and continue developing the SHA-256 side, but in a few weeks, when my budget on Braiins runs out, I won't be able to test fixes myself anymore.

Thanks for reading.