shash42's submissions

1.		Meta releases open data to train General AI Co-Scientists (huggingface.co)
		1 point by shash42 19 days ago \| past
2.		Training AI Co-Scientists Using Rubric Rewards [Meta Superintelligence Labs] (alphaxiv.org)
		1 point by shash42 24 days ago \| past
3.		Next general training environment for superintelligence? (shash42.substack.com)
		2 points by shash42 63 days ago \| past \| 1 comment
4.		Meta Superintelligence Labs Presents: Compute as Teacher (twitter.com/dulhanjay)
		4 points by shash42 4 months ago \| past \| 2 comments
5.		The Illusion of Diminishing Returns: Measuring Long Horizon Execution in LLMs (arxiv.org)
		3 points by shash42 4 months ago \| past \| 1 comment
6.		Counting Down Capabilities to AGI (shash42.substack.com)
		1 point by shash42 6 months ago \| past \| 1 comment
7.		Log-Linear Scaling Is Economically Rational (lesswrong.com)
		1 point by shash42 9 months ago \| past
8.		Great Minds Think Alike and This Undermines AI Oversight (model-similarity.github.io)
		3 points by shash42 11 months ago \| past \| 1 comment