MPI professor: I can advise 6 PhDs but only 2-4 AI agent projects at once
Michael_J_Black · x · 2026-10-18
Michael Black (Max Planck Institute) shares part 3 of his "PhD Advising vs Vibe Coding" series. He can comfortably supervise 6 PhD students and manage 8 while working hard, but struggles to advise 6 AI projects at once.
The reason is cadence: AI agents work fast and return with answers or questions needing his analysis multiple times a day. Meeting each agent weekly like a PhD student would defeat the efficiency gains. His conclusion: 2-4 concurrent AI projects is the sweet spot given his other responsibilities.
More from AGI Musings
- 21 Humans Now Have Neuralink Implants, Using Them 6.5 Hours a Day — Aiden_Tech_Ai · 2026-10-18
- Agentic AI Is Being Overbuilt, Overpriced, and Oversold, Argues New Essay — DavidLinthicum · 2026-10-18
- 1986 headline of teachers picketing calculators echoes today's AI debates — ExcitingImage9211 · 2026-10-18
- Teen solves years-old math problem in a week with AI and Lean, writes 15-page paper — IgorCarron · 2026-10-18
- GTA V on an iPhone and weekend vibe-coded games: the new piracy era — aigleeson · 2026-10-18
- South Park underpants-gnome meme skewers the missing step of AGI utopia — Puzzleheaded-King584 · 2026-10-18