a16z GP: agents can't 'make me a million dollars' — human direction still unlocks the next hill
lennysan · x · 2026-09-09
Lenny Rachitsky shared highlights from his podcast with a16z GP, pushing back on Silicon Valley's "permanent underclass" narrative:
- Agent limits: Anish Acharya explains why "Hey Claude, make me a million dollars, make no mistakes" fails — agents hill-climb to a plateau, but humans are still needed to envision the bigger unlock; nothing in the tech shows otherwise.
- Jobs story hasn't played out: things are better than ever by most measures; radiologists have been "cooked" for 20 years yet demand keeps rising; open engineering roles at all-time highs.
- Market structure: unlike mobile's winner-take-all, every AI layer has 20 strong competitors; moats are discovered, not designed.
- Other topics: staying current without being chronically online, the "/loop make me happier" opportunity, and companies becoming series of loops.
More from AGI Musings
- Monasteries for human skills: keeping math and coding alive without AI — toptickcrypto · 2026-09-09
- Two years from o1-preview to superhuman math: RL scaling now cracks open research problems — jam3scampbell · 2026-09-09
- A plea to lab researchers: don't launch a superintelligent RL run without understanding its mind — peterwildeford · 2026-09-09
- Inside-lab take: OpenAI hasn't internalized the stakes, Anthropic understands but is racing anyway — peterwildeford · 2026-09-09
- Critics see déjà vu as Zuckerberg pitches only AI's positive future — krishnan · 2026-09-09
- Percy Liang's CS336 slide: open models are what make teaching LLMs from scratch possible — stanfordnlp · 2026-09-09