q4y2u35uk

No you're right, this code is terrible. Running a linter like clang-tidy would surface many issues. Not sure why the LLM didn't think of ever doing that.

People who think this do not understand LLMs. LLMs thrive in abstractions just like humans. Assembly code introduces far too much entropy, requires far too much context and documentation, and is far too easy to get horribly wrong. Even after generating large amounts of synthetic training data, it will never be practical.

The difference is Codex couldn't shake its preference for making flashy, loud, obnoxious marketing websites that look like they're trying to sell you a SaaS. This website looks like it wants my money. Meanwhile the website from op is simple, focused, and far more comedic than codex's slew of AI-generated "jokes".

Of course I am familiar with learning from Q&A with LLMs. I'm talking about agents creating full interactive courses with a thought-out curriculum. It's typically been avoided due to LLM's tendency to hallucinate, but these days, swarms of Opus agents corroborating are near-infallible at deciding fact from fiction.

Looks useful. Was AI used to generate the content? Not an accusation, I'm just personally curiosity about a future where AI can be used to upskill us rather than the opposite.

I don't doubt there's a human touch, but the point is art school and being a designer are orthogonal to what would have made this project impressive in the past. This project interfaces with deep and partially undocumented macOS internals. A few years ago this project would have been a pet project of a master systems programmer.

I don't think that making software should be gatekept. I suppose it's a bit melancholy that software has to be appreciated differently now.