sheeshkebab

5 ms·

1.
by sheeshkebab 11d ago
so we finally got our "phd level ai" maybe, semi consistently move blocks around at the level of a 1 year old child. good fucking job everyone, congrats.
2.
by sheeshkebab 28d ago
writing native gui code is atrocious, add to that llms being horrible with ui dev and you end up with with what you describe. but at least if we get more non-ui stuff written in more efficient languages there is going to be some benefit.
3.
by sheeshkebab 29d ago
llms make it trivial writing code in rust, without even knowing rust. There is hope…
4.
by sheeshkebab 1mo ago
It’s pretty bad except that any other way is even worse.
5.
by sheeshkebab 1mo ago
This. The same goes for “skills”, skill type “subagents” and other bullshit - powerful models don’t need any of that anymore I noticed.
6.
by sheeshkebab 1mo ago
The idea that agi would help us cure all the human deceases, esp house built into our generics or developed over millions of years gotta be the top indicator of our own stupidity.
7.
by sheeshkebab 1mo ago
What? psychologists are for that (or drugs and alcohol).
8.
by sheeshkebab 1mo ago
What’s with fascination with agi/superintelligence? It seems people working on it never had kids and just want to compensate for that. It’s a really horrible thing to try to “solve”.
9.
by sheeshkebab 2mo ago
Treasuries are $ denominated -> sell them for $$, buy yen back, yen/usd drops. or have Uncle Sam buy yen.
10.
by sheeshkebab 2mo ago
good.
11.
by sheeshkebab 2mo ago
Try doing a refactoring of some sort or larger new feature using just an agent on a moderately sized codebase, 256k will be compacting every few minutes, and result will be unusable.
12.
by sheeshkebab 2mo ago
My cat is more conscious than whatever latest frontier model is out there, and it runs on a little bit kibble and water.
13.
by sheeshkebab 2mo ago
Barrels of tokens sold by thieves that copied every book sure looks like selling to me.
14.
by sheeshkebab 2mo ago
Someone going to public library, copying every book out of it and then selling them as their own are thieves.
15.
by sheeshkebab 2mo ago
there are plenty of orgs where writing more code is not a good thing, in fact it's the last thing they want. but yet these orgs would still employ 80%+ of all developers, not necessarily to write code though.
16.
by sheeshkebab 3mo ago
unless they launched 10t param models, or figured out some amazing new way to compress as many params into say 100b, I doubt it's anywhere near "mythos level". and I have no idea how many params mythos has but that was just s
17.
by sheeshkebab 3mo ago
it’s not been useful anywhere else - no self driving cars, no laundry helpers - just some idiotic animatronic carcasses that are barely able to walk around, and semi autonomous killer bots on ukraine frontlines.
18.
by sheeshkebab 3mo ago
Qwen models are actually very competitive with frontier models, and you can run them on your local computer. Gotta have a decent graphics card and by that time the current cost of the rig may not justify it over paying $100/month for c
19.
by sheeshkebab 3mo ago
Well, it was great while it lasted - I had fable build me a bunch of stuff this week that opus was just screwing up too much and could never finish. Good thing there are plenty of choices now even if US gov fucks up US AI.
20.
by sheeshkebab 3mo ago
Pro/max subs are not as flexible as bedrock in api use and don’t seem to run the same models either - often times they are notably dumber (quantized I guess) than bedrock equivalent.
21.
by sheeshkebab 3mo ago
I’ll ask it to write me some win32 ui crap when I get hands on it, it will need all its brainpower to get that idiocy right.
22.
by sheeshkebab 3mo ago
Opus regularly bitches and wines to me how long something will take and that I should think before asking it to do it. But then it does it anyway in 15 minutes.
23.
by sheeshkebab 3mo ago
considering they work with any architecture/configuration given enough compute, just more or less efficiently - then maybe it's fundamental, in the same sense as why electricity works...
24.
by sheeshkebab 4mo ago
I like gpt oss - great model even if not too smart.. runs on my laptop at over 100ts has a certain tone that I like over all these qwens stuck up their asses.
25.
by sheeshkebab 4mo ago
27b is slow as molasses vs 35b on local stuff I have (m5 max). Mtp doesn’t make any difference either.
26.
by sheeshkebab 4mo ago
…but they reason well enough given enough context (using their matmuls).
27.
by sheeshkebab 4mo ago
One of these days you’ll learn about “enterprise” code
28.
by sheeshkebab 4mo ago
Mist of human written codebases are unusable for llm dev by that definition.
29.
by sheeshkebab 4mo ago
It’s great laptop to mess around with llms, it won’t replace claude opus or even sonnet.
30.
by sheeshkebab 4mo ago
You can only run heavily quantized models on all 3/4/5 rtx gpus (with 32gb or less vram) - and you probably want moe versions like Qwen 35b for this to run at speed somewhat comparable to Claude. It’s still not there to be honest

More ›