What are A.I. skills?

They're in demand

There’s a lot of chatter about the demand for A.I. skills, A.I. proficiency, and A.I. certifications presumably so that everyone can be A.I. first, and not get left behind. But I still struggle to understand what having A.I. skills means.

I assume it has something do to with knowing which incantations to use with which models and agents in which circumstances using which set of extra instruction files granted which set of permissions to which systems and which ways to verify it all in order to get it to do useful work without costing too much. But even that is in a constant state of flux, and wouldn’t be a solid basis as a skill. The ability to continuously monitor and experiment with A.I. models, agents, tools and techniques as they evolve?

If A.I. worked as advertised, you wouldn’t need A.I. skills. You would ask it a question, and it would give you the right answer. If you asked it to perform a task, it would do it correctly, ask for clarification, or say it didn’t know how. But it doesn’t do any of those things reliably.

That’s where A.I. skills come in. They are all the ways that you need to compensate for A.I. deficiencies. All the personas, and agent files, the clever phrasing and re-prompting, the secondary verification tools, the proof reading (and knowing what to look for), the special environments, the favorite models, and clever orchestrations of multiple agents.

If A.I. changes, those skills will need to change. If it get’s better the need for those skills may well go away. Are those skills a good investment? Are you betting A.I. pretty much stays the same complex beast that you’ve managed to tame (at least long enough to cash in)? Are these skills that no one else will be able to easily acquire in the future to catch up or leap frog you? Will that time you spent honing outdated A.I. skills leave you behind?