Economist after Anthropic visit: recursive self-improvement is close, lab arms races leave everyone worse off
paulnovosad · x · 2026-10-16
Economist Arpit (@arpitrage) shares takeaways from spending days at Anthropic:
- RSI is near: models building models may trigger rapid takeoff scenarios with both risks and benefits. Evidence includes sharply higher code output per Anthropic employee and AI handling a growing share of model R&D — a challenging transition for humanity to manage.
- Arms races between AI labs leave everyone worse off: labs should invest in safety, avoid distillation risk, and cooperate on safety, but the frontier-model race prevents it.
The thread was boosted by economist Adam Ozimek (@ModeledBehavior), who called Arpit one of the best and wisest economists alive.
Related event: Anthropic Believes Recursive Self-Improvement Is Near, Visitor Reports(3 posts)→
More from AGI Musings
- Everyone has a precise AGI definition—until a model gets close to meeting it — VraserX · 2026-10-16
- Viral take: If AI gives us time back, its greatest gift is revealing what we mistook for the cost of living — r0ck3t23 · 2026-10-16
- "AI automation = slop, AI augmentation = super intelligence" — the line CEOs should see — jacob_posel · 2026-10-16
- Reddit take: Building a successor AI species would itself be humanity's purpose fulfilled — Flashy-Disk311 · 2026-10-16
- When Simple Rules Beat AI: A Frozen Ride-Hailing State Machine at Dulles — shashib · 2026-10-16
- Nobel laureate Jennifer Doudna on AI's role in science in Big Think interview — ChiaraStellata · 2026-10-16