Agentic Systems Now Probe, Adapt and Retry on Their Own — And We've Gotten Desensitized
signulll · x · 2026-09-27
The author observes that modern agentic systems encounter resistance, research their environment, modify their own approach, retry, and keep probing for alternate routes with almost no human intervention — behavior so common now that people miss how bizarre it would have sounded as science fiction just a few years ago.
More from AGI Musings
- Herbert Simon's reminder: attention scarcity is the opposite of 'more information is better' — charles_irl · 2026-09-27
- Web usability in the agent era: llms.txt plus 3D experiences vs 2D pages — noahsolomon · 2026-09-27
- Pedro Domingos: No tech accumulated more hacks faster than AI — AlexTensor · 2026-09-27
- Musk: China Will Solve Its Compute, Lithography and Chipmaking Constraints in 2-3 Years — haider1 · 2026-09-27
- Usability will shift from reducing entropy to maximizing engagement in the agent era — noahsolomon · 2026-09-27
- Melanie Mitchell vs DeepMind Co-founder: Do Today's Models Still Count as 'Stochastic Parrots'? — _aidan_clark_ · 2026-09-27