AI Frontiers
Agents, alignment and superintelligence - what is known, and what is not.
- 1Agents are loops with toolsAn AI agent is a language model in a loop, allowed to call tools and read the results. That's powerful, and it opens three new ways to fail. Learn to spot them.Runs in your browser14 min+35 XP
- 2Getting what you asked forAI systems are very good at doing exactly what you measured, which isn't always what you meant. Meet specification gaming, see how chatbots are trained on human feedback, and why honest checks matter.Runs in your browser15 min+35 XP
- 3Superintelligence, honestlyWhat AGI and superintelligence mean, what scaling laws actually measured, and what thousands of AI researchers say about where this goes - including how much they disagree.Runs in your browser18 min+40 XP