Pools centralizing isn't the problem. Polls controlling what the miners mine, *since* they centralize, is the far more concerning issue.
In other words, if a pool couldn't attack the network even if they have 60% of the hash power, then the existence of the pool isn't that big of a deal. This is the point of DATUM.
This is a feature that very few hashers are seeking, as it increases the complexity of running a mining operation.
The only incentive for hashers to take on the task of block construction is altruism at a very low time preference.
The only way to really get this resolved would be to make it super easy at the firmware level of the ASIC server, and that would pose a non-trivial overhaul of mining architecture.
What’s being proposed is all increase in overhead with minimal payback on a reasonable time frame, let’s say 5 years.
