so first the first thing is it's not just about optimizing the

tokens so it can't it can't just be as an organization of that size if you try and optimize for one particular number you'll, you'll, you'll not, you'll not get the right, the right outcome.

So you need to look at the efficiency across the organization.

So is everybody working in tiny little, you know, silos of, uh, they've got a little bit of AI enhancement in their job, but it's in isolation.

It's not part of a joined up workflow and the old way of doing it still exists.

So you're still paying for that old way.

So you end up in a position where if you optimize those tokens, you're not actually saving anything.

To kind of start to wrap up, Lee, if you had to make a prediction for where we are with local AI in the next six to 12 months, where do you think we are?