Latent Space / episode intelligence
Speakers in the public record
Claim mix
Evidence policy
Every row below preserves an exact excerpt. Identified speakers are linked; unresolved voices are labeled and excluded from people counts.
Claim ledger
24 published records
01 / belief
“I think to me the most interesting thing about long inference is like, You're shifting the cost to the customer depending on how much they care about the end result.”
02 / belief
“I, I think what is confusing a lot of people about stability is like they're, they're, they're pushing a lot of things in stable codes, stable l and stable video diffusion.”
03 / belief
“I think the last one we did, the four wars of AI, was the main kind of mental framework for people.”
04 / belief
“I think there's a lot of money that goes into building what GRUK has built as far as the hardware goes.”
05 / uncertainty
“I'm not on LinkedIn, so I'm gonna do horrible on LinkedIn, because I don't know the metagame of LinkedIn.”
06 / uncertainty
“I was gonna say, maybe it's not here, but I don't know if we want to talk about diffusion transformers as like in the alt architectures, just because of Zora.”
07 / observation
“The only way a podcast consistently grows is through your word of mouth, and that helps us book incredible guests and attend great events in our second year.”
08 / belief
“You know I think, I mean, obviously the Voyager paper is like the most basic example where like, now their artifact is like the best planning to do a diamond pickaxe in Minecraft.”
09 / commitment
“If you've seen that code interpreter, that's pretty hard for users to do. So we focus on data and that use case, and we will do that.”
10 / belief
“I think we have a Also, this was the month that we did Compute Provider Month, I think relatively successful.”
11 / belief
“Our third listener is RJ Honecke, who comes from a data science background, but wanted to ask about how we think about learning in public in AI, and how that informs the context with which latent space is created.”
12 / belief
“I am on more of a Yann LeCun side than I am on Jim's side, in the sense that I think that is the vision or the hope that these things can build world models.”
13 / uncertainty
“Maybe, maybe it's the audio in here. I don't know. It depends what, what we get up to later.”
14 / belief
“I think the The challenge, well not the challenge, what they need to figure out is like how do you keep the rag piece always up to date constantly, you know, I feel like the models, you put all this work into pre training them, but then at least you have a fixed artifact.”
15 / recommendation
“Julius is an AI data scientist that helps you analyze datasets, make visualizations, Get insights from the data, and really dive deep into all sorts of data that we have in real life. So, we launched about six months ago, and since then have grown to 300, 000 users several thousand users using us daily to analyze datasets, create visualizations and get insights.”
16 / evaluation
“They have not really commented on why. They released this the headline feature is that it has a 1 million token context window that is multi modal which means that you can put all sorts of video and audio And PDFs natively in there alongside of text and, you know, it's, it's at least 10 times longer than anything that OpenAI offers which is interesting.”
17 / evaluation
“I think that I ki I had to feature this because the, one of the top news of January was that Gemini Pro beat GPT-4 turbo on LM sis for the number two slot to GPT-4.”
18 / evaluation
“5, which we still haven't covered, is very interesting because Gemini magically has fixed all these problems with perfect haystack recall and reasonable latency and cost.”
19 / evaluation
“Whereas this one, I think that the focus for the five directions is just on research that seems more proMECEng than others, because there's all sorts of papers published every single day, and there's no organization.”
20 / evaluation
“One thing that came up on the last day of February, which I should also mention, is EMO coming out of Alibaba, which is also a sort of like video generation and space time transformer that also involves probably a lot of synthetic data as well. And so like, this is of a kind in the sense of like, oh, like, you know, really good generative video is here and It is not just like the one, two second clips that we saw from like other, other people and like, you know, Pika and all the other Runway are, are, are, you know, run Cristobal Valenzuela from Runway was like game on which like, okay, but like, let's see your response because we've heard a lot about Gen 1 and 2, but like, it's nothing on this level of Sora So it remains to be seen how we can actually apply this, but I do think that the creative industry should start preparing.”
21 / evaluation
“I think if anything, RAG's complexity goes up and up the more you use it, you know, because you have more data sources, more things you want to put in there.”
22 / evaluation
“There's been a lot of excitement lately. And, and, like one AI year ago apparently when we launched our 7B AI model, there was a lot of excitement in the buzz, because for the first time, an attention free model beat other transformer models at one trillion tokens at a 7B class.”
23 / commitment
“nd I think the last one that I've recently like heard that was unique for us and is a lot more important is that ultimately this model is owned by everyone because We put it into the Linux Foundation.”
24 / evaluation
“I think that a lot of discussion this month actually was also centered around this concept that People always say, oh, like, transformers don't scale because attention is quadratic in the sequence length.”