Speakers in the public record
Claim mix
belief 26evaluation 7uncertainty 4recommendation 2prediction 2observation 1commitment 1preference 1
Evidence policy
Every row below preserves an exact excerpt. Identified speakers are linked; unresolved voices are labeled and excluded from people counts.
Claim ledger
The useful parts, with receipts.
44 published records
“Yeah. And the big thing was like the mixed trial price fights, you know, and I think now it's almost like there's nowhere to go, like, you know, Gemini Flash is like basically giving it away for free.”
- Publisher
- Latent Space
“Sombra to congregate, and then the AI engineer summit. And that's why when I look at our growth chart, it's kind of like a proxy for like the AI engineering industry as a whole, which is almost like, like, even if we don't do that much, we keep growing just because there's so many more AI engineers.”
- Publisher
- Latent Space
“Uh, I, yeah, I, you know, I think We will go over the individual, uh, talks in a separate episode.”
- Publisher
- Latent Space
“We don't know what the small model size is, but yeah, it's probably in the double digits or maybe single digits, but probably double digits.”
- Publisher
- Latent Space
“I think the collective terminology has been inference time, and I think that makes sense because test time, calling it test, meaning, has a very pre trained bias, meaning that the only reason for running inference at all is to test your model.”
- Publisher
- Latent Space
“As someone who also does charts, XAI is continually snubbed because they don't work well with the benchmarking people.”
- Publisher
- Latent Space
“I think in AI you see this a lot, which is like a lot of stars, a lot of interest at a rate that you didn't really see in the past in open source, where nobody's running to start.”
- Publisher
- Latent Space
“I think that's like one of the hardest things about NeurIPS. It's like the long tail.”
- Publisher
- Latent Space
“Everyone knew, like, F1 we had a preview at the Fireworks HQ, and then [01:37:00] I think some other labs did it, but I think R1 and QWQ, Quill, from the Quent team, Both Alibaba affiliated, I think, are the leading contenders on that front end.”
- Publisher
- Latent Space
“What do you pay for node? Like, I don't know what the game world was like. Maybe the starting price was 1B.”
- Publisher
- Latent Space
“Interestingly, I think the same person responsible for artifacts and canvas, Karina, officially left Anthropic after this to join OpenAI on the rare reverse moves.”
- Publisher
- Latent Space
“I think even, maybe if you ask them, like, they wouldn't necessarily draw a straight line.”
- Publisher
- Latent Space
“I feel like this is almost like, you know, Now the agent thing needs to happen. And I think that's really the unlock.”
- Publisher
- Latent Space
“I think all these episodes that are like [01:45:00] summarizing things that people care about, but they're disparate.”
- Publisher
- Latent Space
“I think every, I mean, there should be, just like every consumer product is going to have a, going to eventually want a gateway, you know, for, for managing their requests and ops tool, you know, that kind of stuff, um, code interpreter for maybe not exposing the code, but executing code under the hood for sure.”
- Publisher
- Latent Space
“Another discovery, I think, um, Ilya actually worked on a previous version called GPT 0 in 2021.”
- Publisher
- Latent Space
“Uh, I think that, I think you've been, uh, the, the, a dream partner to, to build Lanespace with.”
- Publisher
- Latent Space
“I think like a lot of AWS customers, you know, they do this big reserve instance contracts and now they got to use their money.”
- Publisher
- Latent Space
“I think that is where, uh, At least one hour of software experience to go, then it will be another few years [01:33:00] for that to happen in real life outside of the screen.”
- Publisher
- Latent Space
“I think the point of the talk was like everybody, we're scaling these chips, we're scaling the compute, but like the second ingredient which is data is not scaling at the same rate.”
- Publisher
- Latent Space
“People have actually, I would, I would say I've been surprised by how well received that was.”
- Publisher
- Latent Space
“Voice mode, yeah. I think most people have tried it by now. Because it's generally available.”
- Publisher
- Latent Space
“NVIDIA, I think we continue to talk about, I think I was at the Taiwanese trade show, Comtex, and saw him signing, you know, You know, women body [01:30:00] parts. And I think that was maybe a sign of the times, maybe a sign that things have peaked, but things are clearly not peaked because they continued going.”
- Publisher
- Latent Space
“I think if your, if your, your job is a, at least AI content creator or VC or, you know, someone who, whose job it is to stay on, stay on top of things, you should already be spending like a thousand dollars a month on, on stuff.”
- Publisher
- Latent Space
“OpenAI now needs then more money because they need to support those products and I think maybe their bet is like 1 billion we can get to the thing.”
- Publisher
- Latent Space
“Um, and that one, I would say Risa is obviously a star, but she's been on every episode, every podcast, but Isamah, I think, you know, actually being the guy who worked on the audio model, being able to talk to him, I think was, was a great gift for us.”
- Publisher
- Latent Space
“I think like what we have today is not secure enough because it's like normal security when like this is literally a state level interest.”
- Publisher
- Latent Space
“I think we can also use this as part of a general theme of the safety wing of OpenAI leaving.”
- Publisher
- Latent Space
“Uh, I think, um, I have been asked my opinion on this before, and I said, I think I said it on a podcast, which is like, the main layer that you need is [01:38:00] the separate off roles, so that you don't assume it's a human, um, doing these things.”
- Publisher
- Latent Space
“There was maybe like a hype winter, but I don't know if that counts as a real winter.”
- Publisher
- Latent Space
“I think the quote that I highlighted in AI News was that it is the best, like Blackwell is the best selling series.”
- Publisher
- Latent Space
“Ilya saying the word agentic on stage at Eurips, it's a big deal. Satya, I think also saying that a lot these days.”
- Publisher
- Latent Space
“I think today small is actually a more nuanced discussion, you know, that people weren't really having before.”
- Publisher
- Latent Space
“I mean, I'm curious, I think Dylan, At the debate he said SweetBench 80 percent was like a soap for end of next year as a kind of like, you know, watermark that the moms are still improving.”
- Publisher
- Latent Space
“Bash transcription, I would say Whisper, is something that you should be using on a as much as possible.”
- Publisher
- Latent Space
“Everybody needs to run code, right? And I think now all the products and the everybody's graduating to like, okay, it's not enough to just do chat.”
- Publisher
- Latent Space
“Um, and I think today the, the problem is that, Yeah, the agents are, that most people are building are good at following instruction, but are not as good as like extracting them from you.”
- Publisher
- Latent Space
“Yes. I prefer to look at actual downloads, not at stars on GitHub. So if you look at, you know, Lanchain still growing.”
- Publisher
- Latent Space
“So likewise at Meta, likewise at OpenAI, likewise at the other labs as well. So like the GPU ultra rich are going to keep doing that because I think partially it's an article of faith now that you just need it.”
- Publisher
- Latent Space
“I think to me that the most interesting is like rest and GraphQL is almost more interesting in the world of agents because agents could come up with so many different things to query versus like before I always thought GraphQL was kind of like not really necessary because like, you know what you need, just build the rest end point for it.”
- Publisher
- Latent Space
“Oh yeah, magically going to close the gap between the closed source and open source. So basically I think my advice to people is keep track of the slow cooking of benchmark language because the labs that are not that frontier will keep measuring themselves on last year's benchmarks and then the labs that are actually frontier will Tell you about [01:09:00] benchmarks you've never heard of and you'll be like, Oh, like, okay, there's, there's new, there's new territory to, to, to go on.”
- Publisher
- Latent Space
“I think inference time compute is bad for open source just because, you know, Doc can donate the flops at training time, but he cannot donate the flops at inference time.”
- Publisher
- Latent Space
“Because I think there's a lot of doubt or questions about where ML engineering stops and AI engineering starts.”
- Publisher
- Latent Space
“I think the Researchers that I talked with at NeurIPS were kind of positive on this because basically you need private test [00:14:00] sets to prevent contamination.”
- Publisher
- Latent Space