01:58:05 · Stacy Rasgon
[01:58:05] S6: Yeah. I'm and and again At least for Oracle. Making any comments on on Oracle, I mean, those are the for for any of these investments, those are the worries that you're gonna have. Right? Either because again, especially if it's OpenAI. So Altman's committed to a tremendous amount of capacity. Mean, it's just between on the chip side between NVIDIA Broadcom and AMD, it's 26 gigawatts. Yeah. Right? And just for some context, I think the whole global data center installed electrical capacity, want to say something like 70 gigawatts, something in that ballpark, right? So it's a lot. I mean, was even making comments he wanted to deliver, what was it, a gigawatt a week I at one mean, it's a lot, right? And so you already are wondering, can he actually deliver on that or not? Those are valid questions. By the I wouldn't count Altman out like he's aggressive, but he's got like large aspirations. But then with someone like an Oracle who doesn't have the balance sheet or the income statement that some of the others do and you look at the size of the I can't even remember what it was, 400,000,000,000 or something. Was a massive step up, 300 in one quarter. Like, you you wonder where it was coming from. It it's always around, like, can they deliver it and can they pay for it? And and those are the general worries, I think, for anybody that's announcing something.
01:59:17 · John Coogan, Stacy Rasgon, Tyler
[01:59:17] S0: Yeah. How do you analysts try and dig into, you know, there's quality of earnings, quality of revenue, quality of backlog feels like an important thing to get into. Yeah. But I don't know that the contracts are available. [01:59:31] S6: They're they're not. Right? So, look, I I mean, you you talk to whoever you can and, you know, you've you've been doing this while you look for pattern recognition. I mean, look, nothing in semiconductors has ever really baked. [01:59:43] S0: Sure. [01:59:43] S6: Right? I mean, in fact, we just discovered like during COVID, lot of these companies were experimenting with non cancelable orders. [01:59:49] S2: Sure. [01:59:50] S6: And we saw what happens. It's like, yes, maybe I could contractually force my customer to take a year's worth of parts they don't need. Yeah. Am I really gonna do that? Historically, the answer was no. Yep. During COVID, for some companies, the answer was yes. And and those particular companies are are still paying for it. [02:00:05] S0: Yeah. Yeah. [02:00:06] S6: Because, okay. Great. I just took all my parts. I don't need to order anything from you for for a year. Right? [02:00:10] S0: Yeah. That makes a lot sense. [02:00:12] S6: And so I'm always a little hesitant of of of backlog and and commit. And even some of these, by the way, even with these open day ideals, so they put out big numbers, but it's not like the whole thing is committed. Like, maybe the first gigawatt at this point Totally. Which is how it should be. Yeah. [02:00:25] S0: Yeah. Yeah. No. That makes a lot of sense. It's great.
02:00:25 · John Coogan, Jordi Hays, Stacy Rasgon, Healey Cypher
[02:00:25] S0: Yeah. Yeah. No. That makes a lot of sense. It's great. [02:00:28] S1: Are you AGI pilled? [02:00:30] S6: No. Probably not. I'm not exactly sure what AGI means yet. You know, there's a lot of different definitions, and sometimes it tends to change. [02:00:38] S0: Well, do you find AI useful in your day [02:00:40] S6: to do. Day I have to be a little careful because from a regulatory and compliance standpoint, it's not like I can just start throwing stuff the I'm very limited. We have some internal stuff that we can use, which is [02:00:51] S0: Enterprise plan. [02:00:52] S6: Yeah. And it's ring fenced and everything. But I mean, simple stuff. Right? I mean, what it was, this NotebookLM, can literally take I could take a YouTube of this, what is it, three hour podcast and toss it in there. Yep. And it'll give me a a summary, not just a summary. [02:01:06] S1: Then we can take that summary, put it into another model, and say, make me a three hour. [02:01:10] S6: Sure. Yeah. Pretty soon, it's just AI is making and watching podcasts. Yeah. It's a slurry [02:01:14] S3: of content. [02:01:15] S6: But that's just one example. Like, so there there were there were clearly [02:01:17] S0: Yeah. [02:01:18] S6: Ways where where it's influenced my my experience in my in my workload. And I'm I'm probably on the low end, frankly, of of of of of what I could be using if again, I am I am constrained Sure. On what I can can do with this. I'd I'd love to use it more [02:01:31] S0: Yeah. [02:01:32] S6: If I could.
02:01:32 · Stacy Rasgon, John Coogan, Jordi Hays
[02:01:32] S6: If I could. [02:01:32] S0: That makes a lot of sense. Well, thank you so much for coming by. Yeah. You bet. The studio. This is lot of fun. [02:01:36] S1: To have you. [02:01:37] S6: Hey. You bet. [02:01:37] S0: Have a great rest of your day. We'll talk to you soon. Cheers. While he's hopping off, let me tell you about Linear. Linear is a purpose built tool for planning and building products, meet the system for modern software development, streamline issues, projects, and product road maps. Speaking of turning things into podcasts, Christophe has some fake news here on the timeline. I love Christophe, coastal futurist. Christophe says, still crazy to me that Steve Jobs slash Apple invented the word podcast and that it's a mix of iPod and broadcast. And apparently, this isn't true. Apparently, a BBC journalist, Ben Hammersley, coined it in a 2004 Guardian article, but then it got ported back to Apple and the word and the term podcast was adopted. But it is it is funny that [02:02:26] S1: Podcasts. Podcast. It did it did come from It's always felt like a very boomer term. [02:02:31] S0: Yeah? [02:02:32] S1: And it makes sense. I I never Radio show? I never knew that it was it was a combination of [02:02:36] S0: Oh, didn't know. It was iPod plus broadcast. Yeah. Yeah. What else? [02:02:41] S1: What does that that make The [02:02:42] S0: cast casting is very, very popular. [02:02:46] S1: Anyway. What's this, article from
02:02:46 · Jordi Hays, John Coogan
[02:02:46] S1: Anyway. What's this, article from [02:02:49] S0: Evan Armstrong over at The Leverage published an article on who actually makes money when robots work, looking into some of the venture funding that's flowing into humanoid robotics. And he is trying to create a field guide for separating the real companies from the grifters. Talks about one x. The home robot costs $220,000 upfront or $4.99 a month. The website looks like every other VC funded d to c brand from 2015, millennial beige, sans serif, typography, the works buried in the fine print and tech cover tech coverage is the actual product for most chores. A human in a call center will drive the robot around your house via teleoperation while the system records training data for autonomy that doesn't exist yet. And so, there's a big question about how like, what will the actual margins be? And he was trying to dig into that. So, you can go read Evan's piece over on the leverage if you want. You get a seven day trial. [02:03:52] S1: Who actually makes money though when they work? Is it the call center operators?
02:03:52 · Jordi Hays, John Coogan
[02:03:52] S1: Who actually makes money though when they work? Is it the call center operators? [02:03:57] S0: I believe I mean, I I I would not bet against Elon on this. I believe that the the hardware manufacturer will ultimately be the one that makes money in the term. I think that the everything else is more commodity in the stack, but there is going to be a a compounding advantage to actually having the manufacturing capability to build the robots at scale. Now this is years away, but I'm I'm certainly AGI built in the sense that the the teleoperation will become will become less and less important. And and the Tesla model of having a a economically producible product will be very, important. So that would be my take. [02:04:41] S1: In other humanoid news, Brett Adcock was putting a company in the truth zone. [02:04:49] S0: What did say? [02:04:49] S1: Ensign based company, UB Tech, claims to have completed the world's first mass delivery of humanoid robots. And the Shenzhen based company has secured apparently over 112,000,000 in Walker s two orders this year. Brett Adcock says, look at the reflections on this bot and then compare them to the ones behind it. The bot in front is real. Everything behind it is fake. If you see a head unit reflecting a bunch of ceiling lights, that's a give that's a giveaway. It's CGI. So he's putting he's putting UB Tech in the truth zone, and then Christopher comes over the top and puts him in the and it says, such an embarrassing pose for a CEO at this level to make. So people don't necessarily believe that, the the CGI claims.