High Signal Podcasts Evidence ledger
Method
Browse
How the evidence works
← All source episodes

Machine Learning Street Talk / episode intelligence

DeepMind Genie 3 [World Exclusive] (Jack Parker Holder, Shlomi Fruchter)

5 Aug 2025 15 published claims 1 attributable person
Open original episode ↗ Read the evidence

Speakers in the public record

Tim Scarfe

Claim mix

belief 7evaluation 6preference 2

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.

15 published records

01 / evaluation

But in Genie 2, there was an ST transformer, so a special temporal transformer, which was conceptually quite similar to like a VIT. And there was a latent action model, which means even from non interactive data, you could infer some low cardinality action space.

“But in Genie 2, there was an ST transformer, so a special temporal transformer, which was conceptually quite similar to like a VIT. And there was a latent action model, which means even from non interactive data, you could infer some low cardinality action space.”
Speaker
Tim Scarfe
Publisher
Machine Learning Street Talk

04 / belief

I think probably similar to language that there are some fundamental things about the world that you want to remain consistent. So with the language model, I think even though, as you said, they they can be like sarcastic models, if there are things that stated as facts in their in their context, they will still probably recall them correctly.

“I think probably similar to language that there are some fundamental things about the world that you want to remain consistent. So with the language model, I think even though, as you said, they they can be like sarcastic models, if there are things that stated as facts in their in their context, they will still probably recall them correctly.”
Speaker
Not verified from transcript
Publisher
Machine Learning Street Talk

05 / belief

I think that's what you're alluding to. So I think in in the case of we're still in a similar place, I would say, when we think about the world models because you have to provide the description of the world that you want to to maybe walk in into an experience.

“I think that's what you're alluding to. So I think in in the case of we're still in a similar place, I would say, when we think about the world models because you have to provide the description of the world that you want to to maybe walk in into an experience.”
Speaker
Not verified from transcript
Publisher
Machine Learning Street Talk

06 / belief

I think there is definitely this kind of like a puristic approach, or we should have just 1 model to do everything. But I think when when, you know, a lot of of the challenges with modern machine learning comes from actually building there is a lot of engineering and, software and hardware design that actually you know, to build those things, right, train them and run inference.

“I think there is definitely this kind of like a puristic approach, or we should have just 1 model to do everything. But I think when when, you know, a lot of of the challenges with modern machine learning comes from actually building there is a lot of engineering and, software and hardware design that actually you know, to build those things, right, train them and run inference.”
Speaker
Not verified from transcript
Publisher
Machine Learning Street Talk

07 / belief

I think what we can say about the architecture that might be interesting is that definitely because of the interactive nature of the problem or the setup, then the model is autoregressive.

“I think what we can say about the architecture that might be interesting is that definitely because of the interactive nature of the problem or the setup, then the model is autoregressive.”
Speaker
Not verified from transcript
Publisher
Machine Learning Street Talk

08 / belief

In terms of things like more more abstract representations, I think we're still kind of exploring it, to be honest. But it definitely has a slightly different feel to it because especially for the first person view that you get often when you're experiencing it.

“In terms of things like more more abstract representations, I think we're still kind of exploring it, to be honest. But it definitely has a slightly different feel to it because especially for the first person view that you get often when you're experiencing it.”
Speaker
Not verified from transcript
Publisher
Machine Learning Street Talk

09 / belief

The real world is fundamentally populated by people and other agents. And this is something that we can gain from training on this general purpose wormhole that we just have no other approach, I think, to scalably get this data in a safe way as well.

“The real world is fundamentally populated by people and other agents. And this is something that we can gain from training on this general purpose wormhole that we just have no other approach, I think, to scalably get this data in a safe way as well.”
Speaker
Not verified from transcript
Publisher
Machine Learning Street Talk

10 / evaluation

We are here at Google DeepMind in London, and you guys have just demoed to me something which I think I'm more impressed with this, I think, than anything I've seen probably ever before.

“We are here at Google DeepMind in London, and you guys have just demoed to me something which I think I'm more impressed with this, I think, than anything I've seen probably ever before.”
Speaker
Tim Scarfe
Publisher
Machine Learning Street Talk

12 / evaluation

Perhaps in the future, we might have an outer loop, which makes the system more open ended. But right now, my opinion, Genie 3, like all AI, gives you exactly what you asked for in the prompts and isn't creative on its own.

“Perhaps in the future, we might have an outer loop, which makes the system more open ended. But right now, my opinion, Genie 3, like all AI, gives you exactly what you asked for in the prompts and isn't creative on its own.”
Speaker
Tim Scarfe
Publisher
Machine Learning Street Talk

14 / evaluation

I've also noticed that this model is is even better than v o 3 at things like text. It's because you you would think that they would they would dumb down the model to make it interactive and to make it this sophisticated.

“I've also noticed that this model is is even better than v o 3 at things like text. It's because you you would think that they would they would dumb down the model to make it interactive and to make it this sophisticated.”
Speaker
Tim Scarfe
Publisher
Machine Learning Street Talk

15 / preference

You know, I'm a huge fan of open endedness, for example. And certainly at the moment, when we prompt models, if we're quite generic in what we put in the prompt, then we tend to get quite simplistic answers.

“You know, I'm a huge fan of open endedness, for example. And certainly at the moment, when we prompt models, if we're quite generic in what we put in the prompt, then we tend to get quite simplistic answers.”
Speaker
Tim Scarfe
Publisher
Machine Learning Street Talk
Search evidence