Ask HN: What are your use cases for o1 so far? | Hacker News Reader

Ask HN: What are your use cases for o1 so far?

8 pointsfsndz4 comments

Feel like there aren't a lot of use cases where o1 radically changes the reliability issue of LLMs in a way that could make human in the loop approaches less necessary. And played with o1 a bit, but so far it hasn't changed my workflows. Saw some people with the same sentiment on X, wonder if this is general ?

View on HN

It's very different from the others. Ideally you 0 shot it and give it complete instructions. Not part of a task, but basically a full commit, if not a full PR.

It does worse than sonnet 3.5 on simple instructions.

Ideally, you use o1 to give you instructions. It's a very powerful brainstorming tool as it can actually handle the context and isn't just context tokens. Pass it that doc from upwork and just do what o1 tells you. It may still get some steps wrong, but the human in the loop role has been reversed. It's still pair programming or brainstorming or something, but now o1 is the senior dev guiding your hand.

Write proper docs as if you were explaining something to a human. Or a deck to an investor. Pass it to o1 without edits. Writing docs really sucks, but do it without the aid or interference of AI.

I’ve been experimenting with o1 as well, and while it hasn’t radically changed my workflows, I did use it to create a blog webpage with minimum human effort required. It helped automate much of the content creation process, reducing the need for constant oversight.

You can check out the project here: https://github.com/mergisi/openai-o1-coded-personal-blog

Why is the ai slop from o1 better?

Since it's just in preview I haven't been able to get much done with it before hitting the limit. So it's true value to me remains to be seen.

Reply on news.ycombinator.com