Reddit user asks: can you trust LLMs to teach you when they miss adjacent topics?
veggiesouptales · reddit · 2026-09-27
A non-coder who uses LLMs for Python and school projects asks whether prompting the AI to teach itself is a reliable way to learn best practices.
Their hesitation comes from repeated "scope laziness": models often overlook adjacent concepts unless explicitly pointed out — e.g. Claude ignored a listed app while troubleshooting a PC glitch until the user suggested it. They're looking for best practices to push models to explore more comprehensively.
More from Apps
- Muse agent books a family's entire LA trip — flights, SUV, nanny and all — in one day — armand_ruiz · 2026-09-27
- RowRun launches AI automation engineer that turns weekly processes into agent workflows — paw_lean · 2026-09-27
- Law professor uses model 5.5 to generate a full AI history class with reusable slides — technollama · 2026-09-27
- You're Probably Prompting AI Wrong: Common Prompting Mistakes Examined — kevinsurace · 2026-09-27
- Grok launches Finance integration to manage your bank and investment accounts — NicoVerderosa · 2026-09-27
- Turn your question into an infographic to find out what you're really asking — revodavid · 2026-09-27