00:19:46 · John Coogan, Jasper Zhang, Jordi Hays, S2
It it it is gaining a lot of ground among among the the futurists [00:19:51] S4: Yeah. [00:19:51] S1: And the and the science fiction fans. The racing simulation hypothesis is is getting getting popular. This is cool. [00:19:57] S0: That's not all. [00:19:58] S1: Wait. Wait. Really quickly on Genie four, is it available yet? What... Because you've played with some of them before. [00:20:04] S2: Yeah. I'd love to get access. I I don't have access to Genie 4 yet. I think it's still on... You have to, like, apply to their, like, early access program. [00:20:12] S1: Okay. I I'm I'm excited for that to to roll out and be able to play. Jordi, what what is the next Google? [00:20:18] S0: They got another model. Multiple models this week. Sundar is back. He says, I did say we'd be releasing our models periodically, introducing Gemini four Carbon, their next frontier model, which follows Argon's release just two weeks ago. Carbon shows the pace at which we're pushing the frontier, and it ranks number one on the recently updated V5 of the Artificial Analysis Intelligence Index at 59. And it also leads across cybersecurity evaluations, scoring 88.2% on SCCBench pro for finding vulnerabilities and 54.8% on exploit gym. Yeah. So very excited to see the reaction here. Seems like a great launch. Elon says it's impressive. And I think, again, they have to at this point have been learning from the sort of series of of benchmark disappointments that they seemingly have had for [00:21:17] S1: Yeah.
00:21:17 · John Coogan, Jordi Hays
[00:21:17] S1: Yeah. [00:21:17] S0: Semi outs About year. Shots. Right? Just about a about a year straight where every single time it's amazing on benchmarks and then, the the sort of lived experience that people have is less impressive. [00:21:30] S1: Thought Gemini three had, like, a good flavor to it. Maybe it was a bit overly benchmaxed, but it felt like it wasn't too verbose. It felt like it sort of answered questions in, like, one page of text, and so I could, like, see it and see the answer and kind of parse it easy as opposed to getting, like, pages and pages and pages, but it was something LLMs do. It could be a little bit much. I was confused by this because it feels like the the the hierarchy... Like, with with, like, Luna, Terra, Sol, Astra, I kind of understand that, like, the moon is smaller than the Earth. The Earth is smaller than the sun. The sun is smaller than the celestial bodies. Same thing with, Anthropic, where it's like a sonnet is shorter than an opus. A mythos is bigger. I kind of understood that. Yeah. This is interesting because they're going in maybe the opposite direction depending on how you rank them. So argon is the eighteenth atomic number, whereas carbon's number six. And so does that leave them open to do a a lithium, a high... A helium, a hydrogen model that's, like, bigger? Are they gonna go that direction? I like that they have a name Do [00:22:43] S0: you think they're gonna do a helium model?
00:22:43 · Jordi Hays, S5, John Coogan, S2
[00:22:43] S0: you think they're gonna do a helium model? [00:22:45] S5: Yes. I I think the helium model could be the biggest model from DeepMind in years. It could change everything. It would be the... It would be significantly bigger than their new carbon model, [00:23:09] S1: Yeah. But it it does sort of beg the question... Yeah. Like, what is the what what is the taxonomy here? It's kinda cool that they left a lot of space between carbon and argon for sodium, magnesium. [00:23:24] S0: Sodium. Sodium. Honestly... [00:23:26] S1: Neon, aluminum. I love salt. Silicon's in there. Phosphorus? Sulfur? [00:23:30] S2: I would have liked if [00:23:31] S1: if if would have liked [00:23:32] S2: some more fear mongering. They could do plutonium Yeah. Fluorine. [00:23:35] S1: Well, that's really deep. So in the... This is what I'm saying. In the deep mind taxonomy right now, plutonium would sound like a really scary model, but it would actually be much smaller than carbon because it goes from carbon's the biggest and best, then argon at eight... At the 18 atomic number is smaller. And then you go way down to plutonium and it's actually like the least effective, the least smart model in theory. I don't know. It's cool. [00:24:03] S0: They should do Gemini five nuke. [00:24:06] S1: Nuke. Yeah. We... Wait. Don't you have a button for that? I thought we had a button for nuke. [00:24:11] S0: No. I think we I think we removed.
00:24:11 · Jordi Hays, John Coogan
[00:24:11] S0: No. I think we I think we removed. [00:24:12] S1: I think it's too too too for the show. Oh, here we go. Tyler has the truth nuke. Gemini plutonium is coming in any minute now. No. The truth nuke is good. Really slow still. The longest soundboard effect we have. But this this all feels like, okay. A year ago, people were talking about, like, the naming schemes at the labs don't make any sense. OpenAI was probably the worst offender with, like, four o, and then, like, you'd go to four five, which was bigger a model, but it was slower. And then there was, like, Omni versus Fastmo, and it was, like, too much. And everyone was like, I don't get it. And then and then it was, like, five four, five five, five six, And then, eventually, OpenAI said, okay. We're we're going to use a human legible, not just purely numbers. We're gonna go Sarah, Luna, Soul, Astra. And that, I feel like, makes a lot of... A lot more sense to people. It's just more intuitive. And so I like that OpenAI has adopted that format. I like that Google has adopted that format. And I hope that the version numbers kind of melt away. And you're just in the world of, like, okay, like, big model for this task, small model for that task. And and it's a little bit it's a little bit easier to just intuitively get a grasp of whatever model is capable Anyways. [00:25:34] S0: But... Amazing benchmarks. Good luck. Excited to see the reactions. Yeah. And... [00:25:38] S1: And the ramp data. [00:25:39] S0: Yeah. In other news. Yep. OpenAI.
00:25:39 · Jordi Hays, John Coogan
[00:25:39] S0: Yeah. In other news. Yep. OpenAI. [00:25:43] S1: First, let me tell you about Cisco. Critical infrastructure for the AI era. Unlock seamless real time experiences and new value with Cisco. Take me through it. [00:25:50] S0: OpenAI plans to buy Google's chips as part of its AI build out. The new... Like, first we had... So now we have TPUs. Mhmm. Last week, it was it was OpenAI models on GCP. Yeah. Feels like a start of a new era? [00:26:10] S1: Yeah. I mean, sort of a new era. I don't know. I mean, yeah. I guess OpenAI was founded because they were like, Google's bad. Right? [00:26:19] S0: Yeah. That's what [00:26:19] S1: I'm saying. Yeah. That's a good point. Now, it's like, they're not that bad. But this is the same thing when I get there off again, Elon. It's like, you know, very very diametrically opposed, you know, politically. [00:26:30] S0: But yeah. And... [00:26:31] S1: They got compute when you compute. Yeah. So Yeah. Business above all. Business solves everything. Let's give it up for business. [00:26:39] S0: Let's give it up for business. [00:26:41] S1: Business solves everything. [00:26:42] S0: Dylan Patel says semi analysis clients had this already. Woah. Screenshot. [00:26:48] S1: Little victory lap.
00:26:48 · John Coogan, Jordi Hays, Ben Liu
[00:26:48] S1: Little victory lap. [00:26:49] S0: AI loves largest lease ever more massive deals in the pipeline. AMD taking share, question mark. OpenAI's quarter is one eight gigawatt multi phase commitment in Ohio with NVIDIA backstopping 4.25 gigawatts. It's the largest single leasing commitment we have ever seen. Cool. In addition, we believe that OpenAI is discussing with NeoClouds for both GPU a as a service and TPU as a service. We think there's a one gigawatt TPU commitment, and OpenAI is looking for sites. Yep. Anthropic leased 1.7 gigawatts this quarter, with notably three sixty megawatts in Queensland, Australia. Let's give it up for Compute Down Under. Signed on the September 16 with operations starting in the first quarter of twenty eight. The largest overseas leased by an AI lab this quarter. The 1.7 gigawatts includes one gigawatt backstop by Google, but also a large site where we aren't sure yet who will provide the credit support. [00:27:42] S6: Do you [00:27:42] S1: think AI researchers are gonna be okay with doing inference in Australia? That's kind of controversial, right, because they're known for throwing another shrimp on the Barbie. That's rough. You're gonna have to do a truth. It's like, no more throwing shrimp on the Barbie. Just fosters. Just enjoy your fosters. [00:27:59] S0: Enjoy your fosters.