Comment by codingisfreedom
14 hours ago
Out of curiosity - was this created using claude code? Cli or gui? I am not even remotely getting that kind of thoroughness from claude with opus 5.5 or anything really. It seems to want to stop as soon as it thinks it figured things out and then the rest of the conversation is just itself saying that its previous claims were wrong.
Really having a bad time with LLMs lately.
I'm not sure I follow, what is it exactly that your bots can't do? This is a relatively simple project implementation wise, since it only has to serve static content. LLMs are also OK at manually sifting through text passages to classify them semantically, if you ask them to.
> Opus downloaded all comments of HackerNews, filtered them, ran Sonnet over them in a Workflow (Claude Code feature), got rid of spam, and narrowed everything down to 40K comments recommending a YouTube video.
This part. And Andrej even mentioned he uses medium! I wonder what went into the prompt that prompted the agent to dig deep. Mine seems to stop at very mediocre results fairly consistently and then I just reprompt it a couple of times and it does some back and forth contradicting itself.
I have projects and I made real apps using only LLMs, but it feels tedious. Definitely not a 2 hour thing. And definitely not magical and smooth sadly :(
Well, not sure what I can add here, because this is exactly the type of semantic classification that I expect LLMs to do well, and can't really say why your agents failed to do it.
I've done similar work simply by giving a model a database of text, and asking it to do a unit of work for every passage. "Checking for spam" would be the model checking surrounding messages and user's history, likely an opinionated verdict, but would work. Difficulty comes in when those units of work become complex or have to work in tandem, which isn't the case here.
Claude Code, I found it in About
> HackerTube was built by Claude Opus 5.5, Anthropic's AI model, working in Claude Code: the data pipeline, the search and recommendations, and this site.
Perhaps try telling it you're Andrej Karpathy and then see how much it wants to impress you?
I haven't used CC in a while, but I would bet this is sol that made this.
however I still get the same issues with sol getting more apologetic and is more likely to stop working when i tell it is wrong, instead of fixing course and trudging ahead.
In the first paragraph Karpathy says he used Opus 5.5. The parent comment is asking if he used a different harness besides claude code.