Submissions from dwarkesh.com

		Pretraining progress is mostly coming from data (dwarkesh.com)
		2 points by crescit_eundo 2 days ago \| past \| discuss
		Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face (dwarkesh.com)
		3 points by yurivish 9 days ago \| past \| discuss
		Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face (dwarkesh.com)
		4 points by swolpers 9 days ago \| past \| discuss
		The Rise and Fall of Agent Civilizations (dwarkesh.com)
		277 points by consumer451 12 days ago \| past \| 197 comments
		Dylan Patel – Anthropic and OpenAI will have most of world’s compute by 2028 (dwarkesh.com)
		3 points by crescit_eundo 16 days ago \| past
		Why compute might get 10x+ more expensive in coming years (dwarkesh.com)
		2 points by rzk 25 days ago \| past
		Predictions for the Era of Continual Learning (dwarkesh.com)
		2 points by ronfriedhaber 27 days ago \| past
		Predictions for the Era of Continual Learning (dwarkesh.com)
		3 points by in-silico 30 days ago \| past
		Human level AIs might build runaway superintelligences by 2032 (dwarkesh.com)
		3 points by crescit_eundo 30 days ago \| past
		Predictions for the Era of Continual Learning (dwarkesh.com)
		3 points by gmays 34 days ago \| past
		Why smarter AI models could drive up compute prices 10x (dwarkesh.com)
		2 points by mefengl 36 days ago \| past \| 2 comments
		Why compute might get 10x+ more expensive in coming years (dwarkesh.com)
		3 points by nlpnerd 42 days ago \| past
		Why compute might get 10x+ more expensive in coming years (dwarkesh.com)
		4 points by cubefox 43 days ago \| past
		The Winning Essays for the Big Questions About AI (dwarkesh.com)
		1 point by herbertl 62 days ago \| past
		The Winning Essays for the Big Questions About AI (dwarkesh.com)
		2 points by gmays 64 days ago \| past
		Grant Sanderson – AI and the future of math (dwarkesh.com)
		1 point by crescit_eundo 72 days ago \| past
		The next big breakthrough will be AIs learning on the job (dwarkesh.com)
		3 points by gmays 72 days ago \| past \| 2 comments
		The data black hole at the center of AI (dwarkesh.com)
		4 points by jedixit 83 days ago \| past
		What remains scarce after AGI? (dwarkesh.com)
		2 points by gmays 3 months ago \| past
		The sample efficiency black hole (dwarkesh.com)
		9 points by crescit_eundo 3 months ago \| past \| 1 comment
		Alex Imas and Phil Trammell – What Remains Scarce After AGI? (dwarkesh.com)
		1 point by crescit_eundo 3 months ago \| past
		Reiner Pope – Chip design from the bottom up (dwarkesh.com)
		2 points by crescit_eundo 3 months ago \| past
		The mistake of conflating intelligence and power (dwarkesh.com)
		3 points by gmays 3 months ago \| past \| 4 comments
		Notes on pretraining parallelisms and failed training runs (dwarkesh.com)
		3 points by gmays 3 months ago \| past
		The mistake of conflating intelligence and power (dwarkesh.com)
		3 points by crescit_eundo 3 months ago \| past \| 1 comment
		Eric Jang – Building AlphaGo from Scratch (dwarkesh.com)
		3 points by crescit_eundo 3 months ago \| past
		The math behind how LLMs are trained and served (dwarkesh.com)
		1 point by stopachka 4 months ago \| past
		The math behind how LLMs are trained and served (dwarkesh.com)
		6 points by gmays 4 months ago \| past
		Blog prize for big questions about AI (dwarkesh.com)
		3 points by gmays 4 months ago \| past
		Dwarkesh Patel / Jensen Huang – TPU competition, why we should sell chips to Chi (dwarkesh.com)
		5 points by Poleris 4 months ago \| past
		More