AI, Surveillance, and the Fight for Digital Sovereignty | Near's Illia Polosukhin

**David** (0:04)
Illia, last week, Entropic removed access to Fable 5 after the US government issued an export ban.
They cited safety concerns about possible jailbreaks. Is this the last models that are going to be available to the general public? Are there no more models that are coming to me in my clod because the government is saying no?

**Illia Polosukhin** (0:24)
So, I mean, there's a lot here to unpack, right? First, Entropic released this model and kind of neutered it on a number of dimensions. And everybody was outraged about that because effectively, if you asked it about cryptography, if you asked it about some bioscience, it would fall back to the kind of opus. And initially even not very transparently, right? So I think we get into a stage where there is different parties who are effectively deciding how you can use models, right? And so realistically, this goes back against some of the fundamental, like policing the thoughts, not policing the actions. And so I actually predicted that government will start stepping in because like they kind of feel they need to, even though it's violating some of the fundamental, like freedom of speech kind of basis.
So yeah, I think we'll start seeing the very strange things. And already, this is unrolling now, where every, like, so there's a few things that are interesting here. One is, Entropiq effectively now dictates how you can use these models, right? They also changed their privacy policy for the Fable, right? Which says, hey, even though you're paying us to not keep your data, we'll keep your data for safety purposes, right? So that was kind of one side of this. Then, you know, if you read the reasoning, you know, from David Sachs and kind of some of the other folks from the government side, they're like, hey, they said, it's super dangerous. We have like a reasonable way to use it for effectively finding exploits. Like they didn't patch it. Now, the fact that they also then enacted this export control is a very kind of, of this strange reaction. It's also now puts every single other government in the world on alert that at any time, not just AI, effectively every single internet service can be shut down. Like if you think of it right now, like if you take your phone out of their pocket, it runs on a lot of services that are provided by US.
And if government enacts export control on any of them, it's not just it can be shut down like for your country. It's also like you can shut down everything. Like you can shut down everything because it's also on internet hard. Like if you as a US citizen go abroad, like did you give your phone to somebody else? Like how are they going to detect that? So it's a very strange precedent, honestly, for effective the internet because it violates the whole premise that the internet is borderless. So I think that that was a very strange action to use that methodology.
Even if they were trying to like, hey, you guys should fix this problem, that you said yourself is, hey, security and safety is a big deal. So now going back to the models, I think the Pandora box is open. I think people who are, I think like obviously, Mythos came out, but it was only like available to a few parties. But we in crypto saw hack after hack after hack. It wasn't because people were using Mythos. It was because the models are there.
A lot of them are open source now, are open weights. You can put them in the right harness where you effectively inspect every line for invariants for all these properties, and you can find runnabilities. Like this is happening right now.
Obviously, from kind of experiments and from what I heard, people, Fable is like a good step forward on capabilities. But I really believe it's like a whatever, four-minute mile, as soon as one party can do it, everybody else kind of pulls behind it. And a lot of it is just like how much compute you can put behind it to really keep reinforcement learning and update. So I think to your point, is this the last model you'll be able to access? I think we will have this level of model from open source one way or another within next three to six months. We just saw GLM 5.2 came out at least from benchmarks and Coursery usage. It looks kind of like 4.7, so one generation down from Opus. So still kind of a couple months behind, but this is already something that two months ago was actually like, hey, this is a great product. So I think Pandora's Box is open, this intelligence at scale is coming. I think we as crypto actually always thought in this way, where we assume there is somebody very intelligent on the other side who is trying to break in. The rest of the world hasn't built like that. And so I think this mindset of effectively like, hey, we have a government adversary who is equipped with potentially quantum computers trying to break us. We need to kind of apply that across the society. I've talked about this, I call it DDoSing the government. Right now, government is not designed to actually withstand somebody who has intelligence and an immense amount of patience.