subarctic · HackerTrans

3 時間前議論

I haven't seen it do it by itself that much so far. Even then, a lot of the time is spent waiting for some shell command to complete, which isn't inference. I guess being able to do parallel inference is like having multiple cores in a cpu, you can get by without it by just sharing that single core, and having it just gives you a performance boost under certain circumstances

subarctic

4 時間前議論

> Most useful LLM work is done in parallel

I guess what I'm doing is not considered that useful then? I usually only have zero, one, or occasionally two things actively doing inference at a time, be it claude code sessions or one of the chatgpt/claude web interfaces, and i bet that's true for like 95% of people using llms. And anyway i bet even the hardcore people using a bunch of parallel agents would appreciate having access to local, private inference for some things.

You're obviously right though that cloud inference isn't going away anytime soon

subarctic

昨日議論

You learn that you can build a version of Postgres that passes the tests and improves on these benchmarks in rust. You gain access to a codebase that does it which you can learn a lot from too. Definitely not zero like someone else said

subarctic

一昨日議論

Ya the second one should be Artificial Artificial AI. So close lol

subarctic

一昨日議論

What on earth, this is the first I've heard of this. I thought the EU was all about privacy?

subarctic

一昨日議論

I mean you can learn a lot. You can learn what's possible with less effort to build a proof of concept. It's kind of like you had another engineer do it for you, you don't completely learn how to do it yourself but you can still learn a lot with much less effort

subarctic

一昨日議論

I mean if it's actually just another Postgres, you don't have to worry about network effects as much because anyone could use it in place of Postgres. But sure, the more people you know using this, the safer it would feel to use it

subarctic

一昨日議論

Definitely a feature, but also ruins it for this person's use case

subarctic

一昨日議論

Agreed, if there was something actually better to switch to then I'd be interested. But seems unlikely for that to happen now - it's easier than ever to build a new github, at least the app itself but i doubt someone's gonna bother with the business effort needed to actually build it up as a reliable, trustworthy option that you know will be around for a while, which is a process that takes years when you know you're going to get disrupted. It would probably have to be open source to get early adopters to use it but somehow be nicer to use than GitHub, and there's basically no money to be made

subarctic

3 日前議論

That's how i feel about it too. But i haven't used it because there's no point in getting used to a new model if it's just gonna get taken away again. I wonder if they're also doing this to get some sort of signal on which types of users everyone is - for example someone else might always be chasing whatever the best model is

subarctic

5 日前議論

Never been happier that i bought a kobo instead of a kindle years ago

subarctic

5 日前議論

"Without ai assistance" - ok, but what about with ai assistance?

subarctic

5 日前議論

Artificial AI = stuff like mechanical turk where they get humans to do stuff computers can't do and make it look like it's "AI"

Artificial Artificial Intelligence = using computers to do mechanical turk jobs

subarctic

5 日前議論

Same, I basically assumed the community around it had died out. But i guess the sole maintainer/creator is still around.

I never used elm except for doing a tutorial, but lately I've built a full stack gleam app (using coding agents for the most part, with a lot of control in the beginning on the structure of the code) and have found that process works quite well

subarctic

5 日前議論

I'm confused because this snippet implies he's doing the trading himself, whereas later in the article it says it's done by a third party on his behalf that (he says) isn't informed by his personal knowledge. So it's not as clear but insider trading as is implied here

subarctic

9 日前議論

Definitely thought you meant claude opus but now from reading a couple other comments it sounds like you mean something else called opus?

subarctic

10 日前議論

I thought codex was open source https://github.com/openai/codex

subarctic

12 日前議論

I've re-read the comment again and I'm convinced it's not ai. It uses a pattern that ai adopted but it seems like an actual genuine comment. This is based on my spidey senses so it's not something I'm going to debate logically but it seems quite a few people agree with me. And I've seen lots of annoying ai-written stuff, and of course use Claude and Chatgpt regularly

subarctic

12 日前議論

Wow that sounds really dystopic... hope your joking

subarctic

12 日前議論

I kind of agree with that perspective, if you're testing them on something in the exam, why isn't the course material teaching it to them? After all you're paying a lot of money for tuition. I went to a good university where the assignments were hard, but as long as you did them the final exam ended up being easy without having to study for them because you already learned everything.

But i also did a semester abroad at a German university and i found there was a different philosophy there where the assignments don't teach you much, and everyone skips class and then studies really hard for the final exam which is often 100% of your grade

コメント