00:12:09 · Joe Weisenthal, John Coogan, Jordi Hays
[00:12:09] S0: Yeah. Households are spending. They're going... They're either, like, selling based... If they're on the top part of the k, they're, like Yeah. Selling based on their equity accumulation. If they're on the lower side of the k, they're like Yeah. Going deeper into consumer credit. Yeah. Businesses are spending either from the AI build out or just, like, spending on AI. Yeah. And the government is spending on all people. [00:12:29] S1: Although, you see Torsten, Slok over at Apollo say latest? His latest is that, basically, Fortune 500 having a little bit of indigestion around AI spend outside of the Mag seven. Mag seven is seeing positive ROI. If you're a tech company, you can immediately deploy the AI agents, improve your software, see an ROI. But if you're the... You know, you're in the s and p 500, but you make tires or you do forestry or you do oil and gas, a lot harder to actually improve revenue or actually replace a job it [00:13:00] S2: lot of... I think my my stance on this is, for this for the overall trade to work, you have to have a massive capabilities jump. Like, soon. Like, in q four of last year, everyone everyone everyone was like, woah. It doesn't seem like chatbots are [00:13:18] S0: 4.5. Yeah.
00:13:18 · Joe Weisenthal, Jordi Hays, John Coogan
[00:13:18] S0: 4.5. Yeah. [00:13:20] S2: Code and all this. And then you agents. And it's like we we... The the trade right now requires that. Like, Elon Elon's saying, like, yeah. Like, you have some deceleration, you know, like, Anthropic q three, you see a little bit of deceleration. It's still growing tremendously. And and Elon is saying, like, you know, basically, all the SpaceX projections are, like, just focus on supply, right, which is, like, we're gonna bring on insane amount. Don't worry about demand. We're bringing on insane amount [00:13:48] S1: of supply. [00:13:49] S2: Right? And and that could prove to be an amazing decision. Right? Yeah. But we need that capability Yeah. Like, fast. [00:13:59] S1: Yeah. Right now, there's no one outside of an engineer that can go to their boss and say, I wanna spend $10,000 [00:14:05] S2: a month on AI. [00:14:05] S0: Yeah. Pretty hard if you're See, this is in marketing. My colleague at Bloomberg today, Todd Gillespie, has had a good thing. Goldman is clamping down on its nontechnical employees Yep. Like five coding, [00:14:17] S1: including its traders Yep. [00:14:18] S0: Which I thought was pretty telling. Yeah. And and it said it was for both risk and productivity. So if was just risk, you know, I know there are some issues with people... There's still anxiety about data sharing and stuff like that or data retention or whatever. But it's also about the productivity. Yeah. So it's like maybe not every trader needs to be five coding [00:14:36] S1: up their own dashboards and stuff. Dashboards and stuff
00:14:36 · John Coogan, Joe Weisenthal, Jordi Hays
[00:14:36] S1: up their own dashboards and stuff. Dashboards and stuff [00:14:39] S0: like that. Yeah. I thought that was a telling story that speaks to the tourist and [00:14:43] S1: not Yeah. Yeah. There's something about an engineer who's actually seen what a month of engineering work [00:14:48] S0: Oh, [00:14:48] S1: yeah. And how hard that is, and they're maybe a little bit more judicious with the [00:14:52] S2: Yeah. The thing [00:14:52] S0: is, like, as we know, like, I'm a very big believer in what some economists... And I wish I could cite the person who like first either wrote the name, but like they call it the messy jobs thesis. [00:15:02] S1: Yeah. [00:15:02] S0: We all have like, you know, within a job, we have like 30 tasks, And so forth. And so you could say, well, like, AI is really good at this task, but it's not really a replacement for someone's job. [00:15:14] S1: Yeah. Totally. [00:15:15] S0: Except maybe if you're just producing lots of code, then maybe your job isn't so messy. [00:15:20] S1: There are some jobs that work a lot like repeat tasks. [00:15:22] S0: Yeah. And so... Yeah.
00:15:22 · Joe Weisenthal, Jordi Hays, John Coogan
[00:15:22] S0: Yeah. And so... Yeah. [00:15:23] S2: I think I think what we need to see is, like, real agency. And I don't mean, like, age... I don't mean agents of, like, you're telling an agent to go and do something. Yeah. Yeah. You need, like... You know, right now, we have the equivalent of, like, every company has access to like a 100 PhDs, they just like sit in the room and they don't even talk or they don't Right? Even look so the... I do think the next leg up can come from like, hey, I'm setting a budget cap. Now use that most effectively to improve my business. Right? [00:15:50] S0: You know, I've played around with it. [00:15:51] S2: And that's effectively what you do on Meta, which is like, you're like, here's how much I'm willing to spend. Mhmm. Get me as many customers as you [00:15:57] S0: can. Right. [00:15:57] S2: Yeah. And I do feel like what... What's missing is still true agency. Right? And so agents are, like, almost like... [00:16:07] S0: No. That's true. The agents that [00:16:08] S2: we have today are, like, really all, like, sub agents in some ways. [00:16:11] S0: Right? That's right. And... [00:16:12] S2: They're only they'll only do things that they're told to do. And so... [00:16:15] S0: Well, they're also they're also clones of each other. Like, this is like another issue, which is like, you think, like, oh, if you gathered a 100 PhDs in a room and told them to go Yep. Accumulate customer, do whatever. It might actually work. But if it was a 100 of the exact same PhD who had all done the exact Alex who's over at Google. [00:16:33] S1: Love him.
00:16:33 · John Coogan, Joe Weisenthal, Jordi Hays
[00:16:33] S1: Love him. [00:16:34] S0: Yeah. He's a cool guy. He's talked about this. Like, part of the reason Yeah. This hasn't really... Like, people haven't really cracked, like, sort of, like, autonomous multi agent teams Yeah. Is that it's actually kinda all the same agent Yeah. In many cases. You sort of need to get a Thunderdome, like, battle... Make them battle out with, like... Get them, like, genuinely different points of view. [00:16:53] S1: Yeah. And I don't [00:16:54] S0: know if anyone's, like, gonna crack the right orchestration for them. [00:16:56] S1: I mean, people used to just, turn up the temperature on these. Do you remember this? Yeah. Like I [00:17:00] S0: forgot about temperature. No one's sure. No one's Yeah. Some of these things [00:17:03] S1: still feel like silly. Going back and forth for just get... Getting, like, a weirdness up and get it... Get... To let... Let's get some more hallucinations. Yeah. I mean, people were saying, like, hallucinations are sometimes like a feature more more than a bunch of, like, not... You don't always wanna hammer it out because you want some creativity. [00:17:18] S2: Yeah. Are you how are you thinking about your refrigerator at home? Is it Wi Fi connected? [00:17:24] S0: It's not WiFi connected. But I've had a... [00:17:26] S2: Even if it was, you would you would say that it wasn't. My... [00:17:30] S1: So you're a Target.... [00:17:31] S2: Lock in and heat up all your beers. [00:17:33] S0: Oh, god. No. At any... Stumping someone's beers [00:17:38] S1: is diabolical.
00:17:38 · John Coogan, Joe Weisenthal, Ryan Serhant, Jordi Hays
[00:17:38] S1: is diabolical. [00:17:39] S0: I I bought my, refrigerator, like, two years ago. Amazingly, the ice maker works. I've never had it... My whole life, I've never had an ice maker works that long. Luxury. [00:17:46] S1: Can you make a hassle? Can [00:17:47] S0: you Yeah. Make No. My life is pretty good this day. [00:17:49] S3: I got [00:17:50] S0: the ice maker still. [00:17:51] S1: I do [00:17:52] S0: have a Wi Fi enabled Dude. Air conditioning. I hate it. So, like, I actually could... Yeah. Yeah. Yeah. Because I've, like, you know, if I'm coming into town or something like that, [00:18:00] S3: I can turn on the air conditioning before I [00:18:02] S1: get home. That's great. That's great. [00:18:04] S0: I can never... Like, when the the Internet is down or whatever, like, I don't know where my actual remote controls are. [00:18:10] S2: Yeah. Use case. Agent runs, detects your location at JFK. [00:18:15] S0: So this is what we're talking about. Now that would accumulate some tokens. You know, like, this would get some demand going. Constantly checking Joe and whether we think he's near your home [00:18:24] S1: or not. Dinner reservation with artificial intelligence yet? [00:18:27] S0: No. The... I have I have not [00:18:30] S2: yet. Yeah. [00:18:31] S0: I have... And I have not booked a flight yet either. [00:18:34] S1: Okay. [00:18:34] S0: I've tried to book a flight. [00:18:35] S1: Okay. [00:18:36] S3: And I was just like,