Dev warns against chatting with LLMs: 'your sense of self atrophies the more you do it'
DionysianAgent · x · 2026-10-30
An AI developer says he codes with LLMs but hasn't used one conversationally in years. He argues there's a fallacy in spending too much time talking to LLMs: people lose their sense of self, think less independently, and lean on the model like a pocket calculator. He insists his insights come from his own mind and would only revisit LLM conversation for testing or training his own models.
More from AGI Musings
- Essay on technological progress and the meaning of work gets an FT shout-out — soumitrashukla9 · 2026-10-30
- New essay traces work and identity from 1760s weavers to 2026 mathematicians — soumitrashukla9 · 2026-10-30
- Genie 4 sparks debate: robots' GPT moment nears, home humanoid adoption doesn't — FuSheng_0306 · 2026-10-30
- Distillation debate: is Chinese AI pricing 'involution' or rational catch-up? — teortaxesTex · 2026-10-30
- Proton CEO calls security pentest requests 'loser thinking' in AI push — steipete · 2026-10-30
- Investor Stewart Alsop III Slams the 'Yapper vs Builder' Dichotomy as Fake — StewartalsopIII · 2026-10-30