Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
pajop
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
31.
▲
Compare 75 AI Models on 200 Prompts Side by Side
(aimodelreview.com)
18 points
by
pajop
2y ago
|
3 comments
32.
▲
What do you have "unreasonably" high standards for?
(twitter.com)
2 points
by
pajop
2y ago
|
0 comments
33.
▲
Transformer by hand 5-min tutorial by Anna Rahn
(twitter.com)
1 points
by
pajop
2y ago
|
0 comments
34.
▲
MuMath-Code on ArXiv
(arxiv.org)
2 points
by
pajop
2y ago
|
0 comments
35.
▲
Gen AI: Too much spend, too little benefit [pdf]
(goldmansachs.com)
48 points
by
pajop
2y ago
|
30 comments
36.
▲
by
pajop
3y ago
Here you go: https://github.com/quassnoi/explain-extended-2024
37.
▲
Studying Bias in GANs Through the Lens of Race
(arxiv.org)
1 points
by
pajop
3y ago
|
0 comments
38.
▲
by
pajop
3y ago
Interesting commentary on the video: https://x.com/sumlenny/status/1749870142547255525 What if they revise and re-issue the video based on the comments? The alternate titles suggested are also interesting: “ A goo
39.
▲
Is there no way to just post an index.html page somewhere for free?
(twitter.com)
2 points
by
pajop
3y ago
|
4 comments
40.
▲
Mini-Canvas-Editor Example
(img-js.github.io)
1 points
by
pajop
3y ago
|
0 comments
41.
▲
Jennifer Hochschild walks back on comments about Harvard Extension School
(twitter.com)
1 points
by
pajop
3y ago
|
0 comments
42.
▲
by
pajop
3y ago
from the paper: https://arxiv.org/pdf/2310.08659.pdf "LoftQ aims to solve the problem of the discrepancy between the quantized and full-precision model in the context of quantization and LoRA fine-tuning for Large
43.
▲
LoftQ: LoRA-fine-tuning-aware Quantization
(github.com)
3 points
by
pajop
3y ago
|
4 comments
44.
▲
A tilt toward safety stopped a scientist's virus research
(npr.org)
2 points
by
pajop
3y ago
|
0 comments
45.
▲
Math Models from Program Search (Google FunSearch) Using LLMs
(nature.com)
4 points
by
pajop
3y ago
|
0 comments
46.
▲
Google DeepMind used an LLM to solve an unsolvable math problem
(technologyreview.com)
7 points
by
pajop
3y ago
|
0 comments
47.
▲
Martin Shkreli Twitter Spaces on SBC Going to MDC Brooklyn Jail
(twitter.com)
3 points
by
pajop
3y ago
|
0 comments
48.
▲
by
pajop
3y ago
see also: https://duostories.org/tok-en
49.
▲
Plausible theoretical foundation for LK99 RTAPS calculated via supercomputer
(arxiv-vanity.com)
17 points
by
pajop
3y ago
|
3 comments
50.
▲
by
pajop
3y ago
Didn’t Someone solve 1LLL already? https://www.iflscience.com/the-full-1lll-a-rubiks-cube-holy-... https://www.reddit.com/r/Cubers/comments/whuhkq/i_learned_fu...
51.
▲
Toward Actionable Generative AI
(blog.salesforceairesearch.com)
2 points
by
pajop
3y ago
|
0 comments
52.
▲
Enabling cross-chain swaps from any token to any chain with just one click
(devfolio.co)
1 points
by
pajop
3y ago
|
0 comments
53.
▲
Meta Releases MusicGen
(huggingface.co)
1 points
by
pajop
3y ago
|
0 comments
54.
▲
Falcon on Huggingface.co
(huggingface.co)
1 points
by
pajop
3y ago
|
0 comments
55.
▲
OpenAI’s plans according to Sam Altman
(web.archive.org)
2 points
by
pajop
3y ago
|
0 comments
56.
▲
The Next 50 Ideas to Shape the Future circa 2023
(ie.edu)
2 points
by
pajop
3y ago
|
0 comments
57.
▲
SF’s Most Competitive ‘Hacker House,’ Where Workers Will Eat/Sleep/Breathe Tech
(sfstandard.com)
1 points
by
pajop
4y ago
|
0 comments
58.
▲
The Legendary 24-hour Winter Marathon 2022 (chess)
(lichess.org)
1 points
by
pajop
4y ago
|
0 comments
59.
▲
Repo is unofficial ChatGPT API. It is based on Daniel Gross's WhatsApp GPT
(github.com)
2 points
by
pajop
4y ago
|
0 comments
60.
▲
Chrome extension to Automatically rotate through open tabs. By Ben Hedrington
(chrome.google.com)
1 points
by
pajop
4y ago
|
0 comments
More ›