logoalt Hacker News

abeppu • today at 4:38 PM • 2 replies • view on HN

What is reasoning if examining a problem and constructing a program to solve it doesn't count? Changing the original represention into one that is more familiar (particular sequences of tokens) seems like it is all about "intrinsic ability".

I'm not saying they don't have a point, but I think this is actually not great evidence.


Replies

thelaxiankey • today at 5:05 PM

There's a distinction between reasoning with access to a turing machine and reasoning without access to a turing machine. In one case, you must sometimes become a turing machine yourself and execute "algorithms", which is something LLMs can struggle with.

When I was learning combinatorics, I was aggressively discouraged from using induction to prove things. Not because it was bad, or because it didn't work, but because if I used induction, I wouldn't be practicing 'combinatorics', I'd be practicing induction. You can make similar observations about solving ODEs numerically, or using mathematica for all your integrals.

One way to think of this is that banning coding prevents the agent from outsourcing some of the skills it's bad at (discrete sequential reasoning) to a computer. We are grading precisely the thing it's outsourcing because we suspect improving that might make it better.

freecodeio • today at 6:25 PM

someone is teaching a kid to do multiplication and removed the calculator, why are you so surprised with this concept?