I expect rapid progress but not towards general superintelligence

Argues the near-term bottleneck shifts from engineering execution to research ideas as coding agents master distributed GPU work, which affects how you plan around model progress.

• articleThe Cyber Risk Discourse is Broken
• articleDebating RSI, the US-China Gap, and Jaggedness with JS Denain of Epoch AI
• articleThe current balance of power in open models
• articleWhere I stand on RSI

• videoRyan Greenblatt – What happens once AI can automate AI research?Dwarkesh Patel
• articleThe twilight of the chatbotsEthan Mollick
• podcastRyan Greenblatt – Human level AIs might build runaway superintelligences by 2032Dwarkesh Patel

Checking sign-in…

Loading comments…