Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I would imagine that it would do better if it was told to build it up step by step, and critique its own work as it goes along.


It would but what I like is how difficult the task is. If the model becomes able (like today) to generate passable results from such a simple prompt, then that shows something, IMO.




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

Search: