No Priors: Artificial Intelligence | Technology | Startups
00:27:49 · Elad Gil
[00:27:49] S2: Oh, no, was just curious. And then I guess another it's always hard to predict these things, but I'm a little bit curious. Do you think there's structural advantages for Chinese companies to stay at the frontier of at least open source and open weights, or do you think US open weight models are likely to catch up in the near term?
00:28:04 · George Kurtz
[00:28:04] S0: Well, are some open weight models coming out. I'm sure you may be invested in them as well. I think having great US open weight models is going to be very important. So you know, one of the reasons why we worked with Nvidia was, of course, Neumotron is open weight and open source, not just open weight, but having other alternatives is going be very important. And I think there's, you know, again, just being in technology long enough, you kind of look at how things unfold like the operating system of the Internet is Linux, like period, Okay? And there's a reason for that, and there's going to be an operating system of AI and it's probably in my mind it's going to be, you know, an open source open weight when that happens and how it all unfolds. I mean, we'll see and it doesn't mean there isn't room for any frontier model. It doesn't mean that people are not going to make a bunch of money off it. You know the application stack sort of around it. You know the harnesses and how things are connected and how you think about enterprise. All those sort of things are going to be really important, but ultimately like there's just too much cost involved right now. When you look at a lot of these frontier models, I mean, they're being commoditized. If the Chinese models are three months behind, you know a lot of companies are going like three months is not what what am I going get out of it? I does everything I want. And that's the challenge.
00:29:27 · Sarah Guo
[00:29:27] S1: A top five financial said to me recently, like a, you know, a large technology buyer there that there was no chance that their organization had the capability to absorb a new model release every ten days anyway. And so three months behind, you know, was faster than the accepted well, like, regulated pace of their organization. So it made no difference to them to to to some degree. I I think open is here to stay as a big part of the puzzle. You know, one one question. I I know we're gonna keep asking you to look into the future, and it's kind of impossible. But given your history in in security, this is the first time in my career in technology where some very smart, well educated technologists are like, we need to turn the Internet off. We just gotta air gap everything because the Internet is gonna fall apart in the next year, and it will be unsafe for critical systems to touch it. How how much do you believe that to be, like, a reasonable probability in the future where there are lots of critical systems that society cannot rely on touching the Internet?
00:30:36 · George Kurtz
[00:30:36] S0: I mean, I think it's nonsense. You know? Like, is security and what agents can do and the risks that they pose, is that serious? Of course. That serious business. We should all look at that and go. Yeah, we got to make sure that we lock things down, you know, but is the world going to end because of these agents anytime soon? No. And when you look at what the agents are doing, I think there's this sort of mythical view of what an agent does. Like an agent doesn't just beam itself from your computer to another computer. It actually just does the same thing a human would do, just better and faster, right? So if you think about what companies like CrowdStrike have done for a long time, we prevent against these sort of attacks. Now they might come faster. There might be more lower level sort of vulnerabilities chained together and those sort of things. For the adversary, not great for the defender, but it's building the technology that can deal with all these sort of attacks. So what is AI done? It's created more vulnerability so you have more of them and you know you have more exploitation coming faster, so it's more and faster, but there isn't like a unique class that all of a sudden an agent just beams itself somewhere and I think people just get wrapped around the axle of like there's some new class of technique that's taking there. There isn't. It's just been trained on what we know. It hasn't developed its own yet, and that's why I kind of look at it and go security again is going to be an enabler for AI, which I think is incredibly important for the security industry, because normally it's you know you're putting the brakes on. You know we have to take time. You're not letting the business run as fast as they want, and now everyone's coming to CrowdStrike and others going, Well, how do you make AI secure so we can go faster in AI? So that's the way I would look at it. AI is here to stay. I think some of the concern is there, and it should be there, but sort of like there's a lot of kind of doomerism around some of this stuff, and I just don't think it's necessarily warranted.
00:32:40 · Elad Gil
[00:32:40] S2: I know you're also sometimes involved with companies as an angel investor and other parts of the ecosystem. Are there other areas that you're very excited about right now in terms of key areas of change or focus in the AI world in addition to the security world? Like what are some of the things that you think are really exciting that either just happened or are coming?
00:32:59 · George Kurtz
[00:32:59] S0: I think, you know, chips used to be kind of not in fashion. Obviously, we know they are in fashion today, and you're only going to need more of them and faster of them and more specialized. And I do think there will be some big breakthroughs in the chip world. You know, when we think about just sort of the speed, take, you know, Moore's Law might be a bit dated, but at the end of the day, you know, the chips came out and things just got faster and cheaper. I don't know the trajectory, but I can guarantee you it's going to be faster and cheaper in the future like that I can see. So as these specialized chips are developed at, you know, focus in maybe particular areas as new science is developed around the mathematics. Know there can be step function increases in performance. You know less power. Those sort of things. I think chips is a big one. The open weight models. I think there's a lot of room to grow. Certainly The US models, and I do think there's going to be a lot of advancements in taking these models and shrinking them down, you know, and being able to run them on your phone as an example. So I mean, we probably all know the history. Remember the days where you know it took a football field sized room of computers to to, you know, launch the space, you know, first rocket like we have more power in our iPhone today than they had in the entire room. So that's going to come. I think the area where you can make AI faster, cheaper, more efficient, you bring it to the desktop and you bring it to the phone where you're doing it at the edge, I think all of those are sort of exciting opportunities because that's really where I see. You know, you're probably going have a hybrid model, but you can do a lot of it now locally. Again, just looking at the, you know, the TPUs, MPUs, etcetera. Like, these are... They're they're all on systems that we have today, and and a lot of times they're going to waste. But I I don't think they're gonna be going to waste that that much longer.