Jabrov

6 ms·

31.
by Jabrov 5mo ago
I think part of the problem is that a lot of designers are simply not that good at what they do
32.
by Jabrov 5mo ago
Yes multiple GPUs absolutely help with inference even for a single model instance. Some models are simply too big to fit on the largest available GPU. Check out tensor parallelism
33.
by Jabrov 6mo ago
I can’t wait until the distant future where strategy games will have actually good and interesting AI that can communicate and reason
34.
by Jabrov 6mo ago
Such a mistake should never happen, but it's not even about the mistake. It's more about how absolutely awful their support is to revert the mistake.
35.
by Jabrov 6mo ago
Holy. Time to leave GoDaddy. What's the best alternative?
36.
by Jabrov 6mo ago
Kimi is the base, but they've done tons of finetuning on top to produce a really good completions model.
37.
by Jabrov 7mo ago
What a super neat application of computer vision. Cool writeup. Thanks for sharing the code and making it open source too!
38.
by Jabrov 7mo ago
Can't you just make it opt-in? No? Because no one would opt-in, you say? Wow. It's almost like this is a user-hostile feature that breaks the implicit promise behind a "private" repo.
39.
by Jabrov 7mo ago
The lack of GenAI integration is actually refreshing
40.
by Jabrov 7mo ago
RuneScape millionaires?
41.
by Jabrov 8mo ago
I do the same thing, but with a Markdown file which I add a section to every day in a roughly append-only fashion
42.
by Jabrov 8mo ago
What jurisdiction does a data center in space fall under, anyway? The one of the nation that launched it?
43.
by Jabrov 9mo ago
In the article it says “ While most of the killings were carried out by IRGC and Basij forces, reports received by Iran International indicate that proxy forces from Iraq and Syria were also used in the crackdown. The deployment of non-loca
44.
by Jabrov 9mo ago
A lot of it is being done by mercenaries brought in from Afghanistan and Iraq
45.
by Jabrov 9mo ago
What's wrong with VLC?
46.
by Jabrov 10mo ago
Lmao this will not work
47.
by Jabrov 10mo ago
You're absolutely right!
48.
by Jabrov 11mo ago
Ehhh just calling a raw LLM is not going to replace anyone and be prone to hallucination, sure. But lawyers are increasingly using LLM systems, and there's law-specific products that are heavily grounded (ie. they can only respond from
49.
by Jabrov 11mo ago
I don't know if that's so much a mistake as it is ambiguity though? To me, using the viewer's perspective in this case seems totally reasonable. Does it still use the viewer's perspective if the prompt specifies "Pu
50.
by Jabrov 11mo ago
Yes casual racism is very amusing
51.
by Jabrov 1y ago
More sites should be like this
52.
by Jabrov 1y ago
No one’s saying anything about it being a cause though … association is not cause
53.
by Jabrov 1y ago
What makes you think it’s a deferrable workload? Companies don’t buy all that expensive hardware to just have it sit there inactive.
54.
by Jabrov 1y ago
Interplanetary colonization is a much dumber idea than manned spaceflight
55.
by Jabrov 1y ago
This is the best comment I’ve ever seen
56.
by Jabrov 1y ago
What demographic? Do you have a link to or some additional information about the data source? Or are you just speculating?
57.
by Jabrov 1y ago
Has there ever been one that survived for a really long time?
58.
by Jabrov 1y ago
Not exactly true ... KV and prompt caching is a thing
59.
by Jabrov 1y ago
The latest versions of Python 3 already have or can do most of these things.
60.
by Jabrov 1y ago
Very interesting methodology, but the sample size (16) is way too low. Would love to see this repeated with more participants.

More ›