So this one is a hard one to answer specifically, but I'll say, like, we've been for over two years at Khosla, we've been a big fan of agents, like, even before the term was popular. But the reason agents are interesting is because it's it's like literal work replacement. Like, you're automating real work, and you're you're not automating, like, low value work. You end up automating high value work. So I think the potential when you ask, like, potential for ML and enterprise applications, I think most of it is going to be what are the big rock areas of value that you can actually automate and make much cleaner and simpler to do, And how does that actually unlock future potential that just wasn't achievable before because you are bottlenecked on call it latency, speed, cost, etcetera? So, like, it's it's kind of a a broad answer, but I think you're going to see a lot of change in the way enterprises actually do work, and a lot of that's going to be driven by agents. And one one very great example of this is, like, a lot of you are probably engineers and you use Cursor or, like, you played with Devon. Right? TBD, if those are actually the things that exist five years from now, but I think coding, at least for, like, certain aspects of coding in the next five to ten years are probably not going to exist in the form that they exist today. Like, a large number of the companies I invest in have no front end engineers, and it's all just they don't even know React. It's all just AI generated. So there is definitely classes of area that are going to be like, areas, that are going to be disappearing in the future.