Hacker Newsnew | past | comments | ask | show | jobs | submit | snats's commentslogin

personal website: snats.xyz weblog: weblog.snats.xyz

I also did a couple of experiments with pruning LLMs[1] using genetic algorithms and you can just keep removing a surprising amount of layers in big models before they start to have a stroke.

[1]https://snats.xyz/pages/articles/pruningg.html


I suspect this applies to human beings as well.


There's https://en.wikipedia.org/wiki/Hydrocephalus# and there are cases of people living normal lives, not realizing they're missing most of their brains, until this gets discovered on some unrelated medical test. Or people who survived an unplanned brain surgery by rebar or bullet. Etc.


Well, humans don't have “layers” in the first place…


Of course not, that’s ogres.


i get it. i want one of those. the problem is that most cellphones are not actual cellphones, they are entertainment machines. they are a pocket tv / social media feed place. most usage for my normal friends is for that.


I am working on the https://moviemovie.club/about, it's a tiny website about film review.

It works like a run club, where you have to make a review first to see other people's reviews.

I am currently implementing watchlists, comments and a mural to make it feel a bit less lonely. Right now I like the UI but it feels to lonely.


This seems like it would only work if “reviews” would be something rare to come by. Like some forums where you have to contribute to be able to download attachments, or see higher level subforums.

But reviews are everywhere, good ones too so it will be a hard chicken egg problem to solve.


Not an insider but imo the work on diffusion language models like LLaDA is really exciting. It's pretty obvious that LLMs are good but they are pretty slow. And in a world where people want agents you want a lot of the time something that might not be that smart but is capable of going really fast + searches fast. You only need to solve search in a specific domain for most agents. You don't need to solve the entire knowledge of human history in a single set of weights


It's pretty funny to test in-distribution for AI models. But they fail horribly once you push them a bit[1].

I recently made LLMs play Minesweeper and ALL LLMs that I tested had a pretty bad win to loose ratio. Like the only model that won more than 3 times was R1 (mind you there were 50 games).

[1] https://snats.xyz/pages/articles/minesweeper_bench.html


yup, if i went to do a PhD interpretability is the only interesting subject for academia IMO right now


It's more of a distilled model, not a fair 1:1 comparison


I use .XYZ because it was pretty cheap when I bought it


I use .xyz because I have a very common first and last name, and nearly all of the permutations of them were taken on .net, .com, .org, and .us; .xyz seems to price based on how desirable they think the name is, so I still couldn't get $FIRST-$LAST.xyz for a reasonable price, but I got something close.


I do too, aesthetically it's great. Unfortunately the rise in phishing from xyz domains means if you use it to send email your deliverability is likely to suck.


I post about ML/DL and SWE at https://bsky.app/profile/snats.xyz


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: