Dario Amodei on Claude, AGI and the future of AI & humanity

Why it matters

A five-hour deep conversation with Anthropic CEO Dario Amodei (first half), plus segments with alignment researcher Amanda Askell and interpretability pioneer Chris Olah. Dario is the most systematic exponent of the scaling path: why models will keep improving, why safety is an engineering discipline rather than a slogan, and the 'compressed 21st century' vision from Machines of Loving Grace. The clearest account of how Anthropic's methodology genuinely differs from OpenAI's.

Key takeaways

• Scaling hasn't hit a wall: until models top out at things humans are good at, headroom remains.
• 'Powerful AI' may arrive 2026–2027: systems above Nobel-laureate level across most relevant fields.
• Safety as competitiveness: the Responsible Scaling Policy (ASL levels) turns caution into an engineering process.
• Mechanistic interpretability: we can already 'open' models and read features — understanding internals is alignment's core weapon.
• The optimistic future: if strong AI accelerates biology, economics and governance, the 21st century could compress into years.