What do you make of the idea that this is a pretty simple recipe, there are open-weight models, and there is a China, and that all the pricing power will be destroyed because a midrange open-weight model can do everything that you probably want it to do on a Mac. Right now, as you and I are talking, it’s not clear to me that they can get to AGI before they run out of money or before CoreWeave collapses or before the bubble writ large comes to some other conclusion about the state of the Economic landscape. California,” where I’m just openly skeptical of these claims and you guys are fully AGI-pilled. So training budgets may fall, but inference budgets will be robust for the foreseeable future. I mean, these are still very young companies.