00:17:58 · Joe Weisenthal, Alex Altmann, Tracy Alloway
[00:17:58] S2: Versus- Incredible token deflation that's happened every year. [00:18:01] S1: It might be one of the fastest deflating assets in history. It makes those NFTs back in 2021 too. [00:18:06] S2: But GPU rental rates have been going up. They have. [00:18:11] S1: Yes, absolutely. And so there is clearly a two-way debate. And I think it again goes back to the evolution of this broader AI narrative. It's not just necessarily a one-way train, rising tide lifts all boats. There's going to be degrees of differentiation. But you asked about open weight models. And I think where one of the leading edge questions right now is not even just open weight, but combine that with locally run open weight. And this is something which I like to call being long or short human creativity. [00:18:42] S0: Right. [00:18:43] S1: In so far that if, when the iPhone came out in 2007, stay with me here, uh, that if I had, if someone from the future had said 20 years from now, you're going to have whatever the orders of magnitude compute differential And the battery is going to be the same size. That 2007 person would have said impossible. You'd need an iPhone the size of this table, right? So clearly we got there amongst many other things. And so the idea that we necessarily need to ping a bunch of hoppers or Vera Rubins in an Ohio data center just so that we can sort of get our. [00:19:21] S0: Agent to- Manage my email for me. [00:19:23] S1: Yeah, or to send an angry response letter to the hotel. [00:19:27] S2: You're going to be able to do that with the chip on your phone by next year.
00:19:27 · Joe Weisenthal, Alex Altmann
[00:19:27] S2: You're going to be able to do that with the chip on your phone by next year. [00:19:30] S1: Right, if not already. So we've all kind of got supercomputers in our pocket. And the ability to run, say, multi-billion parameter models locally using whether an open-weight model or a closed-weight model, but that's provided by NVIDIA's Nemotron or whatever Meta's putting out, which Meta's already put out a- multi-billion parameter model that's run locally on your laptop. [00:19:52] S2: It's crazy. Right. [00:19:54] S1: But it's not because that technology, right? And so I think that's the other point of the debate is it's not even so much about open weight. It's about the fact that you're moving along this efficiency curve And that was also bringing in the revenue question as well, which ties in with those frontier models as well. Frontier models, we're not going to be producing, you and I, with the greatest respect to everyone in this room, we're probably not going to be discovering the new biotechnological drugs.
00:19:54 · Alex Altmann, Joe Weisenthal
[00:19:54] S1: But it's not because that technology, right? And so I think that's the other point of the debate is it's not even so much about open weight. It's about the fact that you're moving along this efficiency curve And that was also bringing in the revenue question as well, which ties in with those frontier models as well. Frontier models, we're not going to be producing, you and I, with the greatest respect to everyone in this room, we're probably not going to be discovering the new biotechnological drugs. [00:20:38] S2: I was actually, I was talking to someone when I was out in San Francisco recently, and I actually want to ask you about East and West Coast views on the technology, maybe. But I was talking to someone in San Francisco who had this term, he called it the frontier sink. And it's like, what are the industries that actually need PhD level and beyond technology? And there's a few that we could come up with pretty easily. One of them is probably biotech, right? You can probably never have too much intelligence when it comes to medicine. Another one actually is probably semiconductor design itself. You'll probably never be too smart, et cetera, for that. Maybe another one of them has to do with quant trading where you can never be too smart. I'm not sure. Who knows? But it actually does get to be hard to think about. There might just be a lot of industry for which there is a certain level of intelligence that is enough.
00:20:38 · Joe Weisenthal, Alex Altmann, Tracy Alloway
[00:20:38] S2: I was actually, I was talking to someone when I was out in San Francisco recently, and I actually want to ask you about East and West Coast views on the technology, maybe. But I was talking to someone in San Francisco who had this term, he called it the frontier sink. And it's like, what are the industries that actually need PhD level and beyond technology? And there's a few that we could come up with pretty easily. One of them is probably biotech, right? You can probably never have too much intelligence when it comes to medicine. Another one actually is probably semiconductor design itself. You'll probably never be too smart, et cetera, for that. Maybe another one of them has to do with quant trading where you can never be too smart. I'm not sure. Who knows? But it actually does get to be hard to think about. There might just be a lot of industry for which there is a certain level of intelligence that is enough. [00:21:34] S1: I mean, for the moment, again, Do you remember when the, I like going back to the iPhone as an analogy, because it really was a technological revolution at the time. And do you remember when we thought like peak iPhone in 07, 08 was when you had that app, which had like the pint of beer in it. [00:21:51] S0: And you could drink it. [00:21:52] S2: I didn't, but no, everyone is like, Oh, this is so cool. You can pretend. I know exactly. [00:21:57] S0: Joe, you're allowed to take joy in small things. No, no. [00:22:01] S2: This looks like whimsy to me. I don't get this. Oh, yeah. [00:22:04] S0: Joe hates whimsy.
00:22:04 · Tracy Alloway, Joe Weisenthal, Alex Altmann
[00:22:04] S0: Joe hates whimsy. [00:22:04] S2: I hate fun. But keep going. But I remember how, yo, why is everyone going crazy about the fact that you could, like, this is a very good point. Keep going. [00:22:11] S1: And so I just don't think we know what the future of AI really looks like. So, yes, you're right. Today, we probably say we just don't need more than 100 IQ persons sitting in our pocket, managing our emails, sending angry messages to the car rental company because they charged me $ 2, 000 for a ding. I do think the consumer layer AI is going to be, people ask me all the time, like, what will get you really constructive on AI again? I said, when we figure out that killer consumer application layer. Because at the end of the day, the consumer is everything. It's the economy. It's where, to the point about revenues, like enterprise revenues, I just don't think is going to plug that gap. Really? I just don't see how. [00:22:51] S2: Yeah, no, you're right. The numbers. I don't see.
00:22:51 · Joe Weisenthal, Alex Altmann, Tracy Alloway
[00:22:51] S2: Yeah, no, you're right. The numbers. I don't see. [00:22:53] S1: It can probably get there eventually, but it can't get there in the time frame that the market wants to see the ROI. And you need to bust into the consumer layer. And so back to the point about frontier, the consumer probably doesn't need frontier for the time being to do all of these things. And so my big mantra is that one of the best technological changes that we experience as humans is actually not necessarily some new advanced tech. It's changing the application of existing tech. And so it's very possible that we take the 100 IQ agentic workflow that still kind of becomes almost ubiquitous within the next 12 to 18 months. And we find really cool ways to use it. And I think that's probably where you're really going to get the next upswing in AI. [00:23:42] S2: Can I just say, by the way, Tracy, last several days, my wife was traveling out of town. My kids are at a very easy age. It's not a big challenge or anything to solo parent for a few days. But it happened to be a very busy schedule for them, birthday parties and a few other things. And it was very nice having ChatGPT being able to say, where should I be in the next hour? Where should I be in the hour? Because there were a lot of things going on. And it was very easy. It was very nice. [00:24:12] S0: I saw your tweet on this, Joe. [00:24:14] S2: No, you saw the tweet on the right. [00:24:15] S0: This is part of the problem. You tweet everything before you tell it to me, and so I can't react to it in real time.