[bastawhiz]

I expect the answer is "no". I wouldn't expect a tool like this to "discover" assembly without being trained on the compiled output. The model has no notion of how or where the code runs.

After decades of compiler research and super compilers chugging away, we're sort of at a point where discovering novel optimizations with results that are more than a smidge of improvement is almost impossibly unlikely. Compilers today are really good.

That said, I think the value that something like this might have is being able to optimize the intent of the code. If it can determine that I'm sorting some numbers, it can rewrite my code to use a faster sorting algorithm that has the same functional properties. If I'm storing data that never gets used, it can stop storing it. It has a view of the code at a level above what the compiler sees, with an understanding not just of what is being done, but why.

[xavxav]

> After decades of compiler research and super compilers chugging away, we're sort of at a point where discovering novel optimizations with results that are more than a smidge of improvement is almost impossibly unlikely. Compilers today are really good.

I agree when it comes to peephole optimizations, but there's still a lot of juice left in exploiting language guarantees (immutability, non-aliasing, data-parallelism), however most compiler developer energy is spent propping up C/C++ and consequently optimizations are developed with those languages in mind.