Hacker News new | past | comments | ask | show | jobs | submit login

At the end of the day, coding with AI is still coding.

LLMs need to get better at it - more precision, less hallucination, larger context windows, and stricter reasoning. We also need a clearer and more domain-specific vocabulary when working with them.

In short LLM will end up as one more programming language - but one that’s easier for humans to understand, since it operates using the terminology of the problem domain.

But humans will still need to do the thinking: the what-ifs, the abstractions, the reasoning.

That’s where AI becomes unsettling - too many people have been trained not to think at all.




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: