High Signal Podcasts Evidence ledger
Method
Browse
← All source episodes

Latent Space / episode intelligence

Beating Google at Search with Neural PageRank and $5M of H200s — with Will Bryk of Exa.ai

10 Jan 2025 19 published claims 3 attributable people

Speakers in the public record

Claim mix

belief 9preference 3uncertainty 2evaluation 2prediction 2commitment 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.

19 published records

01 / uncertainty

Uh, which felt like it hadn't changed in a decade, uh, because it really hadn't. And it, like, you would give it a simple query, like, I don't know, uh, shirts without stripes, and it would give you a bunch of results for the shirts with stripes.

“Uh, which felt like it hadn't changed in a decade, uh, because it really hadn't. And it, like, you would give it a simple query, like, I don't know, uh, shirts without stripes, and it would give you a bunch of results for the shirts with stripes.”
Speaker
Will Bryk
Publisher
Latent Space

04 / belief

Um, and so I think if you remove that whole, like, oh, that's impossible, and you just think really clearly about like, what's now possible with like what, what they've done with O1, it's easy to see how that scales.

“Um, and so I think if you remove that whole, like, oh, that's impossible, and you just think really clearly about like, what's now possible with like what, what they've done with O1, it's easy to see how that scales.”
Speaker
Alessio Fanelli
Publisher
Latent Space

05 / belief

I would say also this, like this slop that is returned by, by traditional search engines, like Google and Bing, you have the slop is then, uh, transferred into the, these LLMs in like a search GBT or, you know, our other systems like that.

“I would say also this, like this slop that is returned by, by traditional search engines, like Google and Bing, you have the slop is then, uh, transferred into the, these LLMs in like a search GBT or, you know, our other systems like that.”
Speaker
Will Bryk
Publisher
Latent Space

07 / belief

I would say if you were to take the exit being O1 for search, literally, you really need to prioritize search trajectories, like almost maybe paying a bunch of grad students to go research things.

“I would say if you were to take the exit being O1 for search, literally, you really need to prioritize search trajectories, like almost maybe paying a bunch of grad students to go research things.”
Speaker
Shawn Wang
Publisher
Latent Space

08 / belief

I think, well, I mean, there's definitely a question of like inference speed and if certain things like O1 takes a long time, because that's the thing.

“I think, well, I mean, there's definitely a question of like inference speed and if certain things like O1 takes a long time, because that's the thing.”
Speaker
Will Bryk
Publisher
Latent Space

09 / belief

I'm like, I'm sure there's not what event, you know, but I'm curious, like, just like how people there's always like this, I think for a little bit, it went away about like startups and kind of like hustle culture and like all of that.

“I'm like, I'm sure there's not what event, you know, but I'm curious, like, just like how people there's always like this, I think for a little bit, it went away about like startups and kind of like hustle culture and like all of that.”
Speaker
Alessio Fanelli
Publisher
Latent Space

11 / commitment

Um, but I mean, ultimately we want to build a perfect search engine, which definitely for a lot of queries involves your, your personal information, your company's information. And so, yeah, I mean, the grandest vision of Exa is perfect search really over everything, every domain, you know, we're going to have an Exa satellite, uh, because, because satellites can gather information that, uh, is not available publicly.

“Um, but I mean, ultimately we want to build a perfect search engine, which definitely for a lot of queries involves your, your personal information, your company's information. And so, yeah, I mean, the grandest vision of Exa is perfect search really over everything, every domain, you know, we're going to have an Exa satellite, uh, because, because satellites can gather information that, uh, is not available publicly.”
Speaker
Will Bryk
Publisher
Latent Space

12 / preference

Saying, even if you've crafted a great system prompt, you want to be part of the process itself. Uh, because the system prompt doesn't, it doesn't capture everything.

“Saying, even if you've crafted a great system prompt, you want to be part of the process itself. Uh, because the system prompt doesn't, it doesn't capture everything.”
Speaker
Will Bryk
Publisher
Latent Space

14 / prediction

So we've basically developed the next version of Exa, which is the ability to get a near perfect list of results of whatever you want. And what that means is you can make a complex query now to Exa, for example, startups working on hardware in SF, and then just get a huge list of all the things that match.

“So we've basically developed the next version of Exa, which is the ability to get a near perfect list of results of whatever you want. And what that means is you can make a complex query now to Exa, for example, startups working on hardware in SF, and then just get a huge list of all the things that match.”
Speaker
Will Bryk
Publisher
Latent Space

16 / evaluation

ChatGPT was an interesting like thing that expanded people's understanding of search because you start using ChatGPT for a few hours and you go back to Google and you like paste in your code and Google just doesn't work and you're like, oh, wait, it, Google doesn't do work that way.

“ChatGPT was an interesting like thing that expanded people's understanding of search because you start using ChatGPT for a few hours and you go back to Google and you like paste in your code and Google just doesn't work and you're like, oh, wait, it, Google doesn't do work that way.”
Speaker
Will Bryk
Publisher
Latent Space

18 / prediction

You could have a world and we are going to have this world where you have like GP five level systems and beyond that could like answer any complex request.

“You could have a world and we are going to have this world where you have like GP five level systems and beyond that could like answer any complex request.”
Speaker
Will Bryk
Publisher
Latent Space

19 / preference

And, uh, okay, if we need a nap, then we should get a nap pod. It's crazy to me that there aren't nap pods in lots of companies because like I get tired all the time.

“And, uh, okay, if we need a nap, then we should get a nap pod. It's crazy to me that there aren't nap pods in lots of companies because like I get tired all the time.”
Speaker
Will Bryk
Publisher
Latent Space
Search evidence