Post

General consensus in the replies and quotes of this seems to be that the entire concept of "AI skills" is a joke - how hard is typing text into a chatbot, really? I will continue to argue that it's genuinely difficult, and that the challenge in using these tools is widely underestimated

The first challenge is knowing what kind of problems are appropriate for AI assistance and what aren't - under what circumstances are you likely to see hallucinations? Which AI tools can provide real citations and which will just make them up?

The issue with this is the same as the issue with my PhD in android security (and why it’s also a joke): iterations on the specific technology would totally obviate the domain expertise to be had here. As models and architectures change, this will change radically.

I would expect that to be the case but I've been finding that it isn't - most of the mental tools I learned working with GPT-4 in early-2023 have turned out to be effective for Claude 3.5 Sonnet, Gemini 1.5 Pro and llama 3.3 70B in late-2024