Who's making the decision or who owns the decision about what work stays human and what work stays AI? And what happens if nobody makes that decision? So who's kind of owning that work, I don't know, delegation, if you will?

Well, we think that leaders in the organization need to take hold of that.

It is not a single department's responsibility.

You know, think of it this way, is that as a manager and as a leader in an organization, one of the core skill sets that you have to just get good at is how to recognize good talent, right? How to interview them, how to assess them, how to hire them, how to onboard them, how to train them, support them.

That is our role as leaders within an organization.

The thing that we want people to understand is that when you think about doing that for humans, you have to think about doing that for AI as well.

So when we talk about AI as talent, we actually mean that.

You got to understand it so you can bring people along.

Can you talk a little bit more about what it's actually measuring and how you interpret the score?

The first thing that it's measuring is what we sort of think of as the amplification score.

And really this stems from a basic insight that we had very early on, and I think this actually will resonate a lot for your listeners, is that there's really two different types of usage of AI.

One type of usage is where it's transactional.

It's an answer engine.

It's a gopher.

It's like, "Go do this for me, and I'm gonna move on with my day.

But I've kind of come to the conclusion that if you read the tea leaves, token spend based upon the potential introduction of these open source models, is it not everyone's thought that token spend is going to come down in unit cost dramatically over the next year or two? Or am I not reading that right?