Yet another list of things you could read to learn about and stay up to date on AI

Potential initial reading

• Understand how AI works…
  • CSET’s The Surprising Power of Next Word Prediction: Large Language Models Explained: Part 1, Part 2, Part 3
  • Eight Things to Know about LLMs
    
• Understand AI threat models…
  • What risks does AI pose?
  • GPT-2030 and Catastrophic Drives: Four Vignettes
  • Racing through a minefield: the AI deployment problem
  • AI Safety Seems Hard to Measure
    
• Extra: Very optional! – The AI Safety Fundamentals Governance curriculum provides a great introduction to many of the problems IAPS aims to tackle. It’s likely too long of a list to go through without dedicating serious time to it, but it can be a useful static resource when you’re looking for readings about specific topics. Skim through the curriculum to have a sense of what’s there.

Ongoing

• Podcasts
  • Politico Tech: daily 20-min episodes on the latest tech policy news with a focus on the US and AI.
  • CSIS AI Policy Podcast: fortnightly 40-min episodes on international and US AI policy with a natsec angle.
  • Dwarkesh Podcast: in-depth (>2h) interviews with experts at the frontier of AI development with an industry focus.
  • Future of Life Institute (FLI) Podcast
  • Brookings Institute's Tech Tank Podcast
    
• Newsletters
  • Transformer Weekly Newsletter provides a weekly briefing on AI policy and happenings at the main AI companies.
  • Import AI is a weekly newsletter written by Jack Clark, cofounder and head of policy at Anthropic. It offers synopses of meaningful events in AI policy, and a wonky glimpse into corporate governance and government relations.
  • The Simon Institute’s monthly newsletter on AI governance, especially international governance.
  • Center for AI Policy newsletter
  • Center for AI Safety newsletter
  • ImportAI, by Jack Clark, who is the head of Policy for Anthropic
  • NavigatingAI by SaferAI
  • Zvi’s newsletter is a bit verbose but very informative as a distillation of technical AI safety research – usually worth a skim and often a great reference if you want to follow up on something specific.
  • NextGov can also be good but the occasional signal is drowned out by high noise - better to have someone scan it for you I think!
    
• If you’re into Twitter
  • I have an AI governance list and an AI tech list.
    
• If you're in to books
  • Chip War: The Fight for the World's Most Critical Technology
  • The New Fire: War, Peace, and Democracy in the Age of AI
  • The Coming Wave: Technology, Power, and the Twenty-first Century's Greatest Dilemma
  • The Alignment Problem: Machine Learning and Human Values