Limitless: An AI Podcast

Ejaaz:
Last Tuesday, Anthropic released the most powerful AI model the world has ever

Ejaaz:
seen. And that exact same week, by Friday at 5.21pm, they received a letter

Ejaaz:
from the government asking them to shut down access to their latest Claude Fable

Ejaaz:
and Claude Mythos models.

Ejaaz:
90 minutes later, Anthropic shut down access for everyone, citing that anyone

Ejaaz:
who isn't an American citizen or American by nature doesn't get access to this model.

Ejaaz:
Now, the official reason for this was security researchers external to Anthropic

Ejaaz:
had discovered a jailbreak, a way to

Ejaaz:
access or bypass the safeguards on Anthropic's Fable model to get access to

Ejaaz:
a restricted version, which could be used for nefarious purposes,

Ejaaz:
such as designing dangerous bioweapons or exploiting cybersecurity weaknesses in software.

Ejaaz:
Now, in the hands of the wrong people, this could be potentially dangerous,

Ejaaz:
such as China versus the US.

Ejaaz:
The craziest part about all of this is it came from an internal partner.

Ejaaz:
Amazon, who owns the largest individual stake in Anthropic as an investor were

Ejaaz:
the ones that released details of the jailbreak to the government.

Ejaaz:
They had the option of picking up the phone and calling Dario and saying,

Ejaaz:
hey, there's a jailbreak. Let's figure out how to fix this. But they decided

Ejaaz:
to contact Trump himself and report this.

Ejaaz:
And so it unpacks and unravels this entire story around governments taking a

Ejaaz:
nationalized stake in the Frontier AI labs and determining who gets access to which model.

Josh:
Last week was a roller coaster of a week. I mean, we got Fable 5 finally.

Josh:
The Mythos class model was delivered on Tuesday.

Josh:
And Friday at 5.21 p.m., after everyone is getting ready to go home,

Josh:
the U.S. government drops this notice on their desk, saying that they are initiating

Josh:
an export control directive citing national security authorities.

Josh:
The order was to suspend all access to Fable 5 and Mythos 5 by any foreign national,

Josh:
whether inside or outside of

Josh:
the United States, including Anthropik's own foreign national employees.

Josh:
This, of course, is almost impossible to enforce. How do you tell where anyone is from?

Josh:
So Anthropic says there's no clean way to wall off foreign nationals across

Josh:
global consumer products. So in order to comply, they disabled models for all

Josh:
customers, both Mythos and Fable class models.

Josh:
Following this, they also shared the initiative with AWS, which revoked Fable

Josh:
5, all their cloud services.

Josh:
And now where we sit today, everybody is using Opus 4.8 and Sonnet models because

Josh:
Fable and mythos are no longer allowed to be used this

Josh:
sucks um i was getting really used to using fable as a model it is an unbelievably

Josh:
remarkable model in terms of how powerful it is and it feels like now i woke

Josh:
up today i went to use claude and

Josh:
being faced with opus 4.8 i'm like oh man like it's a good model but i really

Josh:
miss my my crown jewel i miss

Josh:
i feel like i am less capable today than i was yesterday and that that feeling alone kind of sucks

Ejaaz:
We've crossed a very important boundary. And it's the first time we've ever

Ejaaz:
done this. This is the first instance of an AI model which has been publicly

Ejaaz:
released and then withdrawn.

Ejaaz:
And I'm feeling the withdrawal effects.

Ejaaz:
As you know, I've been using Fable 24-7 since it got released.

Ejaaz:
And when it got pulled, I was like, I think it was like out on a Friday night and I saw the news.

Ejaaz:
And that was the other thing, by the way, this happened like at the end of day,

Ejaaz:
at the end of the week, when no one was expecting this.

Josh:
Yeah, we got a lot of complaints to get into about this one.

Ejaaz:
Yeah, there's a lot of complaints that I have and that I'm sure many other people have.

Ejaaz:
But the point is, this is the first clear example of a retail accessible model

Ejaaz:
being pulled. And it was government determined. And so it brings into question,

Ejaaz:
obviously, a host of different topics and discussions, but the main one being,

Ejaaz:
is the government overreach too much? Is this a line too far?

Ejaaz:
And then it gets into an interesting discussion around who gets access to this, right?

Ejaaz:
So the clear statement that's been made by the government is,

Ejaaz:
if you aren't an American citizen or other, if you aren't American,

Ejaaz:
if you weren't born here, you don't get access to this model.

Ejaaz:
So technically on this show, Josh, that would mean that, you know,

Ejaaz:
you get access, there's a world where you get access to this model and future

Ejaaz:
models, and I don't purely because of nationality-based stuff.

Ejaaz:
So it gets into a bunch of different things that i have some pretty strong opinions

Ejaaz:
on but firstly like one earth is a is an export export control like what does that mean exactly.

Josh:
I had to look this up because i wasn't sure myself and it's basically these

Josh:
u.s rules that govern the transfer of sensitive

Josh:
technology and information to foreign persons in a foreign destination so basically

Josh:
you don't want to export the technology or the information to other countries

Josh:
because it is dangerous it can be used for a malpractice same is true with a

Josh:
lot of military information this is being treated as such.

Josh:
So for dual use technology, they call it, it has to run through the Bureau of

Josh:
Industry and Security, which is inside the Commerce Department,

Josh:
which is under the Export Administration Regulation. There's just like a lot

Josh:
of regulation involved in this.

Josh:
But basically, the United States deemed this as something that is not safe to

Josh:
be exported. And because Mythos, Class Models, and Fable are being used by foreign

Josh:
entities, regardless of the foreign entity status, they must be shut down immediately.

Josh:
And because it is nearly impossible to tell where the source of that API key

Josh:
is coming from or where the destination of that API is coming from,

Josh:
the API call, they can't actually determine with 100% certainty

Josh:
All the people that are being delivered tokens are not from foreign entities.

Josh:
They are no longer allowed to serve tokens to anybody.

Josh:
And then there's this nuance worth noting where U.S. citizens and green card

Josh:
holders and other certain protected individuals are exempt from this deemed

Josh:
export rule, but it gets really messy.

Josh:
And it's just this incredibly complicated and impossible situation that I would

Josh:
imagine for Anthropics to navigate because you just can't really tell.

Josh:
So as a result, the models are gone.

Josh:
And we're in this weird limbo period now where the frontier model has actually

Josh:
been removed and the frontier for the first time ever like you mentioned it's

Josh:
actually moved backwards a notch instead of forwards in terms of public facing

Josh:
capability and that's been

Josh:
it's set a really interesting and kind of like unnerving precedent as we move

Josh:
forward through this like as we navigate through this new world that we're going through

Ejaaz:
Well i mean this isn't the first time that anthropic and the government have

Ejaaz:
got into a fight these guys have had a relationship they have they have a past

Ejaaz:
and i think it's important to kind of like recap what that past looks like.

Ejaaz:
So literally in the last 12 months, Anthropic has got into spats and fights

Ejaaz:
with the government, disagreements, and they're trying to kind of resolve their outcomes.

Ejaaz:
But there's a bit of bitterness between the two.

Ejaaz:
So I think running through this, Feb 27th, 2026, so that's literally,

Ejaaz:
you know, a couple of months ago, there was a federal blacklist that was put out for Anthropic.

Ejaaz:
Now, the reason why this surfaced was Anthropic's models were being used by

Ejaaz:
the Pentagon or the Department of War to carry out military operations,

Ejaaz:
to design and strategize around how they execute on that.

Ejaaz:
And this was the first known instance. CLAWD was the default model that was

Ejaaz:
used by every single government institute.

Ejaaz:
Anthropic expressed disdain and disinterest in this, saying that,

Ejaaz:
you know, there should be more regulation about this. There should be more transparent comms.

Ejaaz:
The government did not like that. They got into a back and forth about it.

Ejaaz:
They tried to come to an agreement. Ultimately, they did not.

Ejaaz:
And the government's response was to blacklist Anthropics models.

Ejaaz:
Now, what the blacklist basically looked like was no government entity is allowed

Ejaaz:
to use Claude. They can't use it.

Ejaaz:
And they started signing partnerships with Anthropics rivals such as,

Ejaaz:
XAI, OpenAI, and Google. Now, since then, they've reverted back on that.

Ejaaz:
Why? Because Anthropic created this new model called Mythos 5,

Ejaaz:
which was originally restricted. The original Mythos Preview model was ready.

Ejaaz:
Around this time, actually, I think like in February or maybe just after that.

Ejaaz:
And it's been kind of like within restricted boundaries for a while.

Ejaaz:
And then we have this new safeguarded version, Fable 5, which is also more advanced

Ejaaz:
than Preview that has been publicly released and then withdrawn.

Ejaaz:
And so since then, the government has kind of gone back on their word and saying,

Ejaaz:
well, actually, because this model is so powerful, we need to understand how

Ejaaz:
to harden our own security systems. So they've worked with Anthropic.

Ejaaz:
And it's been kind of like this bittersweet sort of relationship.

Ejaaz:
And finally, when Anthropic started to file for an IPO confidentially and publicly

Ejaaz:
released this new model, apparently, according to reports, the government didn't

Ejaaz:
want them to publicly release it.

Ejaaz:
But Anthropic was like, we need to do this. we need to give this to the public,

Ejaaz:
we'll give you a safe version of this.

Ejaaz:
Like, look at how restrictive our safeguards are. You can't get behind this.

Ejaaz:
And then when this jailbreak was exposed to the government, this was the final

Ejaaz:
line drawn and they've gone back on it and basically saying,

Ejaaz:
I'd probably ban this or there'll be repercussions. And that's what I've probably just had to do.

Josh:
Yeah. And now might be an interesting time to actually note the specifics of

Josh:
what they deemed unsafe in order to cause this ban to happen,

Josh:
because it wasn't just like, oh no we don't like this this is a ban it was actually

Josh:
came in the form of a perceived jail break so the report is that

Josh:
someone who we kind of may might have an idea of it's been publicly reported

Josh:
that we might have an idea of someone someone credible reported a jailbreak

Josh:
to united states government and that is the thing that they blamed it on and

Josh:
like you mentioned i find this really

Josh:
troubling and difficult to wrap my head around because so much of the backlash that came from

Josh:
fable five being released was the fact that it was so sensitive to these bio

Josh:
threats it was so sensitive to cyber security that it actually wouldn't answer

Josh:
a lot of questions that were adjacent like i remember, EJ, as you mentioned,

Josh:
you asked, what is mitochondria?

Josh:
And it couldn't even answer that because it was related to bio.

Josh:
So it routed it through Opus 4.8, the safer model that you know you could trust the safeguards from.

Josh:
So there was every effort in the world to go through and make sure that this

Josh:
was safe enough to be released to the public and then give the public access

Josh:
to this amazing frontier model. That way it wasn't gatekept and held privately

Josh:
only for a specific set of companies that are in Project Glasswing.

Josh:
This clearly wasn't enough. There was an alleged jail break that was discovered

Josh:
that was reported to the government. The government said, no,

Josh:
no, no, we can't have this and then went to anthropic and that's where the problems started

Ejaaz:
I mean, at this point, we should probably reveal who broke the news to the government

Ejaaz:
that led to Anthropic needing to ban their models.

Ejaaz:
It was none other than Amazon, specifically the CEO, Andy Jassy.

Ejaaz:
So to give some context here, Amazon

Ejaaz:
is, if not, I think it's the biggest individual investor in Anthropic.

Ejaaz:
They own around 17% to 20%, but they're also their main cloud partner.

Ejaaz:
All of the chip and card infrastructure that is used to train inference and

Ejaaz:
distribute Claude to all government partners, Enterprise, which,

Ejaaz:
by the way, is Anthropic's biggest revenue income provider, as well as retail,

Ejaaz:
kind of is facilitated through AWS, through Amazon's cloud service.

Ejaaz:
So a major alignment between these two companies.

Ejaaz:
Andy Jassy has a security team, and they were testing out Claude Fable,

Ejaaz:
specifically its safeguards, and they surfaced a potential jailbreak which could be exploited by China.

Ejaaz:
Now, if you're Andy Jassy, you have two decisions to make here.

Ejaaz:
You have probably some form of a government obligation because they've done

Ejaaz:
work with a lot of the government officials to reach out to them and say,

Ejaaz:
hey, there's this jailbreak, by the way. Let's work on this to fix this with Anthropic.

Ejaaz:
And then you're a major investor in Anthropic. You should go to them and say,

Ejaaz:
hey, by the way, there's this jailbreak. Let's try and fix this.

Ejaaz:
Andy Jassy could have gone to both people. He just went to one.

Ejaaz:
It was the government and the Department of Commerce, but I think also the Department

Ejaaz:
of War specifically and said, hey, if China gets access to this particular jailbreak

Ejaaz:
that we found, there's going to be an issue.

Ejaaz:
So Pete Hexeth and Secretary Scott Besson.

Ejaaz:
Panic. And they're like, okay, this jailbreak is real. They validate it internally.

Ejaaz:
And they reach out to Anthropic within the hour. And they basically say,

Ejaaz:
hey, we found this jailbreak. We need to speak to Dario.

Ejaaz:
Now, the funniest part about all of these reports is apparently Dario was at

Ejaaz:
a wellness retreat. Now, I don't know if that is true, but he was unreachable

Ejaaz:
for about an hour and 15 minutes to which he then responded,

Ejaaz:
got on a call with the government, and they had this discussion.

Ejaaz:
Now, Dario had a very important statement to make, which was.

Ejaaz:
If this is a jailbreak, it's likely a non-universal jailbreak.

Ejaaz:
Now, there's two types of jailbreaks. There's universal and non-universal.

Ejaaz:
A non-universal jailbreak is a hyper-localized specific jailbreak that only

Ejaaz:
a few niche actors, if they wanted to exploit it, could end up doing.

Ejaaz:
A universal jailbreak is usually the ones that people get scared about,

Ejaaz:
where anyone and everyone with access to Claude Fable, for instance,

Ejaaz:
would be able to get around it and go to the restricted version.

Ejaaz:
What Dario was saying is this is a very specific instance that not really a

Ejaaz:
lot of people know about. And we can patch and it's not really a big deal.

Ejaaz:
We don't need to impose a worldwide ban on people that aren't American individuals.

Ejaaz:
We can fix this. China's not going to get access to this.

Ejaaz:
And reportedly, there's been a back and forth where, again, both of these entities,

Ejaaz:
Anthropic and the government, can't come to an agreeance on any of this.

Josh:
Yeah, it's a lot of like finger pointing and speculation.

Josh:
And we're trying to like piece together the puzzle based on these public reports

Josh:
from news publications but we do have some official commentary from anthropic

Josh:
themselves they published a blog post all about it and just

Josh:
to kind of double click on that jailbreak feature they walked through exactly

Josh:
what happened and their take on what happened and they said no testers have

Josh:
been able to find a universal jailbreak like you mentioned a jailbreak method

Josh:
that can very broadly bypass the model safeguards

Josh:
unlocking a wide range of cyber capabilities so they said we suspect that perfect

Josh:
jailbreak resistance is not currently possible for any model provider.

Josh:
Every safeguard used in the industry is vulnerable to non-universal jailbreaks,

Josh:
which can elicit some cyber information in specific circumstances,

Josh:
and it is likely that universal jailbreaks will eventually be found in the future.

Josh:
And they stated this clearly when they released Fable 5. Given that perfect

Josh:
jailbreak resistance does not appear to be widely possible today,

Josh:
Anthropic adopted a defense in-depth strategy with Fable 5, where they aimed

Josh:
to make jailbreaks either narrow...

Josh:
In the case of non-trivial jailbreaks or very expensive to produce.

Josh:
So the average person is just not going to be able to do this.

Josh:
And then they combine this with this monitoring to detect and shut down any successful attacks.

Josh:
So they're saying, hey, it's not possible to protect a jailbreak from any model.

Josh:
Like all of these models are susceptible to them, but they're susceptible to

Josh:
them on a very narrow band in which it can be detected and then quickly shut down.

Josh:
And if your basis for shutting down a model is on one of these narrow jailbreaks

Josh:
that is available across every model, then that is unfair and improper.

Josh:
It's kind of the argument that they're making.

Josh:
And that logic seems to make sense. Like if you go on X, there's this guy,

Josh:
Pliny the Liberator, I think his name is. And he always has these very like

Josh:
weird, obscure, narrow use cases in which you can get the model to say funny

Josh:
things that like it's not really supposed to say.

Josh:
But that is a novel case that is narrow in scope. And a lot of his accounts get shut down.

Josh:
It's not a broad facing, you could almost say malware, where there's like a

Josh:
prompt that I could put into my LLM and then it will surface me these unfiltered

Josh:
results and that's kind of where the discrepancies are it feels like anthropics is like

Josh:
dude you guys don't understand like we we know this model we understand the

Josh:
risk we've been very public about the risks of this here's what we're doing

Josh:
to stop them here is how we put in the safeguards

Josh:
And the government is saying, oh, no, but there is risk. And because there's

Josh:
a non-zero percent chance of risk, you got to shut this whole thing down.

Josh:
And they're like, no, but this is this every model has risk.

Josh:
That's the nature of the technology.

Josh:
And for some reason, in this instance, there's a zero tolerance policy for this.

Josh:
And therefore, there's no availability currently to actually use the model.

Ejaaz:
I think it's important to understand like both parties' sides of the stories, right?

Ejaaz:
Like if you're the government, right, your default thinking is if a tool,

Ejaaz:
an unrestricted version of this tool gets into the wrong hands,

Ejaaz:
they could hack our government databases and get access to really important

Ejaaz:
information about military operations.

Ejaaz:
Stuff that we're doing that we don't want our adversaries to know about,

Ejaaz:
security access, all these kinds of things which could be used against them.

Ejaaz:
And their kind of view is these adversaries do not care. They have no kind of

Ejaaz:
like maybe moral ethics around any of this. They're not going to be like,

Ejaaz:
oh, okay, well, yeah, this model could be super dangerous.

Ejaaz:
They're just going to use it, right? And so they take a strict kind of like,

Ejaaz:
black or white policy where if there's any kind of possibility or chance for

Ejaaz:
this to happen, they're going to put a ban on it until they figure out what that fix is.

Ejaaz:
And on the tropic side, they're like, hey, listen, we are the experts in knowing

Ejaaz:
how to protect against this.

Ejaaz:
You need to give more weight to our wording, like to our voice,

Ejaaz:
rather than just outright kind of like ban it.

Ejaaz:
So I see, like, even if there's a 1% chance that this could potentially happen,

Ejaaz:
that 1%, if it does happen, could be catastrophic for the entire kind of,

Ejaaz:
like, world in the government's kind of view.

Ejaaz:
Now, the funny thing about this is, before we started recording...

Ejaaz:
Josh, you mentioned that usually when you have a product release,

Ejaaz:
that's specifically software, the jailbreaks, they come out pretty quickly, right?

Ejaaz:
And that happened in the case of Claude Fable. When it got released,

Ejaaz:
there were one or two jailbreaks.

Ejaaz:
One was discovered by playing the Liberator that you just mentioned,

Ejaaz:
but the other one was this one.

Ejaaz:
I want to give you guys an idea of this is an example of a non-universal jailbreak

Ejaaz:
where you can share your ideas of prompts and examples of how to use a model

Ejaaz:
on a public forum, on a public domain, right?

Ejaaz:
And what one individual realized is the way that these safeguards are set up.

Ejaaz:
So Claude Fable 5 is frustrating or was frustrating to a lot of people because

Ejaaz:
you couldn't ask simple questions about biology, chemistry.

Ejaaz:
How a model works, or even how you might want to build a model,

Ejaaz:
because there were strict safeguards put in place saying that you might be using

Ejaaz:
this for nefarious purposes.

Ejaaz:
Therefore, I'm going to re-divert you to an old model Claude Opus 4.8.

Ejaaz:
And that annoyed a lot of people. So people tried to kind of circumvent that barrier.

Ejaaz:
And what one person realized was there is an AI model reviewer for every single

Ejaaz:
request, and you could inject malicious code or malicious prompts just by triggering

Ejaaz:
that safeguard intentionally.

Ejaaz:
And I'm not going to walk through how this works, but the idea is they're kind

Ejaaz:
of like walk-arounds about how all of these different exploits could potentially happen.

Ejaaz:
And it ranges from like low risk and hyper localized to high risk and like accessible

Ejaaz:
to all. And like at this end is where like the highest risk kind of sits.

Ejaaz:
And that's how the government is treating this current jailbreak that they discovered.

Ejaaz:
The issue is like, I want to know what this jailbreak is. I want to know what Amazon discovered.

Ejaaz:
And I want to know that, you know, what the details of that were.

Ejaaz:
Because if we do have information on that, we'll have an idea of where on that

Ejaaz:
spectrum that I just described, this sits, and it'll inform whether the stance

Ejaaz:
that the government took was actually valid or whether it's not at all.

Josh:
Yeah, and there's some weird things going on with this, this one in particular,

Josh:
because of that history between Anthropik and the government,

Josh:
it seems like there is a very low tolerance and low empathy towards the company.

Josh:
I mean, we had, there was this funny post from Pete Hegseth,

Josh:
the Department of War, the person who's in charge of the Department of War,

Josh:
actually got community noted here, where he was making a comment that said three

Josh:
months ago, the Department of War kicked Anthropik out of the building forever.

Josh:
And then the community note is saying the official statement is not accurate

Josh:
or truthful the department of war didn't kick out anthropic forever they invited

Josh:
anthropic back when they launched the mythos model and even continue to use

Josh:
mythos and high stakes military ops so there's a lot of

Josh:
i don't want to say like psyoping but there's like a lot of social um signaling

Josh:
that's happening that is probably getting in the way of what actually happened

Josh:
at the core and i think that's something that

Josh:
i'm most interested in is is like what's actually going on behind the scenes

Josh:
that is triggering these because anthropic as an entity has said publicly that

Josh:
they that the government should have the power to um

Josh:
Kind of make governmental decisions it's just through a transparent statutory

Josh:
process like something that people can follow something that people can understand

Josh:
and this seems very opaque where there's a lot of finger pointing and there's

Josh:
a lot of name calling and community notes is saying like hey the stuff that

Josh:
you're saying actually it's not true and you're being a little like hyperbolic

Josh:
about how you address this stuff

Josh:
And it sets this kind of dangerous precedent because we are at this point in time

Josh:
in which the AI labs and the federal government are

Josh:
like at odds with each other in a way that is not productive.

Josh:
It's like the AI labs, they feel like the whiz kid child who is like in class,

Josh:
they're the genius, they know everything, they understand the ins and outs,

Josh:
and they're trying to say...

Josh:
How they plan to deploy this in a way that is safe and is good for the public

Josh:
as we move forward. And then it's the government, who is perhaps the teacher,

Josh:
who has been around for a long time, has all the authority in the room as it

Josh:
relates to our country or in the classroom.

Josh:
And they're saying, no, you don't understand. We've been here before.

Josh:
This is not how this works. I make the decisions.

Josh:
I'm going to tell you how this is done. And those two things are at odds.

Josh:
And the student is growing very powerful.

Josh:
And the government or the teacher in this case has the power.

Josh:
And they're just continuing to sort of clash in these ways that are now harming

Josh:
the public. Like I loved using Fable 5. It was an incredible model I used every single day.

Josh:
And now it's gone. And that really sucks. That seems to be the dynamic that

Josh:
we're stuck in now. It's just these two ideas at odds.

Josh:
And our producer, Luke, actually, he made this great comment earlier before

Josh:
we were recording that I thought was interesting.

Josh:
It's like, well, a lot of the people in government, they're like,

Josh:
make America great again, like looking backwards to kind of cherry pick the

Josh:
best parts of it, And then getting back to that.

Josh:
And these AI labs are kind of like, wait a second, like we actually,

Josh:
we're doing really great work and we are looking to make things great in a new way.

Josh:
And it seems like those two things have been at odds with each other is kind

Josh:
of the view that seems like it makes the most sense right now.

Ejaaz:
I think fear drives a lot of it, right? You automatically default to thinking

Ejaaz:
about the worst case scenario and then chastising whoever is creating or distributing

Ejaaz:
the tool that could potentially be dangerous, right?

Ejaaz:
So you kind of like slam your face down and you say like, no,

Ejaaz:
we're making this crazy decision to shut it down until we can figure out what

Ejaaz:
the best move is going forwards. That being said.

Ejaaz:
Whether it was Anthropic that came out with this frontier model that posed itself

Ejaaz:
as a threat, or it was OpenAI or Google or Elon Musk's ex-AI,

Ejaaz:
it would have inevitably happened because a big transition has happened over the last year where

Ejaaz:
the attack vector isn't software code anymore.

Ejaaz:
Number one, no one's writing code anymore, or at least at the top level.

Ejaaz:
They're using AIs to do it all for them, and the AI is being used to check the code themselves.

Ejaaz:
So that means that humans themselves have less of a good understanding as to

Ejaaz:
the code that is being written, less of a good understanding how the code actually

Ejaaz:
works, and then less of a good understanding how the weaknesses are surfaced and exploited itself.

Ejaaz:
So you're kind of like relying more and more increasingly on an AI model.

Ejaaz:
The second thing is the attack vector is no longer code specifically, it's words.

Ejaaz:
We spoke about the meta exploit, I think, like.

Ejaaz:
One and a half weeks ago, where someone stole $8 million worth of Instagram

Ejaaz:
accounts, just because they sweet talked Meta's AI assistant into giving them

Ejaaz:
the keys or resetting the password and sending a setting their email as the

Ejaaz:
kind of like main account owner.

Ejaaz:
And they resold those accounts for like millions of dollars,

Ejaaz:
including Obama's White House official account.

Ejaaz:
So the point being is, we are transitioning from like a world where it's like

Ejaaz:
hard coded zero in ones to a world where you can just like kind of sweet talk

Ejaaz:
an AI and kind of coax it into figuring it out. And that's not stopping.

Ejaaz:
The air models are going to keep releasing, whether publicly or privately.

Ejaaz:
And we need to figure out a way to eventually distribute this technology because

Ejaaz:
it's going to improve facets of every single industry.

Ejaaz:
So if we assume that is the case, you kind of want to see a government that

Ejaaz:
takes a more proactive approach in terms of helping figure out what this framework looks like.

Ejaaz:
And listen, in the government's defense, maybe that's what they're trying to do.

Ejaaz:
Like you said earlier, we don't know the discussions that are happening behind

Ejaaz:
the scenes. I would love to be a fly on the wall, but we don't know.

Ejaaz:
So maybe they're trying to figure out a framework that allows them to re-release

Ejaaz:
Fable 5 to everyone, or maybe they tweak a certain safeguard depending on what

Ejaaz:
that jailbreak specifically is.

Ejaaz:
But the jailbreaks aren't going away because we haven't figured out this new

Ejaaz:
wild west of how prompting works or how to defend against a nefarious attacker

Ejaaz:
that distills an account using a foreign API that gets to query Anthropic,

Ejaaz:
even though legally they're not able to do it.

Ejaaz:
Anthropic hasn't figured this out. Google, OpenAI, Anthropic have all been distillation

Ejaaz:
attacked by Chinese entities, by Chinese AI labs over the last couple of months.

Ejaaz:
So how do we stop that? We haven't figured that out just yet. And,

Ejaaz:
if we assume that these AI models are getting more powerful,

Ejaaz:
which they are increasingly at a frequent rate every couple of months now,

Ejaaz:
you need to kind of take a more proactive approach. You can't just kind of like,

Ejaaz:
the genie's out the bottle. You can't put it back here.

Josh:
And I have to ask the question of like, what happened if this was a different AI lab?

Josh:
Like Anthropic is the most safety focused company in the world.

Josh:
And their entire ethos is built around safety and security.

Josh:
What happens if this was open AI or was Gemini?

Josh:
Would it have been treated as seriously or is this just a vendetta that they

Josh:
have it's a it's really interesting to think about

Josh:
the implications if this was someone else and and it creates this kind of like

Josh:
bummer of a precedent where because anthropic went first because they tried

Josh:
their best to make it safe but they released this new frontier model

Josh:
they were the one to face all the penalties and if if open ai for example had

Josh:
released a mythos class model first

Josh:
would they have felt the backlash of it and does it create this precedent where

Josh:
maybe you should keep the models private for longer because you don't have to

Josh:
deal with the public repercussions of it.

Josh:
You don't have to deal with not only the backlash, but then the actual government

Josh:
slamming down the hammer on you.

Josh:
And who you're allowed to serve this model to and not. So does this create this

Josh:
like slowed precedent in which AI models and AI labs are kind of incentivized

Josh:
to keep their model private for longer because they don't want to be the first mover.

Josh:
They don't want to push the frontier forward because of the backlash and downstream

Josh:
implications of that. And what does that mean for the industry at large?

Josh:
Is that going to push people towards more open source models?

Josh:
Are they going to want to kind of have more control because the uncertainty

Josh:
of these public facing closed models is there like what if you build your business on

Josh:
the next version of fable or chat gpt 9.0 or something and at any given time

Josh:
it could be pulled out from underneath you it sets this like really challenging

Josh:
and difficult precedent that

Josh:
hopefully i mean like with the department of war saga earlier in the year

Josh:
we'll start to build a new framework in which companies can move forward to

Josh:
and stand on a little bit more stable ground but for now we're sitting in this limbo where

Josh:
the best model in the world is is only available for the private government

Josh:
and for private companies, and it is not available to the public,

Josh:
and that's just kind of a bummer.

Josh:
And I hope coming out of this, we start to build a better understanding of where

Josh:
everyone stands and what's allowed and what's not.

Ejaaz:
To draw a bit of a silver lining on this as we round this episode up, and this might be,

Ejaaz:
I think it's ultimately good that it is happening in the way that it's happening right now,

Ejaaz:
because typically these conversations around potential nationalization of a

Ejaaz:
government, you know, government taking equity stakes and controlling how technology

Ejaaz:
is built in their own domain.

Ejaaz:
Happens behind closed doors.

Ejaaz:
Like we don't usually have a public outwards facing type situation.

Ejaaz:
But, you know, thankfully, we have Anthropic that is not only at the frontier,

Ejaaz:
but they are openly speaking about this in blog posts, in commentary online,

Ejaaz:
and trying to keep us up to date with, you know, their different stances and

Ejaaz:
how the government is responding to that.

Ejaaz:
And as a result, the government has to engage publicly.

Ejaaz:
And that gives the public an idea of, number one, how dangerous these bottles

Ejaaz:
are. Number two, what's being restricted and prevented for them.

Ejaaz:
And then number three, how this might shape a framework of interacting with

Ejaaz:
future more powerful AI models?

Ejaaz:
I don't think the debate or question is really about Fable 5 at all at this

Ejaaz:
point. It's about, you know, the future powerful AI models that we haven't created

Ejaaz:
just yet that are currently undergoing training. You know, who gets access to that?

Ejaaz:
How is it determined? Is it by race, creed, nationality, income bracket?

Ejaaz:
Like, what does that look like?

Ejaaz:
And starting to have those discussions is now, right now, is net-net very good for us.

Ejaaz:
And I think that, if it wasn't anthropic, it'll also maybe happening with OpenAI

Ejaaz:
or Google, and I hope we start to see more of these happen between government

Ejaaz:
and AI labs, because I think the public forum of discussion is very important here.

Josh:
Yeah, we're very much witnessing history. This is another week in which we are

Josh:
going to establish new precedents for how things move forward in the world of

Josh:
AI. So it feels like every week there's something.

Josh:
There's something huge that happens that is changing, that is like creating

Josh:
history or altering it. And this is no different than that. So that's the update.

Josh:
If you're watching this, you currently can't use Fable. Hopefully that changes

Josh:
soon. Hopefully we establish some new precedents.

Josh:
But yeah, it's a really interesting story that is still ongoing and evolving.

Josh:
So as we get updates, we will continue to share them on the show.

Josh:
But I think that's the update for today. EJ, anything else that you wanted to

Josh:
add or that we missed or that's noteworthy?

Ejaaz:
No, but I do have a prompt. Now, it's funny. Typically, the prompt for every

Ejaaz:
episode is, hey, we have this new feature, this new model. Tell us what you're using it for.

Ejaaz:
I want you to tell us what you miss from Claude Fable. For those of you who

Ejaaz:
had access to it, who were using it, what do you miss the most?

Ejaaz:
What's the difference that you're noticing more now that you're using an older model 4.8?

Ejaaz:
Like, you know, where's that gap? Like kind of tell us in the comments or DM

Ejaaz:
us or whatever that might be.

Ejaaz:
Tell us, you know, what you were using Fable 5 for and why like this gap,

Ejaaz:
this restriction is kind of bad for you. Or maybe it's useful for you.

Ejaaz:
Maybe you've realized Opus 4.8 is actually a better model in many ways and that

Ejaaz:
you don't really need the Frontier model.

Ejaaz:
That might be a hot take, but let us know in the comments. But aside from that,

Ejaaz:
thank you so much for listening.

Ejaaz:
For those of you out there listening to this right now who aren't subscribed

Ejaaz:
or haven't left us a comment, please do so. We get hundreds of comments on every

Ejaaz:
single episode. We read every single one of them.

Ejaaz:
Maybe Joshua and I replied to every one of them. I don't know.

Ejaaz:
You need to find out. Maybe you need to post a reply.

Ejaaz:
And if you haven't rated us on Spotify or Apple Music or wherever you listen

Ejaaz:
to us, please do give us a five star rating. It helps us out massively.

Ejaaz:
And aside from that, we will see you on the next one.

Josh:
Awesome. See you guys tomorrow.