"Kimi K2 Thinking" is China's Plan To End American AI Dominance

In this episode, we discuss the launch of Kimi K2 Thinking from Moonshot AI Labs, an open-source AI model targeting GPT-5 and Gemini for just $4.6 million. With its impressive benchmarks, there are major implications for the American AI industry amidst rising competition. Tune in for insights on Kimi K2’s innovative architecture and its potential to reshape the future of AI and its economy!

------
🌌 LIMITLESS HQ: LISTEN & FOLLOW HERE ⬇️
https://limitless.bankless.com/
https://x.com/LimitlessFT

Substack:
https://limitlessft.substack.com/

------
TIMESTAMPS

0:00 Kimi K2: The New Frontier AI
1:29 Impressive Specs and Performance
4:54 Cost Comparison with GPT-5
6:59 Mixture of Experts Architecture
8:47 U.S. vs. Chinese AI Models
11:21 Open Source Advantages
13:02 Licensing and Commercial Use
18:37 User Experience and Ecosystems
21:05 Efficiency vs. Precision
22:53 The Consumer Advantage
24:03 Future of Open vs. Closed Source
25:32 Closing Thoughts and Call to Action

------
RESOURCES

Josh: https://x.com/joshjkale

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 world's latest and greatest AI model is 100% free for you to download and run at home right now.

Ejaaz:
Kimi K2 Thinking is the latest reasoning model from Moonshot AI Labs,

Ejaaz:
which is a Chinese frontier AI lab.

Ejaaz:
And it beats OpenAI's GPT-5, Anthropics Claude, and Google's Gemini across pretty much all benchmarks.

Ejaaz:
But that's not even the most shocking part. The most shocking part is that it

Ejaaz:
only costs $4.6 million to train and build, which is only a fraction of the

Ejaaz:
billions of dollars spent by OpenAI to train GPT in the first place.

Ejaaz:
It's also 100% open source, which means that you can download and run Frontier

Ejaaz:
AI right at home where you're sitting right now.

Ejaaz:
But of course, it begs two very important questions. Number one,

Ejaaz:
is open source AI the winning strategy?

Ejaaz:
We've been led to believe that closed source is typically the better strategy

Ejaaz:
when you run a business, but China and the AI models are proving us wrong here.

Ejaaz:
And the second question, the more ominous question is, will the US stock market bubble finally pop?

Ejaaz:
Josh, what have we got here? What is this new model and why is it taking over

Ejaaz:
social media everywhere I look?

Josh:
They did it again. The Chinese did it again. They knocked it out of the park.

Josh:
Grand slam, home run. It's an unbelievably impressive model.

Josh:
And this happens every time.

Josh:
We get this amazing flagship model out of the US.

Josh:
A couple months later, we get the same thing, marginally better,

Josh:
at one-tenth of the cost. Like full orders of magnitude cost less than what

Josh:
it calls for the leading AI labs in the US today.

Josh:
The specs are really impressive. We're going to get into everything.

Josh:
We'll start with, I guess, just like the high level spec sheet.

Josh:
State of the art on humanity's last exam, which is the reference point that

Josh:
we kind of use in terms of benchmarks.

Josh:
It scored the highest anyone's ever scored, 44.9%.

Josh:
It has a bunch of these really cool breakthroughs, but the big thing that it

Josh:
excels at, like it says in the post here, reasoning, agentic search, and coding.

Josh:
Now, there's a few cool things that we could talk about here.

Josh:
Ejaz, maybe we'll just get into the charts because i feel like that's an easy way to visualize

Josh:
how much better this model really is than all of the others and what we're seeing

Josh:
on the chart is that while gpt5 was the best kimmy k2 is now the new best and

Josh:
this is as it relates to thinking and reasoning

Josh:
and this again this is so impressive because

Josh:
one this model is fully open source you can go download the model and run it

Josh:
yourself locally for free what were your first thoughts when you saw this because

Josh:
to me i was like oh my god i was like why would i use anything else?

Ejaaz:
My first thought, if I'm being honest, Josh, was like to look at the stock market.

Ejaaz:
I was like, is this going to crash the entire US stock market?

Ejaaz:
Like when DeepSeek initially released the R1 thinking model,

Ejaaz:
do you remember it was at the end of last year, people's kind of entire bubble

Ejaaz:
and vision of how AI models were trained was completely burst.

Ejaaz:
And since then, China has repeatedly delivered on breaking edge models,

Ejaaz:
one of which is the Moonshot AI lab team, which built Kimi K2.

Ejaaz:
It's such an impressive model for a few different reasons for me.

Ejaaz:
Number one, it can now compete with all the best. And,

Ejaaz:
personally, GPT-5 is something that I use pretty much every day,

Ejaaz:
whether it's for like kind of casual prompts and requests or whether it's kind

Ejaaz:
of like the deeper thinking and research and some of the lines of work that

Ejaaz:
I do. So it's become kind of like...

Ejaaz:
Quintessential for me. Now to have a separate model that I can download and

Ejaaz:
run privately on my own computer at home that I'm showing on this tweet here that costs 60 cents,

Ejaaz:
per million token input and $2.5 output is just an insane cost-cutting average

Ejaaz:
where if I was running a business using an AI model.

Ejaaz:
There's very little reason for me not to switch over to something like this

Ejaaz:
aside from maybe like maintenance and setup and stuff like that.

Ejaaz:
The other really impressive thing for me, Josh, was the team itself.

Ejaaz:
Like, this is only a two-year-old startup, which reminds me of another two-year-old

Ejaaz:
startup, which is Elon Musk's XAI, right?

Ejaaz:
And there's a funny link between these two models, Josh, which is Kimi K2's reasoning,

Ejaaz:
this thinking model, can do so because it does this like really neat little

Ejaaz:
chain of thought experiment where it takes many steps to kind of think to a

Ejaaz:
logical answer versus just kind of like splurging an answer for you.

Ejaaz:
That's something that Grok Heavy 4 did that they pioneered when they launched their new product.

Ejaaz:
So Kimi K2 is kind of like drawn on some of these learnings from XAI to produce a similar model.

Ejaaz:
The other really cool thing is

Ejaaz:
it does this thing called tool use or tool calling whilst it's thinking.

Ejaaz:
So if you imagine, as I'm kind of like trying to think through a complex problem,

Ejaaz:
I will leverage different tools to be able to help me get to the answer.

Ejaaz:
So if I'm doing a maths exam, I can use a calculator. Or if I'm doing a deep

Ejaaz:
research question, I might use Google.

Ejaaz:
This AI model naturally does that and has access to over 200 to 300 different

Ejaaz:
tool calls and tool uses whilst it does its thinking. So just overall,

Ejaaz:
a very impressively new looking AI model.

Josh:
Yeah, Ijaz, you mentioned the cost being 60 cents per million tokens.

Josh:
And I just want to add a little bit of context as to how low that actually is.

Josh:
I was looking at the GPT-5 Pro cost per inputs, and it is $15 per million tokens.

Josh:
$15 for the GPT-5 Pro cost currently.

Josh:
The output is $120 per million tokens.

Josh:
Granted, this is the top of the top. If you're using GPT-5 standard,

Josh:
input is $1.25 per million tokens, output is $10.

Josh:
So any way you scrape it, it's at least a 2x multi-cost reduction up to like

Josh:
100x on the highest end, assuming it can compete with GPT-5 Pro,

Josh:
which all those benchmarks suggest it very well can.

Josh:
So the cost is really like, it's a big deal. And to get kind of dig more into

Josh:
the point that you were making, Ejoz, and how it actually works.

Ejaaz:
Well, we'll get there.

Ejaaz:
Save the memes.

Josh:
Don't spoil the memes yet. We got to get to the funny jokes next.

Josh:
But basically, the way this works is like there's this very complicated diagram

Josh:
on the screen. I'm not going to try to even explain what that is.

Josh:
But there's this fun way that I like to describe it when I was describing it

Josh:
to my friend earlier this morning.

Josh:
Which is that like Kimi K2, it's like this giant school and it has these things called specialists.

Josh:
And in fact, Kimi K2 has 384 specialists. You could think of these specialists

Josh:
as like a math club or a history club, coding club, debate, whatever it is.

Josh:
And when you ask it a question, it doesn't invite the whole school.

Josh:
It doesn't invite all the clubs. It's just, EJS, if you ask for a math question,

Josh:
it will query the math club.

Josh:
And it chooses eight out of those 384 clubs to help combine their answers,

Josh:
pick the experts and decide how it's going to solve this problem.

Josh:
So it has a trillion parameters, but it only uses 32 billion of them at once.

Josh:
And that's how we're able to get the huge cost reduction because it uses this

Josh:
thing called mixture of experts. A lot of people describe it as MOE,

Josh:
but basically what it is, instead of using the entire model's intelligence to

Josh:
answer, what should I have for breakfast this morning?

Josh:
It will take the chef club. It will take the health club.

Josh:
It will combine those together and it will form an answer that should hopefully

Josh:
give you just as good as a result if you took the entire model but it's much more efficient

Josh:
in terms of cost in terms of energy and in terms of the amount of tokens they

Josh:
could generate because it's so much cheaper across the board and i think that's

Josh:
one of the big really exciting things that has been cool to see coming out of china

Josh:
we saw with deep seek we see it with kimmy and it's this mixture of agents architecture

Josh:
where they're really kind of modularizing the entire model

Josh:
and only using the stuff that's important for the specific query

Ejaaz:
They were put in a very constrained position which is they didn't have access

Ejaaz:
to the latest GPUs or NVIDIA GPUs, there's been a bunch of U.S.

Ejaaz:
Tariff restrictions on Chinese labs getting access to these kinds of things.

Ejaaz:
So they've really needed to kind of like work within their bounds and means.

Ejaaz:
And so coming up with an architecture like Mixture of Experts or the one that

Ejaaz:
they did is super important.

Ejaaz:
And it brings me to this meme, Josh, which is, what are we doing here?

Ejaaz:
There is an obvious mismatch between American-made AI models and the,

Ejaaz:
Chinese ones. You've got OpenAI, which is now projected to spend $1.4 trillion,

Ejaaz:
over the next five years.

Ejaaz:
That's trillion with a T versus Kimi training for $4.6 million.

Ejaaz:
Now, I know there's a bit of like clickbaitiness here.

Ejaaz:
That $4.6 million was relative to one training run and usually takes a few training runs.

Ejaaz:
But let's say it took like 20 training runs, right?

Ejaaz:
At $4.6 million, that's still only like 100 mil, right? Or less than that.

Ejaaz:
So it doesn't really matter when you put it into the context that GPT-5 is rumored

Ejaaz:
to have cost $1.7 to $2.4 billion for OpenAI to train.

Ejaaz:
So there's a mismatch that I don't quite understand, Josh. And that's what makes

Ejaaz:
me the most nervous when it comes to what American-made companies and Frontier Labs are doing.

Ejaaz:
I feel like they're missing the mark. I don't quite know what it is,

Ejaaz:
whether it's this mixture of experts thing, but there's someone's being sold

Ejaaz:
a lie and I don't know whether it's me or whether it's me like looking at this

Ejaaz:
Kimi K2 model and being like wow it's so amazing.

Josh:
Yeah, when I think about the role that China plays versus the United States

Josh:
in terms of like open source companies or closed source companies here in the US,

Josh:
the thing that is reassuring to me, at least, is a lot of these innovative breakthroughs

Josh:
that happen on the software level actually do happen in these private AI labs.

Josh:
We do get like chain of thought and reasoning. And there's like this whole slew

Josh:
of new innovation that becomes standard very quickly.

Josh:
That all happens in the United States AI labs. And as far as we're concerned,

Josh:
the AI labs in the US still have they're making the most progress the fastest

Josh:
they are creating the most innovation and then what you kind of see like we

Josh:
described earlier in the episode

Josh:
is that innovation starts to trickle down whether it's voluntary or whether

Josh:
it's stolen and it gets implemented into these new models and they just completely

Josh:
cut out the bottom in terms of cost and efficiency because that's kind of all

Josh:
they're able to do they don't have access to the resources of

Josh:
like millions of gpus from jensen huang and nvidia they don't have the access

Josh:
to 50 billion dollars of capex just to spend on

Josh:
employees just to spend on salaries and compensation um so it seems to me like

Josh:
i mean we're still doing very well it's just

Josh:
China is very good at implementing the technology and applying it at scale in

Josh:
a way that's open sourced and the open source thing there's there's a lot to say for that because

Josh:
it's it's very impressive and it's kind of this community effort that we saw

Josh:
early days with the united states but once they became better they closed it off so what happens is

Josh:
you get innovation in one company like Kimi and then you see it implemented

Josh:
in DeepSeek and then you see it implemented in Quen and then suddenly this technology

Josh:
is it's kind of synchronously growing between the three because it's all open

Josh:
source they're publishing all the code all the open weights

Josh:
and it's much more easier for them to thrive whereas innovation in the United

Josh:
States very much happens behind a closed wall and it's only leaked out

Josh:
at the advent of a new model when they release it to the world and people kind

Josh:
of reverse engineer how it works

Ejaaz:
Um, I was reading an article in the Financial Times where they interviewed Jensen

Ejaaz:
Huang, and he said verbatim that China will win the AI race if they continue

Ejaaz:
down the pot that they're currently on.

Ejaaz:
And if the US doesn't kind of ramp up their energy production,

Ejaaz:
he was making a wider point that their open source strategy is pretty effective

Ejaaz:
in the way that they're up, that they're building these new AI models with the

Ejaaz:
constraints that you just mentioned.

Ejaaz:
And kind of speaking more about the open sourceness and the benefits of this.

Ejaaz:
I've got a tweet up here, which shows that Kimi K2 thinking this new model can

Ejaaz:
basically run on two MacBook M3 Ultras, which is like a couple of thousand dollars

Ejaaz:
worth of cost, which is an insane thing to do to run Frontier

Ejaaz:
AI model at home, privately in your house, trained and fine tuned on any of your own private data.

Ejaaz:
So you don't need to kind of like sell that data to Sam or Mono,

Ejaaz:
whoever, just super cool and super cheap, right? Because you're running local

Ejaaz:
inference at home. So you don't have to worry about anyone kind of like spying

Ejaaz:
on any of your queries or your prompts or your research.

Ejaaz:
It's just all at home, which I thought was super cool.

Ejaaz:
The other part of the open sourceness, which I found interesting,

Ejaaz:
Josh, was the fact that they had an MIT license with this new release or an adjusted MIT license.

Ejaaz:
And we'll dig into that in a second. But the point being, when DeepSeek released

Ejaaz:
their first major open source model and it took the world by storm,

Ejaaz:
there wasn't any major licenses around that.

Ejaaz:
So you could pretty much download and do whatever the hell you wanted for it.

Ejaaz:
You could implement it into your own product, whether you were an American founder.

Ejaaz:
And if, let's say, you scaled that up to a million users that used a feature

Ejaaz:
that was leveraging that DeepSeek model, you wouldn't have to credit that team at all.

Ejaaz:
Kimi K2 kind of like takes a step in a different direction here where they've

Ejaaz:
released an MIT license where I think if you hit

Ejaaz:
I think it's either 10 million or 20 million users for your product you need

Ejaaz:
to show the Kimi K2 label and say that listen I'm using this model under the hood but there's

Ejaaz:
there's some differences with this license right Josh can we can we dig into that?

Josh:
I believe it's modified. I don't know to the extent that it is modified,

Josh:
but I know that there is something different going on here. What does this say?

Josh:
Our only modification part is that if the software or any derivative works thereof

Josh:
is used for any of your commercial products or services that have more than

Josh:
100 million monthly active users or more than 20 million US dollars or equivalent

Josh:
other currencies in monthly revenue,

Josh:
you shall prominently display KimiK2 on the user interface of such product or service

Josh:
that's a fun little marketing ploy

Ejaaz:
Fair enough fair enough you know what it reminds me of josh um what's that it's

Ejaaz:
what meta tried to do with their llama models right,

Ejaaz:
so um meta is the only other major american company that i can think of that

Ejaaz:
went down this open source ai route and the goal or the intended goal at the

Ejaaz:
time was to basically level the playing field,

Ejaaz:
between Meta and OpenAI and other frontier model AI labs, which had raced so far ahead.

Ejaaz:
So if you released all this cutting edge AI tech for free and accessible to

Ejaaz:
anyone, then it kind of drives down the cost premium that OpenAI and all these

Ejaaz:
other frontier AI labs can charge you to access this thing.

Ejaaz:
China's doing that as a vast hole on the American AI stock market,

Ejaaz:
right? So that's why we saw like NVIDIA crash, I think 4.2% on the news getting released and such.

Ejaaz:
I'm curious whether this kind of pops the bubble and the CapEx bubble in America, Josh.

Ejaaz:
Is that a crazy thing to say? I mean, the markets reacted pretty viscerally to this news.

Josh:
I don't think I have a problem with this. I don't think it's popping a bubble.

Josh:
I don't think we're in trouble.

Josh:
I think this is just totally fine so long as we continue to stay slightly ahead or at least at par

Josh:
i think we're really excellent at making software distributing software creating

Josh:
products i think china's really good at shamelessly

Josh:
innovating and deploying without needing to go through all the hoops and intellectual

Josh:
Problems that the united states mostly has um so i don't think this will lead

Josh:
to any sort of bubble popping i think a lot of the frontier innovative stuff

Josh:
still happens in the u.s the place where i will begin to start to get a little

Josh:
worried is when this switches to embodied ai once we

Josh:
start moving from large language models to implementing these into robots or

Josh:
implementing these into physical hardware that's where i think we have problems

Josh:
on the software front we're good we're crushing it everyone's spending tons

Josh:
of money um on the hardware front we don't have the same lead and

Josh:
over the last what 30 to 50 years we've kind of outsourced manufacturing capabilities

Josh:
to other places and therefore are just kind of i mean everyone knows we just

Josh:
can't really make things cost effectively here in the united states

Josh:
if we are at a foot race with china when it comes to making embodied ai like

Josh:
humanoid robots specialized robots whatever it may be that's where things start

Josh:
to get a little bit scary because that's where there is a significant lead and

Josh:
that lead is comes in the form of atoms which are much more difficult to move than bits

Josh:
because you can steal some open source code create this slight innovation on

Josh:
top roll it out to a billion users overnight and that's innovation that does not happen

Josh:
between version two and version three of your humanoid robot you actually have

Josh:
to build it with a factory with real materials and people and places and it's

Josh:
it's very difficult and challenging to do and china very much stands to be the

Josh:
largest winner in that so i think on the software front i feel really confident

Josh:
and as of now that's all that we're battling on but in this near future where

Josh:
things start to become embodied, where AI becomes physically manifested in the world around us,

Josh:
That seems like a place where I would start looking at Chinese investments a

Josh:
little bit more than the American ones.

Ejaaz:
Okay. I think I might push back a little bit and say that there is reasonable

Ejaaz:
evidence to be bearish on the software side before it gets embodied AI.

Ejaaz:
I mean, so a few ways to think about it.

Ejaaz:
There is such a gross discrepancy when it comes to capital expenditure for these things.

Ejaaz:
On one side, you've got the US spending trillions of dollars literally to train

Ejaaz:
AGI or the best AI models.

Ejaaz:
And on this side, you're in like the hundreds of millions of dollars,

Ejaaz:
which is like an order of magnitude less, right?

Ejaaz:
So there's an obvious mismatch here that we aren't seeing.

Ejaaz:
Whether it comes down to training architecture, training design,

Ejaaz:
or just kind of like hardware manufacturing.

Ejaaz:
I don't know where that kind of advantage is being played, but the Chinese have

Ejaaz:
found it and they're able to kind of really push down on that lever to get ahead

Ejaaz:
or on par with the U.S. And they've been able to successfully do this for years

Ejaaz:
now at this point. DeepSeq was kind of like test case one.

Ejaaz:
Now I've seen like, you know, at least 50 open source models come out of Chinese

Ejaaz:
frontier AI labs since then.

Ejaaz:
Number two, it's not like the U.S. government has kind of like not tried to constrain them.

Ejaaz:
We've imposed a number of different sanctions, which include,

Ejaaz:
you know, constraining which GPUs NVIDIA and other manufacturers within the

Ejaaz:
U.S. can sell to China. But that still hasn't stopped them.

Ejaaz:
They've been able to maintain and train these frontier AI intelligences despite

Ejaaz:
all of these different things.

Ejaaz:
So I think if I were to look on the other side of this, it would be,

Ejaaz:
so what if you have an open source model that is super cool?

Ejaaz:
Why aren't you using it right now? Like I'm not using Kimi K2 regularly,

Ejaaz:
even though I use GPT-5 and it might be better than GPT-5. And the answer for me is pretty simple.

Ejaaz:
I'm locked into an ecosystem in OpenAI that I'm pretty happy with,

Ejaaz:
which is it has memory on me. It understands who I am.

Ejaaz:
It has a context of all the previous chats that I have with it.

Ejaaz:
But also most importantly, Josh, if there's an issue with something on my account

Ejaaz:
or something that I'm trying to use, there's a community that I can access.

Ejaaz:
There's a support team that I can speak to. There's a software ecosystem that supports me, right?

Ejaaz:
Versus me jumping ship to kind of Kimi K2 setting it up on my own and then having

Ejaaz:
to like troubleshoot it myself I think a lot of people will be disincentivized to do that.

Josh:
It is difficult, but I mean, we're seeing market forces from both sides, right?

Josh:
Like I saw you included a link here somewhere where Cursor and Windsurf's new

Josh:
AI models, they were using some sort of Chinese models.

Josh:
In fact, they were thinking in Chinese. And I found this really fascinating

Josh:
that like American-made products are now thinking in the Chinese language.

Josh:
So that's certainly a concern in terms of the commercial side where those API

Josh:
costs really matter, where if you can get a million tokens for $0.60 versus

Josh:
$10, that really affects the margins of your business.

Josh:
For consumers like us, there's no real interest to use Kimi K2.

Josh:
And the phenomenon you spoke about earlier, where you can actually run a quantized

Josh:
version of Kimi K2 on two Mac studios running the M3 Ultra chips,

Josh:
it generates tokens at like 13 to 15 tokens per second. So it's very slow.

Josh:
Slow. You're getting like a sentence or two every second, which it's much slower.

Josh:
It's going to feel groggy. It's not going to feel well.

Josh:
There's a case to be made that that changes because this year and it's funny

Josh:
that apple's really the only computer that that supports this now they're releasing

Josh:
the m5 ultra which will be the new version

Josh:
and um it's going to be interesting to see how it plays out what i found interesting

Josh:
this one side note actually that i wanted to share with you just because you

Josh:
might find it cool too is the version that runs on these apple computers the

Josh:
apple studios um it's a it's a slightly quantized version and

Josh:
I heard about this and i learned about this recently in the tesla earnings call

Josh:
that they had the shareholder meeting recently and we're gonna have an episode

Josh:
on this later this week but there's this interesting thing that elon mentioned

Josh:
during the episode where he was talking about quantize versus floating point

Josh:
ai and i was like what the hell is that like what why are you spending so much

Josh:
time talking about this it doesn't make sense

Josh:
and what i realized is a lot of ai models they they use like

Josh:
many, many points after the decimal in terms of data to get more precise results.

Josh:
And that is floating point. When you quantize a model, you remove all of the

Josh:
data to the right of the model and you just go to single integers.

Josh:
So you lose the variance of maybe up to like 60%, but you gain so much faster

Josh:
efficiency, so much better speed improvements, cost improvements,

Josh:
and you can actually run it locally on these things.

Josh:
So I think it's interesting to see the different decisions that people are making

Josh:
in terms of, well, how precise does the model have to be versus how cost effective

Josh:
and how efficient does it need to be?

Josh:
And what we're seeing with Kimi Gay too is it's very easy to over-index on the

Josh:
efficiency, but maybe that's not the stated goal of OpenAI, where if they really

Josh:
wanted to, they could sign up, quantize these models. They could go more to integer type compute.

Josh:
And it was just something I was thinking about is how they approach them,

Josh:
because it could just be, well, Kimi's just kind of optimizing for speed and

Josh:
efficiency, and the downstream effect is it's also really fast,

Josh:
whereas OpenAI kind of hasn't really optimized for that specifically yet.

Ejaaz:
Right. And the counter argument to that point would be, well,

Ejaaz:
Josh, it's crushing all the benchmarks that we've evaluated all the other American

Ejaaz:
models on, right? Yeah. So surely it's much better.

Ejaaz:
And my pushback on that would be like, well, benchmarks don't really materialize in real life use.

Ejaaz:
So what if it crushes 50% on humanities last exam?

Ejaaz:
Is it useful for me to use? Does it understand what I'm trying to say?

Ejaaz:
Does it understand the context of the prompts that I'm putting into it?

Ejaaz:
Um the other side of this um you know on the point of quantization josh is um,

Ejaaz:
i think that a lot of frontier american ai labs like open ai google etc,

Ejaaz:
actually have enough compute to give you the best experience the um the highest

Ejaaz:
floating point uh experience um to put it to put into that context,

Ejaaz:
but they're using the majority of that compute to train the next big model that

Ejaaz:
we haven't even seen yet, right?

Ejaaz:
There was news that broke last week that OpenAI is doing this,

Ejaaz:
right? So technically they have enough compute to give you like amazing service

Ejaaz:
all year round, but they're using 70% of that compute to train GPT-6.

Ejaaz:
So I think it's just a matter of prioritization right now until we reach some

Ejaaz:
kind of parity that these AI models are good enough.

Ejaaz:
But I will say from all of the things that we've discussed on this episode so

Ejaaz:
far, there is one clear winner and And that is the consumer.

Ejaaz:
It's you, I, and everyone listening to the show, which basically gets access

Ejaaz:
to frontier level intelligence for the cost of next to nothing.

Ejaaz:
Download it completely free and run it privately at home.

Ejaaz:
On this tweet that I have pulled up here, it basically says for every closed

Ejaaz:
model, there is an open source alternative in it.

Ejaaz:
And it goes through a list like Sonnet 4.5, you've got GLM 4.6,

Ejaaz:
Grok Code Fast, you've got GPT OSS, GPT 5, you've got Kimi K2 Thinking.

Ejaaz:
And it just goes on and on and on. And if we look at this kind of like a year

Ejaaz:
and a half ago, maybe even two years ago, this list would be non-existent.

Ejaaz:
It would just be Frontier AI Labs on the closed source side and zero open source side.

Ejaaz:
So to see this kind of progress is really, really encouraging.

Josh:
Yeah, it's going to be a race. It's going to be a battle between open and closed source.

Josh:
And perhaps that's not even the battle. Perhaps it's open source until they

Josh:
catch up to closed source.

Josh:
And then it's closed source across the board. So it's going to be interesting

Josh:
to see the developments.

Josh:
We have a new batch of models that are coming. We're kind of in this weird limbo

Josh:
where Gemini 3 is hopefully coming soon. We'll have some new benchmarks.

Josh:
And one of the things that was this harsh truth to kind of wrap my head around,

Josh:
which is what you just mentioned, Ejaz, and the fact that everyone's just compute constraint.

Josh:
Like OpenAI could have made GPT-5, probably twice as impressive if they really

Josh:
wanted to they just have no compute to serve that and it would have been way

Josh:
too expensive and way too slow so it's not that it's

Josh:
they can't it can't be done it's just that people don't have the resources to

Josh:
do it so it's this constant balancing act and it's gonna be fun to see how

Josh:
how companies kind of slot themselves into that that curve of like how much

Josh:
they want to spend on compute versus cost versus just what they have available

Josh:
to actually use to train these models and deploy them at scale to users

Ejaaz:
And that's it for today folks um super fun episode.

Ejaaz:
It is always surprising to me how quickly open source catches up with closed

Ejaaz:
source centralized AI. I always think kind of like it's going to lag a few years

Ejaaz:
and now it's come down to the fact that it's lagging a few weeks.

Ejaaz:
We have a jam-packed week. We have potentially a new nano banana model being

Ejaaz:
released by Google tomorrow. Fingers crossed.

Josh:
I'm praying for that.

Ejaaz:
Fingers crossed. I'm also praying for that as well. And we have a second episode

Ejaaz:
based on tesla's investor day which had some really jam-packed exciting news now listen,

Ejaaz:
If you want the U.S. to win this AI race, and make no mistake,

Ejaaz:
it is a race, you need to subscribe to American AI YouTube channels, one of which is us.

Ejaaz:
Please subscribe, hit the notification button, wherever you're listening to, give us a rating.

Ejaaz:
We are helped by these so much. It is bringing up so much awareness.

Ejaaz:
The algorithm is favoring us. We're getting all these wonderful views and new incomers.

Ejaaz:
We've got a thousand of you from last week, which is just insane.

Ejaaz:
Hello, welcome to the channel.

Ejaaz:
We hope you enjoyed the content and we will see you on the next one.

Josh:
Yeah before i let them off the hook i'm checking i'm doing the stat update 83

Josh:
percent of the people that watched last week were not subscribed if you're watching

Josh:
this on youtube don't come on guys

Josh:
or go on spotify my preferred place of finding this podcast it's the best i'm telling you

Josh:
i don't know how to describe this to people any better spotify is so good you

Josh:
have the video you have the audio you could turn it off and lock your phone

Josh:
without needing a premium membership please go go over there go leave a comment

Josh:
over there because also the comment section is kind of popping too so yeah anyway

Josh:
thank you for all this we do

Ejaaz:
Not pick and choose wherever you listen go for it.

Josh:
There you go all right we will see you guys in the next one thank you for watching

Josh:
as always much appreciated um peace