atomon

Putting aside the details of the product itself, I love the style of this post. I wish more announcements read like this.

Is the main point “let me mathematically prove that it’s impossible to do what I’ve already done 12 times this year?”

Yes, very long workflows with no checks in between will have high error rates. This is true of human workflows too (which also have <100% accuracy at each step). Workflows rarely have this many steps in practice and you can add review points to combat the problem (as evidenced by the author building 12 of these things and not running into this problem)

This looks very cool, will try it out on my next project.

There have been a number of solutions popping up to address this problem, and I think the need is very real. Decomposing these LLM tasks into subtasks seems to be one of the best ways to work around the shortcomings of LLMs in production apps (hallucinations, context window limits, etc). But then you end up with complicated pipelines that are difficult to debug, improve, reason about, etc.

He’s a couple days unemployed and is already producing code that looks like it’s meant to teach people. Very happy to see that! The world will be a better place with more Karpathy educational content.

Can you walk into any NY retailer and do this? Or is it up to the seller whether they support it?

We're having issues as well