Hacker News Reader

Yes I contributed code to a paper on it this spring. It’s hard to measure and “studying” it just means feeding money into frontier models. If anyone has a few billion tokens for a couple masters students hmu so they can get it published.

View on HN

I'd be very interested to read this paper when it comes out.

I will add you to the list, I’ll definitely post it on HN also.

They had originally planned on june and my part is in so…. Me too.

I'm interested too - besides the paper, is it your personal experience that working with llms on ocaml is superior to other languages? I thought with the sheer amount of training data languages like python and java would be "ideal"

No it’s more about speed and quality of the feedback loop. So you want the compiler that catches the most stuff when given the least information.

Can it not be studied with open weight models? This is a genuinely curious question, not an attempt at scoring a point.

I don’t know it’s not my paper. I don’t want to talk shit but I’m also a little confused at this apparent limitation.

Reply on news.ycombinator.com