npmipg

7 ms·

1.
Install.md: A standard for LLM-executable installation(mintlify.com)
109 points by npmipg 9mo ago|118 comments
2.
Anthropic, Meta, and Snap are paying up to 350k+ base for a DevRel(devreljob.com)
3 points by npmipg 1y ago|1 comments
3.
by npmipg 1y ago
Devrel salaries appear to be growing very rapidly. This was unheard of just a few years ago. It's possible it's because previously these roles would be called 'Head of Product Marketing,' and have just been rebranded as
4.
by npmipg 1y ago
lmao yes but the language is interesting
5.
Decoding an SF Craigslist "furniture" listing that appears to be a coded drug ad(twitter.com)
4 points by npmipg 1y ago|2 comments
6.
Fast Cheap Image Captioning Model Trained on Video Frames(cliptagger.inference.net)
1 points by npmipg 1y ago|1 comments
7.
by npmipg 1y ago
We trained a frame captioning model that's 2x faster and 17x cheaper than Claude 4 Sonnet
8.
OS 12B model that beats Claude 4 Sonnet at video captioning and is 17x cheaper(old.reddit.com)
2 points by npmipg 1y ago|1 comments
9.
by npmipg 1y ago
hf: https://huggingface.co/inference-net/ClipTagger-12b blog post: https://inference.net/blog/cliptagger-12b docs: https://docs.inference.net/use-cases/video-understanding se
10.
by npmipg 1y ago
Note that distilling a general model is several orders of magnitude more expensive than distilling a task-specific model, which is what I'm trying to promote here. Smart general models make distilling great task specific models with no
11.
by npmipg 1y ago
Hey, I'm the author of the post. The image has been fixed, and the point I'm making is that proprietary models are almost always ahead, and this gap is widening. OS models that are nearly at the same quality are usually distilled
12.
GPU-rich labs have won: What's left for the rest of us is distillation(inference.net)
88 points by npmipg 1y ago|50 comments
13.
The Economics of Hosting Open Source Models(inference.net)
1 points by npmipg 1y ago|0 comments
14.
Smart Routing Saved Exa 90% on LLM Costs(inference.net)
1 points by npmipg 1y ago|0 comments
15.
The Cheapest LLM Call Is the One You Don't Await(inference.net)
1 points by npmipg 1y ago|0 comments
16.
by npmipg 1y ago
working on this as we speak!
17.
1T Tokens for Sale(twitter.com)
2 points by npmipg 1y ago|0 comments
18.
Late Chunking vs. Contextual Retrieval: The Math Behind RAG's Context Problem(gist.github.com)
2 points by npmipg 2y ago|0 comments
19.
Show HN: Open Source 12ft.io Alternative With Free AI Summaries(smry.ai)
7 points by npmipg 3y ago|0 comments
20.
Show HN: SMRY, AI-enhanced open source 12ft.io and removepaywall.com alternative(smry.ai)
10 points by npmipg 3y ago|0 comments
21.
Show HN: Slang Thesaurus – An AI-Powered Slang Thesaurus Trained on UrbanDict(slangthesaurus.com)
4 points by npmipg 3y ago|0 comments