Grant Harvey

Okay, fair enough, fair enough.

But the general idea, like there are non-traditional skills that are incredibly valuable in this world where you're effectively engaging with a model or an interface on top of a model to try and understand the problem or try and break it down or try and deconstruct it or try and like sequentially map it and do all sorts of complicated things to set up a set of agents using different tools to do some piece of work.

Like that's not sit down, write Python, hello world.

You're like, let's talk about the car.

I'm just like, hey, I'm just chatting on League of Legends.

Okay.

I mean, there's a whole heap of things.

I think thinking it's step one, talk about AI, step two, question mark, step three, magic.

Like, no, it's a very boring set of processes.

A classic example, I mean, we really think of it like sort of left to right data process evaluation agents.

Yeah.

If I'm going into your company, I'm saying, okay, you're an underwriter.

You're an insurance company.

I'm saying, hey, how many claims did you play out in the northwest of England in February 2024 that involved flood?

If your answer to that is, oh, yeah, fantastic.

Let me just go and query.

Done.

You're in a good place.

If your answer is, lol.

How are you going to get AI to do anything?