A software expert argues LLMs help most when users can spot the bullshit
mariofilhoml · x · 2026-07-29
The thread argues that LLMs are most useful when the user already knows the domain well enough to spot mistakes.
The author compares software and medicine, saying a doctor friend enjoys LLMs because they know when to call bullshit, while the author sees the same pattern in software: expertise makes the model more useful because it is easier to judge when it is wrong.
More from AGI Musings
- Gary Marcus says CEO “singularity” claims are hype, not evidence of a real AI milestone — anilkseth · 2026-07-29
- Essay argues AGI is not the Singularity, and the real question is the world it enters — theomitsa · 2026-07-29
- Voice-first agent prompting will be the next platform shift, poster says — wordgrammer · 2026-07-29
- AI Synthetic Empathy Threatens Human Relationships — PierceLilholt · 2026-07-29
- Viewpoint: Real-time Multimodal AI Will Enable Superhuman Communication — ctjlewis · 2026-07-29
- Doubling Time at Technological Maturity: Hours or Days, Not Years — EigenGender · 2026-07-29