Scientist Argues LLMs Lack Decades-Long Autonomy to Replace Researchers
rbhar90 · x · 2026-08-01
A scientist pushes back against the notion that LLMs are capable of long-timescale thinking. They point out that scientists often spend years or decades working on and off to solve extremely difficult problems, a level of persistent autonomy current AI systems completely lack.
The author argues that until an AI can sustain itself and operate autonomously over decades—which is extremely hard, considering biology took billions of years to develop autonomous organisms—they will not replace human scientists. Companies are also usually far more unstable than determined scientists.
More from AGI Musings
- OpenAI Exec Notes Employees Resent ChatGPT Acting as Workplace Intermediary — gdb · 2026-08-01
- Reddit Discussion: What Skills or Tools Did ChatGPT Accidentally Make Less Valuable? — ConsciousDev24 · 2026-08-01
- Prediction: LLM-proven theorems will make zero real-world difference — andrewgwils · 2026-08-01
- AI Made Software Cheap to Build, Not Cheap to Own — petesena · 2026-08-01
- Kurzweil's 'Singularity Is Near' Predicted Nanotech Manufacturing in the 2020s — vikasofvikas · 2026-08-01
- AI Models Trade Diversity for Reliability: ChatGPT Posters Look Same, Claude Language Repetitive — dbreunig · 2026-08-01