[retep_kram]

One of those instances of AI-generated code replicating another library's API, using a more low-level language for the core implementation.

I love the idea of using new low-level languages and making libraries faster. But I wonder if there is a better way to recognize and value the hard work that the original designers of libraries like FastAPI, Zod, Pydantic, have invested in to make those libraries.

Without the API design of those libraries, this turboAPI and dhi would make no sense.

[pennomi]

Honestly that’s one of the best potential uses for LLMs, translating code that was cleverly designed by brilliant humans into lower level languages.

I don’t trust LLM API design in the slightest, but they are decent at the brute force coding part, especially if you can replicate the testing suite.

[alanwreath]

I was asking earlier if DHI has tests - seems they do. As does the FastAPI refactor.

Would be nice though if there was an exact test suite that ran against both that could demonstrate parity. It would be onerous to try and compare both library’s test to see which one is more comprehensive (believable) - by default I’d expect the original to be the most believable (duh) mostly because tests represent historical cases that can up and broke it.