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?

How come supposedly the most advanced computer systems we have ever built can't look up facts, can't do long division and can't count the number of Rs in strawberry? (Except when they can, because some of them can consult external databases or run Python or JavaScript code to use as a calculator)

Did you know ChatGPT now has two entirely different ways of running Python code? The rules of what those can do are utterly baffling simonwillison.net/2024/Dec/10/...

I don't think students should take classes on how ChatGPT Code Interpreter differs from ChatGPT Canvas - that would be like teaching MS Word specific VB macros in "intro to computers" I do think there's a lot of space for helping teach students how to navigate these utterly weird new AI systems

The fundamental challenge here is that there are key skills - reading, writing, thinking - that students urgently need to develop The temptation to outsource those activities to generative AI will clearly harm the acquisition of those skills - they need to be a slog

This is a great point: learning how to ask clear, unambiguous, non-leading questions is vital for getting good results from LLMs and is a valuable skill in itself, independent of AI usage bsky.app/profile/mich...

If you don't believe in AI skills, what's your opinion on Google/search skills? I think learning to effectively use tools like Google is a greatly underrated skill as well bsky.app/profile/rod....

Sure, but is that what software eng hires for now? I've been rejected from more jobs for not remembering how to do something I only do during job interviews, than for failing to display these higher order skills.

I see that as more of a reflection of how terrible we are at designing effective hiring processes!

Hate to break it to you, but Im not sure we're any better at teaching higher order skills! ;)

Yes, but it’s not at all clear that the right way to do that is *via AI systems*. Learning how to use LLM-based tools is a practical skill one builds *on top of* those fundamentals, and while you might be able to deploy LLMs in that loop, they may also short-circuit a lot of its development.

That is: pedagogically, the question is not whether people may have need to develop skill in using web search or AI or any other tool, but when and how those practical skills should be developed—as distinct from more foundational skills in reasoning, problem-solving, etc.