And we also have a "catastrophic" problem with PoW:
No, it you read what I wrote it was a
hypothetical, modified version of AES which does not diffuse to all of the output bits, which is specifically not what real AES is designed to do. No one has identified an actual problem.
Ok, let's make it more practical then. Citing the experts below:
As I said above, no one has identified an actual problem. His concerns are theoretical and other experts disagree. No one has identified an actual problem.
EDIT: I'd add that the DoS issue is a potential problem, but is somewhat mitigated by the optimizing of the algorithm and the code that occurred. Initially (when that comment was written), it took hundreds of milliseconds to verify a hash which is indeed a lot of time and a big DoS vulnerability. These days with the optimized hash code and multithreaded verification, the effective time is well under 10 ms, which is close enough to network latency to serve as an effective throttle on DoS.