BIP110ers are super cerealBIP110ers are super cereal

You may have seen Chris Guida's post a few days ago where he noted that he had prepared Luke Dashjr's older proof-of-work-change code for use "just in case miners decide to collude to betray bitcoin." #1540163

Although the move was quickly seconded by most of the bip110 crew (Mechanic, Dathon Ohm, Kratter, and Dashjr himself), Dathon assured me that the code would only be used in an "unlikely contingency"

What's the point of changing the proof of work algorithm?What's the point of changing the proof of work algorithm?

Bitcoin mining relies on a specific math function (SHA256). Bitcoin miners use circuit boards designed to compute SHA256 on 80 byte Bitcoin block headers and they are pretty good at it, but they are not so good at computing other functions. So, if a chain wanted to stop all the large-scale miners from attacking it, they might decide to use a different function to compute the proof of work for new blocks.

One possible alternative to SHA256 could be RIPEMD160. Large-scale Bitcoin miners who had invested lots of money in ASICs would not be in a very good position to mine blocks on a blockchain requiring proof of work via RIPEMD160. This is what people mean when they refer to "firing the miners."

Again, the BIP110ers assure me that this is not a likely outcome and that they are only preparing the code out of an abundance of caution and that they are super, super cereal.

Out of the pot and into the Spawn Camping Attack fireOut of the pot and into the Spawn Camping Attack fire

However, in the very surprising scenario where most of the Bitcoin world ignores BIP110 and they find themselves forked off on a small, minority chain with very little hashrate, it seems possible that they would opt to change the proof of work algorithm rather than be exposed to 51% attacks.

Should they make such a change, it is possible that they would be opening themselves up to a different sort of attack: one that Vitalik Buterin dubbed the Spawn Camping Attack.

Michael Tidwell described this attack as:

This is when there are no asics for a given mining algo, so any motivated attacker with large access to gpu's can just keep sniping the new chain with the new algo no matter how many times you change it.

You might say it's kind of like a powerful player camping out next to a spawn point in a FPS game and killing everyone the second they spawn.

Keeping in mind that GPUs seem to be easier (if not cheaper) to rent than they were back in 2017, you could see a world where the BIP110ers change their proof-of-work algo, and immediately get 51% attacked by whoever is willing to rent enough GPUs to do it. Jumping to a new algorithm (firing the miners) doesn't work anymore though because wihtout ASICs, all algos are equally equally efficient for all players.

There are probably other ways of dealing with this sort of attack, but it would be interesting to see if something like this plays out.