Just listened to Paul Sztorx's talk on CUSF at the opnext.dev conference.

It is the worst technical concept I've ever seen presented at a Bitcoin conference. 😂

tl;dr: because no-one wants to merge his terrible soft fork proposals – and he's too lazy to rebase his code – Sztorc wants to ship a separate piece of software that uses the invalidateblock RPC call to detect and reporg out blocks that violate his soft forks. Obviously, if miners actually run this, it makes it a nightmare to actually run a real validating node as now this Rube Goldberg bit of complexity now becomes the Bitcoin protocol. And obviously, that makes it likely that a soft fork won't be enforced by economic nodes, resulting in massive losses when miners get bored and stop running it. 

 Read a “challenges” paper that raises questions about Paul’s approach. 

 I ain’t no genius, but dat dere sounds über dumb. 

I have a better idea, how about Paul runs a master node and everyone asks Paul if certain transactions are valid or not. 

 "ship a separate piece of software" - so is he forking Bitcoin, or just randomly filtering mempool for his own miner like those green energy certificate folks?