More responsive indefinite loops with blockchain

💡 Ideas
1

It may be possible, with blockchain technology, to detect and mend indefinite loops:

The state of every relevant hardware register could be hashed, as for example, into an MD5; and these subsequent hashes --
if compiled into a blockchain, could model the system's entropy -- which, in turn, could be used to measure, detect, and reset duplicate operating states

Iff possible, could this basic analysis method be written into the specification ?

2

Look up the halting problem. You won't be able to detect this in the general case. Also, in the restricted memory case, you're almost certainly going to face super-exponential complexity.

Also, what practical wins would this bring that killing a tab or process the one-off time it happens doesn't?