Thanks heaps for the detailed reply! The BigO complexity analysis 'for free' sounds pretty interesting.

I'm wondering if Co-dfns would be a natural fit for an algorithmic design of programs through deep learning. We're already seeing some attempts at using an AlphaGo like architecture (marrying traditional AI such as monte carlo tree search with modern deep learning to learn useful heuristics, or in the case of a recent paper on program design I can't find right now, they combined the network with an SMT solver).

If the compiler naturally gives the complexity of a program, creating a feedback loop with a speed prior ( http://people.idsia.ch/~juergen/speedprior.html ) might work very well for AGI.

Sorry if this far out speculation is off topic, I'm just really keen on exploring tech that could be valuable in AI development.