stephen ryner jr. 🦉
@nuthatch@infosec.exchange
I’ve been working more with LLM models recently. It seems like any other tool, but it’s a very slippery footgun. If you know what you’re doing, it might help, but if you’re trusting it outside your domain, damn, things could go bad real fast. I think exceptional developers may get exceptional results, but proceed with caution.
stephen ryner jr. 🦉@nuthatch@infosec.exchange
My biggest annoyance is the “obvious” solution is usually crap. So you look at what got barfed out and say, e.g. “Can you do that without Combine? Use modern Swift,” and it proudly says “Oh yes this, is much cleaner!”
…why didn’t you give the better solution the first time 😩