Andy, we're just swamped with demand that we shouldn't even be talking about price.

Um, interesting.

Yeah, I mean, so, so a couple things I would, I would point out on that, and, and then I would come back to something Matt mentioned on, on efficiency gains.

Um, what, what I think is, is interesting is that the amount of compute required per use case has gone up significantly.

And so, uh, there, there was a recent analysis done that, um, you know, the difference between, um, the tokens required for, you know, an, an early chat request from ChatGPT three years ago versus an agentic function today is a, about a billion times, uh, more compute required.

And then, um, you know, we, we've seen more than a fourfold increase just in, in token consumption compared to a year ago.

Um, so, so you're kind of seeing those two things pair out, which is that even as we've had efficiency gains on the compute stack, which, which Matt mentioned, um, the demand signals have also in- increased dramatically.

What would be the signals that something isn't quite right, that this, uh, this boom is beginning to, to, to, to peter out or falter, um, and, uh, how would you approach that?