THIS WEEK IN AI: Chat GPT-5.5 Beats Claude Mythos, SpaceX Cursor Rumors, Google's New TPU's

Big week this week. We discuss the launch of OpenAI's ChatGPT 5.5, showcasing its exceptional coding and problem-solving capabilities that surpass competitors like Claude Mythos. Highlights include its implications for fields like mathematics, innovative applications such as a space mission simulator, and a 3D dungeon game. 

We also cover industry news on SpaceX's partnership with Cursor, Amazon’s investment in Anthropic, and emerging competition from Chinese open-source models.

------
🌌 LIMITLESS HQ ⬇️

NEWSLETTER:    https://limitlessft.substack.com/
FOLLOW ON X:   https://x.com/LimitlessFT
SPOTIFY:             https://open.spotify.com/show/5oV29YUL8AzzwXkxEXlRMQ
APPLE:                 https://podcasts.apple.com/us/podcast/limitless-podcast/id1813210890
RSS FEED:           https://limitlessft.substack.com/

------
TIMESTAMPS

0:00 Launch of ChatGPT 5.5
4:53 Comparing with Mythos
5:30 Inference and Pricing
7:27 New Applications
11:04 Impressive Capabilities
13:01 Beyond Coding
15:30 SpaceX Partnership
25:23 Anthropic's Funding
26:35 Rise of Open Source Models
28:12 Anthropic vs Figma
30:44 President's Comments on AI

------
RESOURCES

Josh: https://x.com/JoshKale

Ejaaz: https://x.com/cryptopunk7213

------
Not financial or tax advice. See our investment disclosures here:
https://www.bankless.com/disclosures⁠

Creators and Guests

Host

Ejaaz Ahamadeen

Host

Josh Kale

What is Limitless: An AI Podcast?

Exploring the frontiers of Technology and AI

Ejaaz:
The most powerful model in the world is here right now. In fact,

Ejaaz:
it's so good that it beats Claude mythos.

Ejaaz:
OpenAI just released ChatGPT 5.5 and it crushes Claude on every single benchmark.

Ejaaz:
It's the new number one coding model. It can do 20 hour tasks that expert software

Ejaaz:
engineers sometimes can't do.

Ejaaz:
It's already discovered groundbreaking solutions in maths and frontier sciences such as genetics.

Ejaaz:
And it's cheaper than GPT 5.4. This is the result of two years worth of frontier

Ejaaz:
research released in this one single model.

Ejaaz:
In fact, it's so good that an NVIDIA engineer said, and I quote,

Ejaaz:
losing access to GPT 5.5 feels like I've had a limb amputated.

Josh:
I think a lot of people are going to compare this to Opus 4.7,

Josh:
and that's fair, but I really think the true comparison is to Mythos because

Josh:
Sam Elman recently, he just posted something as the model was coming out that

Josh:
felt very much like a jab at Mythos.

Josh:
And we're going to get into the benchmarks comparing them, many of which will

Josh:
actually beat the Claude model. But what I find most interesting about this

Josh:
post is the second paragraph where he says, we believe in democratization.

Josh:
And he mentioned specifically, we have been tracking cybersecurity as a preparedness

Josh:
category for a long time and have built mitigations we believe in that enable

Josh:
us to make capable models broadly available.

Josh:
So this is very much a dig at Mythos, which is, as we all know,

Josh:
privately available, only gated to the companies that are given allowance to it.

Josh:
ChatGPT and OpenAI are like, hey, we're going to give you the powerful cybersecurity.

Josh:
We're just going to bake in the precautions into the model so that everyone could have it.

Josh:
And it ends by saying it's this really sweet thing. It's like we love you and

Josh:
we want you to win. We believe in everyone having access to this intelligence.

Josh:
And I really respect that. And I think it's an awesome way to set the precedence

Josh:
for what the next generation of these models is going to look like.

Josh:
But before we go any further, let's talk about the model itself. It's out right now.

Josh:
If you have a chat GPT membership, you can go and use it, go and play with it.

Josh:
EJS, what's the TLDR? What are the high-level things that everyone should know?

Josh:
What's most new and noteworthy about GPT 5.5?

Ejaaz:
Okay, so inspired by your mythos comparison, the first question that pops into

Ejaaz:
my head is I use Claude Opus 4.7 every single day. So I'm like,

Ejaaz:
is it better than this? Like, should I be switching back to ChatGPT right now?

Ejaaz:
The answer might be yes. So if we look at the benchmark score right here,

Ejaaz:
GPT 5.5 on the left over here absolutely crushes all the standard benchmarks

Ejaaz:
that these frontier models are weighted against.

Ejaaz:
And if you look on the right over here, Claude Opus 4.7, it either doesn't even

Ejaaz:
measure in a particular category, or it's completely beaten by GPT 5.5.

Ejaaz:
In fact, the only stat that GPT 5.5 doesn't beat Opus 4.7 in is something called

Ejaaz:
Software Engineering Benchmark Verified Pro or something like that.

Ejaaz:
It's like the pro software coding situation.

Ejaaz:
But there's a footnote at the bottom of this blog where OpenAI states,

Ejaaz:
Anthropic has publicly said that they might have gamed that particular benchmark

Ejaaz:
and they need to be re-evaluated.

Ejaaz:
So we might have a complete clean sweep for 5.5 as we see today.

Ejaaz:
So it's an incredibly powerful model.

Ejaaz:
But a question that popped to my head is, does it actually beat Mythos?

Ejaaz:
And we have a direct comparison right here.

Josh:
Yeah, so it shows that it does across some benchmarks. Now, again,

Josh:
these benchmarks are pretty fuzzy.

Josh:
We don't know which ones are gamed to do what. But there is a world in which

Josh:
GPT 5.5 will outperform Mythos on some things, which ones we're not entirely sure.

Josh:
I think as we kind of figure out ways to describe GPT 5.5, it seems as if it's

Josh:
their first attempt at making a model built for autonomy instead of answers.

Josh:
I think a lot of the benchmarks that they're working on is in agent decoding,

Josh:
things like it handles tasks that are 20 hours long. We'll get into that.

Josh:
It's doing 85% of OpenAI's internal work already.

Josh:
And it also helped rewrite the infrastructure that

Josh:
built it there was this amazing quote in the blog post it said open ai

Josh:
says 5.5 itself helped optimize the stack

Josh:
that serves it codex analyzed weeks of production traffic and

Josh:
wrote custom heuristics for load balancing that boosted token generation speed

Josh:
by over 20 so they're using the model to actually build the model and make it

Josh:
maximally efficient based on the data that it's collected from users like us

Josh:
who are interacting with the model on a daily basis so it's very smart it's

Josh:
very clever it's not just there to give you answers.

Josh:
It's there to think deeply and actually solve problems for you in a way that

Josh:
I think Mythos and a lot of these other frontier models are kind of pivoting towards now.

Ejaaz:
The great thing about this model release is it reveals a few things that OpenAI

Ejaaz:
has as an advantage against, say, a frontier lab like Anthropic.

Ejaaz:
It's clear looking at these benchmarks compared to Mythos, which,

Ejaaz:
by the way, the entire world is spiraling because of this model,

Ejaaz:
because it's going to have the cybersecurity ability to take over any kind of government system.

Ejaaz:
This model is pretty close, and Sam is going to be releasing this publicly,

Ejaaz:
or OpenAI is going to be releasing it publicly for everyone to use.

Ejaaz:
So a question that pops to my head is, does this mean that it's a matter of

Ejaaz:
compute, and OpenAI just simply has more of them?

Ejaaz:
Certainly, if you compare Sam Altman's ability to acquire compute and spend

Ejaaz:
all these trillions of dollars to acquire it versus Anthropic,

Ejaaz:
Anthropic has been extremely conservative, and now they're struggling.

Ejaaz:
They recently signed a $5 billion deal with Amazon, which we'll get to later

Ejaaz:
on. But the point is, this is a tale of two stories.

Ejaaz:
Either OpenAI has enough compute and they're about to leapfrog Claude because of that.

Ejaaz:
And they're proving that through this model that is a very good answer to Mythos.

Ejaaz:
Or, and this is the alternative side, Anthropic's Mythos model is just plainly better than 5.5.

Ejaaz:
And these benchmarks are actually verified, which is technically kind of true

Ejaaz:
because I don't know how official these things are.

Ejaaz:
These are just through tests that a small set of users have done.

Ejaaz:
So it's a game of both. I'm sure Anthropic is watching this and thinking,

Ejaaz:
hmm, maybe we should roll out Mythos, but they don't have to compute.

Josh:
Yeah, they don't have the inference. In fact, speaking of the inference,

Josh:
Sam actually made a post saying that he's really...

Josh:
Excellent work by the inference team to serve this model so efficiently he

Josh:
wanted to really highlight the fact that to a significant degree they've

Josh:
become an ai inference company now and i think that's a

Josh:
really big difference than what was previously stated like anthropic has really

Josh:
tough time serving compute and we see that and even if they had mythos available

Josh:
in a way that was safe they can't serve it open ai can and we see it reflected

Josh:
in pricing because i mean we have some pricing for this model right and it seems

Josh:
as if it's roughly at par with 4.7 if not slightly better?

Ejaaz:
Slightly. It's slightly more expensive, but not by much.

Ejaaz:
So for every million tokens input, it's both the same for Anthropic Opus 4.7 and GPT 5.5.

Ejaaz:
It's $5 in, but the output is $30 for 5.5 per million tokens and $25 per million tokens for 4.7.

Ejaaz:
So it's a little more expensive, but here's where you actually have more of

Ejaaz:
a bargain using the more expensive model 5.5. It is cheaper than GPT 5.4,

Ejaaz:
and it uses tokens way more efficiently to think.

Ejaaz:
So what does that mean if you are an enterprise that wants to plug in this AI

Ejaaz:
model and not worry about it and just have it power your entire profit engine?

Ejaaz:
Well, you end up using less tokens, so you hit your rate limits in a much slower

Ejaaz:
rate, which means that you end up getting more bang for your buck as long as

Ejaaz:
you use the model like 24-7 or you use it effectively.

Ejaaz:
Way if you are just kind of out there using 5.5 to like ask questions that you

Ejaaz:
should maybe be asking google this is probably not the model for you but otherwise super powerful one

Josh:
Yeah. And if these prices don't mean anything to you, that's fine.

Josh:
As long as you have a $20 a month subscription.

Josh:
In fact, this is going to be available to freezers fairly soon, I believe.

Josh:
But anyone who is a subscriber has access to this. You don't need to use the

Josh:
API. There's nothing fancy.

Josh:
You open up your app on your phone, you go to the web browser,

Josh:
it's there, it's available, ready to go.

Josh:
Now, there's a few interesting things that you can do with this model that haven't

Josh:
previously been possible.

Josh:
And although we don't quite have access to it just yet, we're recording this

Josh:
right as the model got launched.

Josh:
We do have a blog post from OpenAI themselves who are showcasing a few demos

Josh:
so again take these with a grain of salt these are straight from open ai but

Josh:
they are seemingly pretty impressive and pretty noteworthy as to what they're

Josh:
capable of doing starting with this space mission application which is um pretty

Josh:
cool and very reminiscent of the moon mission that we just had yeah.

Ejaaz:
Um so if you guys don't know um josh has a secret he has many secrets on this

Ejaaz:
show one is he's a massive space fan and when he's not hanging out with me he's

Ejaaz:
doing uh space simulations uh on whatever he can do right well okay maybe maybe

Ejaaz:
be part of that is a bit of a lie.

Ejaaz:
But with this new app that we're seeing in front of us right now,

Ejaaz:
this was completely vibe coded using 5.5.

Ejaaz:
And it's used to simulate a specific space mission.

Ejaaz:
Now, if this looks very similar, it's because we just had a space mission for

Ejaaz:
some we visited or went back to the moon in 53 years, pretty big deal.

Ejaaz:
And we can see a pretty accurate simulation going on right here.

Ejaaz:
So as you can see, there's various different toggles, the physics of the entire

Ejaaz:
thing is very important.

Ejaaz:
And that's another point I want to make about this model. it is being

Ejaaz:
used for frontier research not just in ai but in

Ejaaz:
mathematics in genetics like it made frontier progression on both of these fronts

Ejaaz:
and so what we're showing here is this is a model that goes way beyond just

Ejaaz:
text and telling you what could be it actually implements this into a lot of

Ejaaz:
different things and understands the world around it which is extremely powerful

Ejaaz:
but we have another one here we have a we have an earthquake tracker

Josh:
For anyone who wants to make websites, it's so good at making websites.

Josh:
And this appears to be one of the strong suits. In this case,

Josh:
there's a few things to highlight on this Earthquake tracker.

Josh:
One of them being that it's one, just like a pretty elegantly designed website.

Josh:
But two, all of the graphics are interactive. You'll notice that they update

Josh:
dynamically as you hover over them, as you click. It looks very clean.

Josh:
I assume that it is pulling up-to-date information from an API somewhere that it set up.

Josh:
It is just truly competent and capable of doing these kind of longer tail tasks

Josh:
that are a bit more complicated than a static landing page,

Josh:
but have dynamic data have the richness that

Josh:
you would expect from a high-end high-quality polished website

Josh:
except just built with an ai model from someone who doesn't

Josh:
need to know anything about coding at all and then

Josh:
for the gamers also there's another great example of a dungeon game

Josh:
which is they're describing as a playable 3d dungeon arena prototype

Josh:
built with codex and gpt models now

Josh:
i think this is something novel to this setup where codex handles

Josh:
the game architecture the combat systems the enemy encounters and then the character

Josh:
models the character textures and animations those were created with third-party

Josh:
asset generation tools using something like image gen 2.0 so this is also one

Josh:
of the earlier signs where you can actually merge a lot of these tools together

Josh:
to build something dynamic in a way that you previously couldn't have done before.

Ejaaz:
Yet actually the quality of this game looks like something out of uh league

Ejaaz:
of legends or something like that at least that's what it reminds me of like

Ejaaz:
the these games are getting way more high def than i expected i know it's just

Ejaaz:
like it's pretty basic for anyone that's watching this they can kind of like

Ejaaz:
pick with a finer eye but it's cool but for those of you who prefer like the

Ejaaz:
more traditional side of games

Ejaaz:
this might be something that you can kind of vibe code in a

Ejaaz:
couple of minutes now it may look basic but theoretically

Ejaaz:
this is like a 3D spatially aware game and that's not something you could achieve

Ejaaz:
at least very easily with previous models what I love about this as well is

Ejaaz:
it's also they've also created or included the prompt for all of these things

Ejaaz:
so this is something that you can try right now like look at this And the prompt

Ejaaz:
is no more than like, what's it?

Ejaaz:
One, two, three, four, like 12 lives. 12 lives, dude.

Ejaaz:
And you can have like a fully functioning game. You can probably then add an

Ejaaz:
extra step or extra prompt saying, hey, can you deploy this to Vercel? And-

Ejaaz:
Send that to your friends. Now you can use, you have a game.

Ejaaz:
You're a game creator. You're a game developer.

Ejaaz:
So the applications for this model cannot be understated.

Ejaaz:
I'm going to be very honest. I thought this model was going to be just an iterative

Ejaaz:
upgrade. I didn't think it would get anywhere near Claude Mythos.

Ejaaz:
Two stories have now revealed themselves, which is, one, it's the answer to Claude Mythos.

Ejaaz:
And two, it's really damn good. I am now convinced that compute is everything,

Ejaaz:
but not in the way that I thought it would be useful. I thought it would be

Ejaaz:
largely for pre-training.

Ejaaz:
But to Sam's tweet earlier on, and also in Greg Brockman, the president of OpenAI's

Ejaaz:
recent interview, they're going all in on inference, test time compute,

Ejaaz:
which just means that if you have more compute and if you have a good enough

Ejaaz:
model, it can do the thing.

Ejaaz:
This thing, like I said, built itself. It's a self-improving model. Very, very impressive.

Josh:
It's good for solving hard problems. It's good for thinking for a long time.

Josh:
In fact, they marketed it as a model that can now think for 20 hours coherently.

Josh:
Great which is almost a full day it can work

Josh:
on a problem yeah and what you're noticing from this prompt that's on screen is

Josh:
it doesn't take that much to get it going you don't need to

Josh:
kind of spoon feed it all the way through anymore it can make

Josh:
decisions on its own it can infer conclusions on what

Josh:
you want just based on the the knowledge architecture that it

Josh:
currently has it's amazingly impressive in fact one of the people who got access to

Josh:
it early just posted on x that he's posting

Josh:
live as his um prompt is seven hours

Josh:
into his task it has been running for over seven hours

Josh:
he said this has literally never happened before the models would maybe run

Josh:
for 30 minutes or so wow or or if you

Josh:
really shouted them after two to three hours but he's on seven plus hours i

Josh:
think this is going to be fun for people with complicated things if you really

Josh:
want to make a triple a feeling video game or a simulator or a really complex

Josh:
website this is the model to try out and to use it with codex and see how all

Josh:
these things kind of piece together it's really i mean

Josh:
I wasn't, I didn't have my hopes very high based on the Opus 4.7 to 4.6 incremental improvement.

Josh:
This seems like a very solid improvement over 5.4.

Ejaaz:
Absolutely. And listen, if you are listening to this and you're like, listen, I'm not a gamer.

Ejaaz:
I can't waste my time with that. I focus on more serious things.

Ejaaz:
Well, for you serious people, if you're a manager at a top company or whatever

Ejaaz:
that might be, this isn't just a toy or a model used for coders.

Ejaaz:
A lot of the examples that we just gave are around coding.

Ejaaz:
You can use this for just admin stuff or managerial work, like the capability

Ejaaz:
of this model to think more strategically and long-term and understand the context

Ejaaz:
of the tasks that you're working towards.

Ejaaz:
Like we said earlier, for coding specifically, it can work on 20-hour-long expert tasks.

Ejaaz:
That also applies for administrative stuff or things that are more generalized,

Ejaaz:
white-collar worker work.

Ejaaz:
And so in this example, Noam Brown says, I'm a manager at OpenAI,

Ejaaz:
but I'm using this model to basically manage my entire team and make sure we're

Ejaaz:
focused on the right things.

Ejaaz:
And guess what? but the output of this team and this product has been pretty amazing.

Ejaaz:
So all around really excellent work by the entire team and the inference team

Ejaaz:
specifically, as Sam Altman says here.

Ejaaz:
And yeah, I'm looking forward to using this thing. I don't have access to it right now.

Ejaaz:
I've refreshed my account probably like five times at this point and it hasn't

Ejaaz:
appeared. So maybe it's like a slow rollout.

Ejaaz:
But if you're listening to this and you've tried it out, let us know what you're

Ejaaz:
using it for. Let us know what amazes you. I really want to hear more.

Josh:
Yeah, OpenAX had a pretty incredible week. And this comes on the back of their

Josh:
new ImageGen model that they just released, which was also unbelievable.

Josh:
If you haven't seen that episode, we just recorded it yesterday.

Josh:
So I would go advise you to see because, oh my God, it is amazing.

Josh:
We also recorded an episode on Apple's new CEO this week and what that means

Josh:
for the company, as well as the hardware race and how this, I mean,

Josh:
this model, Opus, no, not Opus, this is GPT.

Josh:
GPT 5.5 is very much part of the AGI class of models that is built on Blackwell

Josh:
chips. and we've recorded an entire episode all about that.

Josh:
Very interesting, very fascinating. Also interesting and fascinating because

Josh:
as always, this is the weekly roundup. We have a few other topics to talk about.

Josh:
We have some news out of SpaceX, which is a pseudo acquisition.

Josh:
Now they haven't quite acquired Cursor being the company in question,

Josh:
but they have at least partnered with them with the option to buy Cursor for

Josh:
either $60 billion or pay 10 billion for the right to actually work together.

Josh:
This seems like a big deal.

Josh:
This seems like, I mean, XAI, we could call it SpaceX, but SpaceX AI is taking

Josh:
AI very seriously. They're currently behind.

Josh:
They clearly don't want to be behind. This is a huge step and a huge kind of

Josh:
trust of support in Cursor with this minimum of $10 billion into accelerating

Josh:
their progress and trying to get themselves into this game.

Ejaaz:
This is actually a genius deal, and there are a few stories why it makes that so. So let me explain.

Ejaaz:
If you're SpaceX AI, which by the way is a ridiculous name now,

Ejaaz:
like we'll just call them XAI, you are currently harboring...

Ejaaz:
One to 1.5 million of the frontier GPUs, mainly NVIDIA, in a warehouse. There's one issue.

Ejaaz:
You're not really utilizing all of it because XAI has had a bit of a slow start

Ejaaz:
to training their models. What's a genius idea?

Ejaaz:
If I rent those out to another company to train their own model,

Ejaaz:
then we can make money from that. Okay, so that's win number one for SpaceX.

Ejaaz:
But then they've thought of another thing which is huh grok

Ejaaz:
isn't really good at coding and we are

Ejaaz:
losing the race every single day we don't update our model

Ejaaz:
at coding because anthropic and chat gpt

Ejaaz:
5.5 is completely running away with it so

Ejaaz:
how did they leapfrog and get ahead they should acquire the company that is

Ejaaz:
using their own gpus to train a frontier coding model so then the question becomes

Ejaaz:
well who the hell is cursor what what's the mode that they have like why do

Ejaaz:
they have a good shot of training a better coding model than Anthropic and GPT-505?

Ejaaz:
Aren't those two companies way ahead? Well, the answer is not quite so.

Ejaaz:
Cursor, for the longest time, was the number one platform and tool for people

Ejaaz:
to use to do their Vibe coding. Why?

Ejaaz:
Not only did they have access to Frontier coding models from Claude and ChatGPT,

Ejaaz:
they also had something called an agent harness.

Ejaaz:
Now, you'll notice in GPT-505, it's really good at coding because of something

Ejaaz:
called agentic That is something that Cursor pretty much pioneered.

Ejaaz:
It's basically the harness, the prompts, the environment that they mold the model,

Ejaaz:
or rather that they mold around the model that makes it so good and intuitive

Ejaaz:
and remembers the context across every single project, like menial things,

Ejaaz:
like understanding your GitHub branches and working on separate flows at the same time.

Ejaaz:
A lot of the top software engineers in the world right now use tools like Curse

Ejaaz:
and Argentic Coding to be able to pull this off.

Ejaaz:
So Elon Musk thought, hmm, if I give you the GPUs to train a better coding model,

Ejaaz:
which gives you a better product, I should have the option to acquire you.

Ejaaz:
In acquiring you, I can integrate you with Grok and Grok somehow becomes the

Ejaaz:
number one coding model over the next year or so, depending on if this deal goes.

Ejaaz:
And if the deal falls through and they create a really bad model,

Ejaaz:
well, you pay me $10 billion for the service.

Ejaaz:
Well i pay you not a bad deal not a bad deal

Josh:
Yeah it seems like they're they're going to be continuing to

Josh:
work with other companies to accelerate in places that

Josh:
they're weak at currently because i mean they they're so strong at

Josh:
building out the hardware and creating these huge data centers they need

Josh:
someone who could take advantage of all those gpus hopefully this will

Josh:
help serve that cause and that's not the only spacex news

Josh:
this week the other is that they have officially filed an s1 which

Josh:
for those who are not familiar it means they're going public it's officially official

Josh:
100 they will be going public this year if there

Josh:
were any doubts please let them be relinquished here we

Josh:
have it spacex will be going published the most interesting thing from

Josh:
this was i think the share structure of

Josh:
how they're going to be organizing this for daddy elon who's going to be getting

Josh:
quite a big payday if he does well so we have on screen here just a series of

Josh:
some of the financials i mean we know starlink as a business has been doing

Josh:
unbelievable they have about 25 billion dollars in cash 92 billion assets 50

Josh:
billion liabilities that's.

Ejaaz:
Quite a lot of liabilities on this my god

Josh:
They got a lot of debt man i don't know we'll see we'll see once they finally

Josh:
publish everything i'm very excited for the first earnings report where you

Josh:
really get a true peek behind the scenes of what's going on there but it looks

Josh:
like it's going to be going public at a 1.75 trillion dollar valuation now in terms of pay structure.

Josh:
Elon is posed to get 60 million shares

Josh:
which is 11 tranches vesting in

Josh:
500 billion dollar market cap increments from 1.1 trillion to 6.6 trillion dollar

Josh:
share price oh um so for those unfamiliar with the current ceiling i think it's

Josh:
nvidia nvidia is what five trillion under five trillion close to five trillion

Josh:
it's like 4.3 yeah okay so not even close they're like 20 away from five trillion.

Josh:
SpaceX needs to be, what is that? Like 20 something percent more valuable than

Josh:
the most valuable company in the world.

Josh:
But if they do, Elon gets 60 million shares.

Josh:
Now I haven't done the math on exactly how much that is.

Josh:
But if we make some assumptions here, the total value at Vest looks like it

Josh:
could be about a quarter of a trillion dollars.

Josh:
So pretty good payday for Elon. I think the most important thing is that he's

Josh:
getting a lot of control over this.

Josh:
It seems as if he's going to have 40 something percent control of

Josh:
the company which is really ultimately what was most important to

Josh:
him as they went public so really exciting news i am hopeful that it happens

Josh:
this june which we can expect and it's without a shadow of a doubt going to

Josh:
be the largest ipo in history i think everyone's going to be talking about it

Josh:
there is a new vehicle in which some people are investing in we're actually

Josh:
going to have the founder on the show soon so keep an eye out for that one.

Josh:
And yeah, the SpaceX news is very exciting.

Ejaaz:
Now, in the world of AI hardware, many people think that NVIDIA has run away with the win.

Ejaaz:
And you could argue that with a $4.300 market cap, not many people are competing,

Ejaaz:
except that there is one company, Google.

Ejaaz:
Now, you might be thinking, Google does all my search engines and stuff.

Ejaaz:
Well, Google is the only vertically integrated Mag 7 company that is involved

Ejaaz:
or has a frontier capability at every single layer of the AI stack.

Ejaaz:
Now, right at the bottom are these things called Google TPUs,

Ejaaz:
Tensor Processing Units.

Ejaaz:
And they're their version of the GPU. In fact, fun fact,

Ejaaz:
Google's Gemini models has never trained on an NVIDIA GPU.

Ejaaz:
It's all been their own internal warehoused infrastructure. And they've been

Ejaaz:
working on this thing for 10 years.

Ejaaz:
Now, just today, or rather this week, they released their latest generation

Ejaaz:
of TPUs, the TPU-8T and the TPU-8i.

Ejaaz:
Now, the TPU-8T, T stands for training or pre-training.

Ejaaz:
It is highly optimized for the pre-training part of an AI model.

Ejaaz:
So this is like the bulk, arguably the more expensive part of training a model.

Ejaaz:
It's like teaching it like, hey, these are words.

Ejaaz:
These are the general fundamental set of facts that you need to know before

Ejaaz:
we can kind of like put you out into the world and present you to our users.

Ejaaz:
TPU AI is specialized or hyper-specialized in inference specifically.

Ejaaz:
Now, the important part about inference is it's being used for so many different things.

Ejaaz:
Number one, it's to answer all your different prompts. Whenever you write a

Ejaaz:
prompt and you submit it to an AI model, it is known as inference.

Ejaaz:
It's getting inference. It needs to query the model and make sure it does the

Ejaaz:
right types of thinking and gives you the right answer.

Ejaaz:
But the other part of inference is post-training, where a lot of people train

Ejaaz:
the model, and then they do more training after the fact by using it to help

Ejaaz:
the model reason and think of other alternative facts before it presents you the actual answer.

Ejaaz:
And that's what that second TPU is. Now, Google's TPUs have been used extensively.

Ejaaz:
In fact, their largest customer is a little-known AI lab known as Anthropic,

Ejaaz:
which currently runs 1.5 million TPUs. So the argument can be made that TPUs

Ejaaz:
are largely responsible for Claude's and Opus' success.

Ejaaz:
So very impressive all around, but there's some other facts about this, right?

Josh:
Yeah, well, I love the dual architecture training setup that they have here being hyper specific.

Josh:
I mean, the AT chip in particular, it's built to reduce frontier model development

Josh:
cycles, they said, from months to weeks.

Josh:
And then we have the AI, which is the reasoning engine, which is specifically

Josh:
served for agentic use to deliver tokens really quick, as fast as possible.

Josh:
And as we know, Anthropic is working closely with them. And also,

Josh:
I mean, Google is making these for themselves.

Josh:
So I think whoever is working with Google, whoever's kind of focused on these

Josh:
accelerators, is probably in for a nice little windfall as it relates to increased

Josh:
velocity of the training and also increased ability to distribute these models.

Josh:
As we know, Anthropik is having a very difficult time with this.

Josh:
Now, NVIDIA and Jensen are probably feeling a little shook.

Josh:
They got to be feeling a little bit of pressure here. and it seems as if

Josh:
that's why they're pushing to be open source because if you

Josh:
are a in a closed source world where everyone is

Josh:
making closed source models on their own architecture then the

Josh:
nvidia edge very quickly disappears and i mean i'm looking at these chips in

Josh:
hand they look beautiful they're ready to be they're taped out ready to be manufactured

Josh:
and i i think you could start getting kind of excited about this new world of

Josh:
accelerated hardware and we're seeing this happen again and again because amazon

Josh:
just made another big investment in who else other than Anthropic.

Josh:
And the deal, I think, is like this has to be close to a record deal.

Josh:
They're owning a tremendous amount of this company now.

Ejaaz:
Yep. So the news here is Amazon announced they're investing $5 billion into Anthropic.

Ejaaz:
They've just raised $5 billion. Congrats.

Ejaaz:
And so the reason why this is important is, well, there's a few reasons.

Ejaaz:
Number one, Anthropic knows that they don't have enough compute.

Ejaaz:
The argument could be made that's why Claude Mythos hasn't been rolled out.

Ejaaz:
Well, hey, hey, presto, now you have $5 billion worth more of compute.

Ejaaz:
Now, for those of you who didn't know, Amazon is a primary investor already in Anthropic.

Ejaaz:
Before this announcement, they owned around 17% of Anthropic.

Ejaaz:
After this announcement, it's closer to 20%. So we're talking about one company

Ejaaz:
that's publicly tradable right now that owns a fifth.

Ejaaz:
Is my math right? Yeah, a fifth of the world's leading

Ejaaz:
AI lab, which is pretty crazy. Now, if we look into the stats of this,

Ejaaz:
this is a five gigawatt deal, which is more than any single data center that's currently live.

Ejaaz:
It's actually a multiple of five. I think SpaceX AI's Colossus 2 is the largest

Ejaaz:
right now with their 1 million TB.

Ejaaz:
So it's going to be 5X larger than the average data center that we're seeing

Ejaaz:
right now for AI specifically.

Ejaaz:
And they're aiming to get one gigawatt online by the end of the year.

Ejaaz:
Now, Now, the reason why this is so good for both teams is Anthropic already

Ejaaz:
has a close relationship with AWS and Amazon's cloud computing department.

Ejaaz:
So spinning up more compute clusters is gonna be so easy for them.

Ejaaz:
They have a working relationship.

Ejaaz:
They're used to training cloud models on this, so it shouldn't be too hard to

Ejaaz:
ramp this up. If you're Amazon,

Ejaaz:
Welcome back. That $5 billion is going to come right back to you.

Ejaaz:
So I don't know what kind of like circle economy this is, but it's back and

Ejaaz:
it's very impressive for them.

Josh:
Is it ironic that today Amazon hit an all-time high? No, maybe, maybe not.

Ejaaz:
I'm holding stock. I got the stock.

Josh:
Clearly, clearly they're doing something right. Amazon is a phenomenal company.

Josh:
They're the largest shareholder in Anthropic.

Josh:
It's hard not to be bullish on them. It's hard not to be bullish on the accelerated

Josh:
computing stack. And I think that's probably what Jensen is getting nervous

Josh:
about. That's why NVIDIA is pushing open source.

Josh:
And the good news is, is he has some help. He has some assistance.

Josh:
From the folks overseas in China who have been pumping out unbelievable models

Josh:
all week long as it relates to Kimi and Quen, our Chinese favorites.

Josh:
We have Kimi K2.6 and Quen 3.6. There's a lot of digits and numbers.

Josh:
All you need to know is that the best open source models in the world didn't exist last week.

Josh:
They now exist this week and they are better at pretty much everything,

Josh:
but exceptional at coding.

Josh:
In fact, word on the street is that some of these models are as good as GPT

Josh:
5.4 was and only a few points off of Claude.

Josh:
I mean, these are pretty amazing open source models that, again,

Josh:
are free to run locally on your machine if you have the machine capability of doing so.

Josh:
That's a big, that's a big game changer.

Ejaaz:
Okay, so typically the story we tell with these open source models is,

Ejaaz:
wow, aren't they so amazing?

Ejaaz:
Yeah, they're the good younger brother. They're not as good as the Frontier AI Labs.

Ejaaz:
That completely changed this week. So Kimi K 2.6 is the latest model from a

Ejaaz:
Chinese lab called Moonshot Labs. I believe it's Moonshot or Moonshot AI.

Ejaaz:
And they released their model, which ends up being as good as coding or at coding.

Ejaaz:
As opus 4.7 and it's 100 open source

Ejaaz:
like you mentioned josh which means that maybe you could run

Ejaaz:
this on a local device now the answer that you would typically get back from

Ejaaz:
this is hey like listen it's uh it's too

Ejaaz:
large to run on my laptop and that is true but with the latest quen model which

Ejaaz:
is a 3.6 version you can run it as an 18 gigabyte sized model slightly quantized

Ejaaz:
on your laptop today so the point that i want to make about these models isn't

Ejaaz:
exactly the specifics but across all benchmarks,

Ejaaz:
it's not as good as the Frontier AI Labs, but it's a few points.

Ejaaz:
That difference and gap has closed massively over the last couple of months,

Ejaaz:
which tells me two things.

Ejaaz:
Number one, China has figured out some kind of groundbreaking way to train their

Ejaaz:
models that they haven't told the West about, and they're going to keep it closed

Ejaaz:
guard and eventually close source their model releases going forwards.

Ejaaz:
And number two, they've figured out a new way to use inference to their benefit.

Ejaaz:
Like one thing I'm going to highlight here is this new Kimi K 2.6 model can

Ejaaz:
code continuously for 12 hours straight using 300 agents.

Ejaaz:
So the unlock here isn't one model itself. It's spitting up 300 versions of

Ejaaz:
itself and getting it to attack the problem.

Ejaaz:
That's something Sam realized and what he's implementing in 5.5.

Ejaaz:
That's something Opus 4.7 realized and is doing probably similarly with Mythos.

Ejaaz:
So I have this question here, which is like, how do they have to try to do this?

Ejaaz:
Well, I think every three months that there's a new open model that gets released,

Ejaaz:
they're making these jumps because they're using these models to train themselves

Ejaaz:
we proved that with kimmy k 2.5 there's too many two point whatevers um and

Ejaaz:
the same thing is happening with quen it's just all around pretty amazing stuff um

Josh:
Yeah chen is crushing okay so before we go we have two quick things to hit the

Josh:
first being one that we missed last week which we need to touch on quickly anthropic

Josh:
has a design tool now if you are a designer if you are interested in building

Josh:
web pages videos graphics slideshows pitch decks any type of visual asset you're

Josh:
claude now has an entire design suite built just

Josh:
for this purpose it's called claude design it exists separately

Josh:
you can access it through the desktop app or on your browser and

Josh:
it basically allows you to build visual assets in a way that

Josh:
you couldn't previously previously with claude you had artifacts an

Josh:
artifact you could generate something dynamic it could kind of

Josh:
build you a web page this takes it to a whole new level you could generate wireframes

Josh:
if you want to try it to use less tokens you could fill it out and create properly

Josh:
created prototypes that are actually clickable it's amazing the video we're

Josh:
seeing on screen highlights a few of them unfortunately there was a big loser

Josh:
in this because uh this sounds like a lot of what that little design company named figma does.

Ejaaz:
Yeah the little company

Josh:
Stock market did not love the reaction to that did it nope.

Ejaaz:
Nope it is down almost 20 on the week um i actually tracked the stock price

Ejaaz:
after the announcement was made so like it wasn't even readily available it

Ejaaz:
was literally just the tweet 20 minutes after it was tweeted the stock was down

Ejaaz:
six percent so So the point being,

Ejaaz:
whether this is market speculation or not, like, listen,

Ejaaz:
Claude Design isn't as good as Figma.

Ejaaz:
They're working with a few of these different partners, such as Canva.

Ejaaz:
Two weeks ago, one of Anthropic's former most execs left the board of Figma.

Ejaaz:
And the rumors was because they were building a competitor. So it's pretty clear.

Ejaaz:
Anthropic is going off to every single sector, whether you're a designer,

Ejaaz:
a software engineer, a mathematician, a research scientist, doesn't matter.

Ejaaz:
They're going off to everything because the model is applicable to everything.

Ejaaz:
And I don't know what this means for certain modes that companies like Figma

Ejaaz:
holds, but it's certainly going to affect the stock price.

Josh:
Can you do me a favor and click the max button real quick for me just to show the chart?

Ejaaz:
Oh!

Josh:
Yeah, minus 86% since IPO for those who are not watching on screen.

Josh:
It's been a pretty bad, rough run for Figma.

Ejaaz:
We have to start naming Anthropic the stock killer, Josh. This is like every

Ejaaz:
single tweet is tanking a stock.

Josh:
No, it's tough. It's brutal. We had one last thing that you wanted to mention.

Josh:
I know. We got to end on this strong. What do we have?

Ejaaz:
How good is your accent or impersonation of your president, of our president, Josh?

Josh:
Pretty horrible.

Ejaaz:
Not good. Okay, well, then we're not going to attempt it.

Josh:
I'd love to hear your British take on it, though, if you're feeling ambitious.

Ejaaz:
Okay, so my British take on this is, this is, albeit hilarious and somewhat

Ejaaz:
terrifying, that the President of the United States is saying this.

Ejaaz:
He commented, okay, on the government's relationship with Anthropic.

Ejaaz:
Now, if you're wondering why on earth he's commenting on it,

Ejaaz:
they're going to be releasing this cold mythos model.

Ejaaz:
It might be a security risk. It's probably good for the government to have access

Ejaaz:
to this thing and prepare necessarily.

Ejaaz:
The government has been having very important conversations with bankers and

Ejaaz:
governments all around the world, just try and figure out, you know,

Ejaaz:
how best to prepare for this.

Ejaaz:
And after having an in-depth discussion with Dario Amore, which by the way,

Ejaaz:
he blacklisted that CEO and Anthropic entirely from the government using it.

Ejaaz:
He's now rekindling it and saying, maybe there's a deal on the line.

Ejaaz:
He goes, and I quote, I'm not going to do the accent.

Ejaaz:
We'll get along with Anthropic just fine. Trump said on CNBC.

Josh:
We'll get along with Anthropic just fine. I think they can be of great use to us.

Josh:
They're high IQ people. Very good. Very good. They tend to be on the left,

Josh:
radical left, but We get along with them.

Josh:
I don't know. That's all I got. But that is what he said.

Ejaaz:
Were you practicing that? That was actually pretty good. I was practicing my head.

Josh:
I was rehearsing.

Ejaaz:
Damn. I closed my eyes whilst you were doing that, whilst I was laughing. Did it feel right?

Josh:
It sounded like him? Good.

Ejaaz:
It channeled his spirit. It was there. It was a good effort.

Ejaaz:
But I believe that's it. That is the end of the roundup.

Ejaaz:
Josh and I are recording this. FYI, it's 4 p.m. over here. Typically, we're morning birds.

Ejaaz:
We deliver this in the morning, but we waited for the announcement of Spud GPT

Ejaaz:
5.5 just for you guys. and we're going to be bringing you the cutting edge news every single week.

Ejaaz:
As Josh mentioned, we had three other amazing episodes that we filmed earlier

Ejaaz:
this week. Definitely go check them out. They're all each 20 minute song.

Ejaaz:
It's your commute to work.

Ejaaz:
It's your gym session if you're not that active.

Ejaaz:
Definitely go check it out and let us know what you think. But yeah,

Ejaaz:
Josh, any final thoughts?

Josh:
Call me crazy, but I like the afternoon recordings. I got good energy.

Josh:
I'm like woken up. I'm 100% right now. I'm rocking and rolling.

Josh:
I'm feeling good. So I don't know. Maybe we'll have to lean into this a little

Josh:
bit more, but that's everything.

Josh:
If you've made it this far, if you're still listening to this and you've heard

Josh:
our other episodes, you're caught up. You're done for the week.

Josh:
You can go touch grass. Enjoy your weekend.

Josh:
There will be a lot more to talk about next weekend. But for now,

Josh:
you have fully synchronized with all of the chaos happening on the frontier of AI and technology.

Josh:
Thank you so much for watching. As always, we very much appreciate it.

Josh:
If you enjoyed this episode or any of our previous episodes from this week,

Josh:
don't forget to share them with a friend who you also might enjoy it, possibly.

Josh:
We have a newsletter on Substack that goes live twice a week.

Josh:
Just went live yesterday, going live again tomorrow.

Josh:
The Friday issue is a recap of everything that happens this week,

Josh:
which is always fun and exciting.

Josh:
In fact, I'm gonna go write that as soon as we finish this episode.

Josh:
So thank you all for watching.

Josh:
As always, don't forget to subscribe, like, comment, all the good things,

Josh:
and we will see you guys next week.