About whether LLMs like ChatGPT can build non-trivial things, after a Wozniak AI-anxiety interview.
Has this happened?
Yes 30 (30%)Not sure 25 (25%)No 46 (46%)
Votes cast 1–2 October 2026: 100,590 votes from 9,694 people.
Hacker News has set AI a lot of challenges over the years. Which ones has it met?
About whether LLMs like ChatGPT can build non-trivial things, after a Wozniak AI-anxiety interview.
Votes cast 1–2 October 2026: 100,590 votes from 9,694 people.
Nothing I wrote was personal or rude. I argued against the content of your comment, not you as a person. HN is only interesting because people disagree with each other here all the time.
go build it
I don't know if my comment implied this or not, but I can't build it because I don't think it's possible. I don't think ChatGPT can help with non-trivial applications yet, and LLMs may never be able to because they can't "understand" things, they can only guess what billions of other people would write next.
That's why Copilot works (pretty) well, but writing an entire novel with a cohesive plot doesn't.