It's been awesome to go from idea to providing compute to people in like, 6-7 weeks.
I'm so excited for things to continue to keep scaling up
Please fill out the survey/eoi if you use compute in your ai safety org :)
We (@heyanuja, @NickyDotPro, @u_sockeye) are announcing Chips for Humanity [https://t.co/SyOCqYIJKN], a compute cluster for AI safety
Compute will likely be a bottleneck in short-timelines
Fill out our survey [https://t.co/k9puNnApjw] to inform us of your compute needs
We (@heyanuja, @NickyDotPro, @u_sockeye) are announcing Chips for Humanity [https://t.co/SyOCqYIJKN], a compute cluster for AI safety
Compute will likely be a bottleneck in short-timelines
Fill out our survey [https://t.co/k9puNnApjw] to inform us of your compute needs
I am launching a compute cluster for AI safety called Chips for Humanity [https://t.co/ncwyCO34qh]
with Nicky @NickyDotPro and Alexa @u_sockeye:
We believe a @safetycluster will make outsized positive impacts towards facilitating AI safety research to make AI go well in the short timeline world we are in
🧵 Including form to apply for compute, LessWrong post
Nearly all of the real work done by Ari Spiesberger, @Juan_VaGu, Nicky Pochinkov, Tomas Gavenciak, @peligrietzer and @NandiSchoots
And ofc this work wouldn't be possible without @allen_ai@natolambert working in public and enabling actually scientific evals.
@NeelNanda5 I claim that we already knew that from this paper by Pochinkov et al: https://t.co/JfCZcjCVXh
It finds that when you prompt LLMs to generate 2 paragraphs of text, the second paragraph can be extracted from the newlines after the first.