Comment by runeks

If LLMs are really as good at writing code as some people say, they should write a script that fetches each challenge, as soon as it's released, and feeds it to an LLM, to produce a solution. Then run this code and submit the answer it outputs.