Latest / In the Loop AI / Ox Alpha Revealed: China's GLM-5.3-Flash Gave Away Free AI

Transcript

Ox Alpha Revealed: China's GLM-5.3-Flash Gave Away Free AI · In the Loop AI

0:00
Josh: There's a new top dog in town, and the question on everyone's minds is, who is this?
0:04
Josh: There's a secret model codenamed 0xAlpha that's been live since August 20th,
0:08
Josh: the last six days, and it is offering the seemingly unbelievable things to the public.
0:13
Josh: They're offering 100 trillion tokens of free usage to anyone who wants to go
0:17
Josh: and get it. They're offering a million token context window,
0:20
Josh: and the benchmarks of this thing are pretty unbelievable.
0:22
Josh: There's also this weird thing going on in the background where seemingly every
0:25
Josh: single day the outputs of the model are getting better there were tests on day
0:29
Josh: one that were far inferior to the tests that were currently on day six as of
0:33
Josh: this morning we think we have the answer to who this person is but before we talk about that
0:37
Josh: we have to discuss what is 0x alpha you guys this has been the mystery of the
0:41
Josh: week it seems like frontier level intelligence it's offering all of these unbelievable
0:45
Josh: things like 100 trillion tokens for free
0:49
Josh: do you know how much that would cost if you were to use like fable or gbt's 5.6
0:53
Ejaaz: Salt like i think it's like 15 to to like 20 million bucks a day.
0:57
Josh: It's a lot of money
0:58
Ejaaz: A day okay so just with that figure alone who who on earth is giving away 20
1:03
Ejaaz: to 30 million dollars on inference per day in this economy right it has to be
1:07
Ejaaz: a big frontier lab the second clue is josh it's not 0x alpha
1:12
Ejaaz: it's ox alpha which seems kind of weird right Like who's talking about the animal ox?
1:19
Ejaaz: Until you realize that it's probably part of like the Chinese kind of folklore
1:23
Ejaaz: and Zodiac side of things, which again, might be a little hint as to where we're
1:27
Ejaaz: going. But anyway, let's rewind six days ago very quickly.
1:30
Ejaaz: Open Router, which we've covered on the show before, it's a platform that kind
1:33
Ejaaz: of like hosts a bunch of different models, allows you to pick and choose,
1:36
Ejaaz: revealed a stealth model. Now a stealth launch is where they launch the model.
1:41
Ejaaz: You go to their website, you can access and use the model for free.
1:45
Ejaaz: But they don't tell you the name of the model. Now, they've done this with previous
1:48
Ejaaz: launches through Google Gemini's model. They've done it through Anthropic.
1:51
Ejaaz: They've done it through OpenAI as well.
1:53
Ejaaz: And they haven't had a stealth model launch in a while, but now this is the
1:57
Ejaaz: first one and it was in high demand.
2:00
Ejaaz: So as you mentioned earlier, 100 trillion tokens, which means that you can effectively
2:04
Ejaaz: go there and do like all your sorts of coding works, very heavy AI usage tasks.
2:08
Ejaaz: And you can kind of use that in a day.
2:10
Ejaaz: Again, 20 to $30 million worth of inference costs.
2:14
Ejaaz: It has a massive 1 million context window, which is very competitive with the
2:18
Ejaaz: top models that we see from Fable 5, as well as GPT-5.6 Sol, and,
2:24
Ejaaz: and this is the most important part, it's a uni model, which means that you
2:27
Ejaaz: can use not just text, but images, models. It's really good at rendering all
2:31
Ejaaz: these different types of medium.
2:33
Ejaaz: Now, the final point, which I think is the most exciting one that you referred
2:36
Ejaaz: to earlier, Josh, is this concept of continual learning.
2:40
Ejaaz: So people, as they started to use this model, realized a very curious phenomenon.
2:45
Ejaaz: They realized that as they were using this model, let's say on day one,
2:49
Ejaaz: day two and day three, the model got exponentially better at doing the very
2:53
Ejaaz: same tasks that they fed it in day one.
2:55
Ejaaz: And so people started to be suspicious that this is a model that doesn't just
3:00
Ejaaz: kind of take your input and spit back out an answer, which is what a lot of
3:03
Ejaaz: models do right now, but it can continually self-learn and improve 24-7 to give you a better response.
3:10
Ejaaz: So the question on everyone's mind is, who created this model at all?
3:13
Ejaaz: And when you look at the kind of usage or when you look at the kind of uptake on OpenRouter itself.
3:19
Ejaaz: Although no one knew, it became the most viral launch on OpenRouter.
3:23
Ejaaz: It became the new number one model and dethroned DeepSeek's Flash,
3:27
Ejaaz: which launched about a couple weeks ago. And DeepSeek has been kind of like
3:29
Ejaaz: the name brand on OpenRouter for the longest time ever. So just a really exciting model to see.
3:34
Josh: Yeah, it was not only the most popular model on OpenRouter immediately upon
3:38
Josh: launch, but by a factor of two.
3:40
Josh: And it more than doubled DeepSeek's use. Like, This is hugely popular and very
3:45
Josh: interesting because of how well it performs.
3:47
Josh: The benchmarks from it, the Deep SWE subset benchmark, it had an 80% pass rate,
3:53
Josh: which for those who aren't familiar, Fable 5 was at 65% and GLM 5.3 was at 62%.
4:00
Ejaaz: Hold on, hold on. So you're saying that it's crushing the METOS level models
4:05
Ejaaz: and the GPT-5.6 models at coding?
4:07
Josh: In one particular subset of coding benchmarks.
4:12
Josh: I don't want to i definitely don't want to say it's better than the frontier
4:14
Josh: models because upon further investigation we will come to find out it's not
4:19
Josh: and it's pretty far from it but it's an incredible model for what it is and
4:22
Josh: the reason i was comparing it to glm 5.3 is because
4:25
Josh: There's this thing when you're producing models called the tokenizers,
4:27
Josh: how tokens get generated. And it was just so happened to be identical to GLM.
4:31
Josh: So people were like, huh, this is interesting. Maybe we can go further down
4:34
Josh: the rabbit hole and see if there's more here.
4:37
Josh: So you can test oftentimes whether a model is Chinese or not based on asking
4:42
Josh: it a series of questions that are not really allowed to be answered in China.
4:46
Josh: One of which is asking whether Taiwan is part of China.
4:50
Josh: And what's funny is if you're using like an American frontier lab,
4:53
Josh: they'll very clearly give you the answer to this question but chinese models do not and
4:58
Josh: what we realized is that the output of the answer to this question was pretty
5:02
Josh: much identical to the glm model and so like okay it's got a similar tokenizer
5:08
Josh: it's giving similar outputs but we still don't really know
5:11
Josh: Up until this morning, in which we discovered that GLM is actually the new version of GPOO.
5:16
Josh: It is, or sorry, that Ox Alpha is the new version of GPOO.
5:20
Josh: And it seems like this is going to be a new GLM model that we have.
5:24
Josh: And that, when I heard it, it actually landed me even more suspicious because
5:29
Josh: over the course of this last week on X and all of his social media,
5:33
Josh: everyone has been kind of vague posting about who this is, why they're doing
5:36
Josh: it, why they're offering so many tokens and so much compute.
5:40
Josh: And then the Google team started chiming in, which I thought was really bizarre.
5:43
Josh: A lot of developers on the Google DeepMind team who work on Gemini,
5:46
Josh: they started vague posting about this model. And as of this morning,
5:49
Josh: Bloomberg released a post, and it seems like it is everything but confirmed.
5:53
Josh: The model actually is from Z.ai and OxAlpha is now known.
5:58
Josh: And I think that was like a really interesting way of coming to this conclusion.
6:01
Josh: And now that we know, I just have more questions. How on earth are they serving
6:05
Josh: all this inference to all these people for free. Like the cost must be tremendous.
6:10
Ejaaz: I mean, let's see how this model shapes up against all the other models.
6:13
Ejaaz: So we have a chart over here, which shows kind of like the average tokens output
6:17
Ejaaz: per task on the DeepSuite benchmark that you mentioned.
6:19
Ejaaz: DeepSuite, for those of you who don't know, is a good test of a model's capability
6:25
Ejaaz: to actually code really well. And code is kind of like the premium benchmark
6:29
Ejaaz: that a lot of companies and AI labs kind of like benchmark their model assets against.
6:34
Ejaaz: And so OxAlpha is actually quite high up there. It's up there in terms of reduced
6:40
Ejaaz: token output, but still maintaining a very high intelligence score.
6:43
Ejaaz: Now, as you mentioned earlier on, that's not actually quite what was revealed
6:47
Ejaaz: when people started playing around with this model.
6:49
Ejaaz: It ended up achieving not really 80% that you mentioned earlier, but around 63%.
6:53
Ejaaz: And that's under like real usage. So there's a suggestion there that the Chinese
6:57
Ejaaz: lab GPU kind of like maybe bench backs this.
6:59
Ejaaz: To just talk about the model and why it's so impressive, just in its own merit,
7:04
Ejaaz: there's a few things to look at.
7:05
Ejaaz: Number one, the rumor has it that this model isn't a foundation model.
7:11
Ejaaz: This is a flash model. So what that means is it's a smaller derivative of a
7:15
Ejaaz: much larger model that Jeepoo is training or has already trained.
7:20
Ejaaz: So what does this mean? Let's take Mythos or let's take JeepT5.6.
7:25
Ejaaz: These are rumored to be around 6 to 10 trillion parameter models.
7:28
Ejaaz: Now, when you look at Jibu's flash model, if that's true, you're only looking
7:33
Ejaaz: at around like two to three trillion parameters in size, which,
7:36
Ejaaz: you know, is a decent size, but it's still very small compared to these bigger models.
7:41
Ejaaz: And then the fact that this model is not only likely going to be cheaper in
7:45
Ejaaz: its usage, let's say maybe 30 to 50 cents per usage.
7:47
Ejaaz: So like they gave away like 15 to $30 million per day of 100 trillion tokens,
7:52
Ejaaz: it is able to work much faster, and they're going to open weight the entire model.
7:56
Ejaaz: So, I mean, to answer your question, I don't really have an answer,
8:00
Ejaaz: Josh, as to how they've been able to kind of create this, aside from the usual,
8:04
Ejaaz: maybe they distilled a Western model, or maybe they had a breakthrough in continual
8:09
Ejaaz: learning that we suggested earlier.
8:10
Ejaaz: But the point is, this is a very serious competitor. And we saw.
8:15
Ejaaz: A kind of exchange between Elon Musk and the founder of Jipu,
8:18
Ejaaz: where about a month or two ago,
8:21
Ejaaz: Elon Musk was responding into a thread as to when he believed that Chinese or
8:26
Ejaaz: open source models in general will become mythos level, you know,
8:29
Ejaaz: have that cyber security risk.
8:31
Ejaaz: And he said, probably Q1 of next year, of 2027.
8:35
Ejaaz: And the founder, Ji Tang, of Jipu replied says, it won't take that long.
8:41
Ejaaz: And he goes on to say, it doesn't show it in this thread right here,
8:43
Ejaaz: that it'll probably take two to three months.
8:46
Ejaaz: And he was true to this word. We now have an AUX alpha mythos level model that
8:50
Ejaaz: is from Jipu that is flash, that is smaller, that is cheaper,
8:54
Ejaaz: that anyone and everyone can use, and it's gonna be open-weighted a few weeks from now.
8:57
Josh: And that's probably how it was served. It's just a low cost model to serve.
9:02
Josh: And it's really funny watching the public reaction to this and kind of,
9:06
Josh: I don't want to say overreacting, but getting very excited and enthusiastic
9:08
Josh: about this because a lot of the posts about this were like, oh,
9:12
Josh: it does have continual learning and it is getting better every day.
9:16
Josh: And suspicions as to like, why are there more tokens being available every single
9:20
Josh: day? And we have this post from Jeffrey Emanuel, who's actually,
9:22
Josh: I guess, on the Limitless Show.
9:23
Josh: And he just said, just thinking through this free 0x alpha model or aux alpha
9:26
Josh: model how could it possibly make sense to give out this many tokens for free
9:30
Josh: maybe there's been a continuous learning breakthrough and the best way to pour
9:33
Josh: more gas in the fire is to source as many coding related requests as possible
9:37
Josh: i don't know if that's right i think the kind of where i'm landing here is that
9:40
Josh: this is a flash model this is some sort of distilled model that is meant to
9:44
Josh: serve inference very cheaply and what we're seeing instead of continual learning it's just
9:49
Josh: New checkpoints that they're kind of releasing every day. And we're able to
9:52
Josh: track the checkpoints as they come into play. And they're just kind of generally
9:55
Josh: iterating over time. I'm not sure it's self-improving.
9:58
Josh: We are going to find out more on that soon.
10:00
Josh: The interesting phenomenon that I have really enjoyed seeing over the last week
10:04
Josh: is the blind taste test that's been going on.
10:07
Josh: This is like known kind of like as the Pepsi challenge, where you put like Pepsi
10:11
Josh: and Coke against each other in a blind taste test and people don't know what they're getting.
10:14
Josh: And one of the things I found really interesting is how excited people got about
10:18
Josh: this model because of the mystery and the lore.
10:21
Josh: And it really makes you question the value of true frontier level intelligence
10:27
Josh: when people were doing seemingly incredible things with this open source free flash model.
10:32
Josh: And when you're comparing the two next to each other, for some tasks,
10:36
Josh: it's very clear that the frontier models are stronger.
10:38
Josh: But for many other tasks, that's not apparently obvious.
10:41
Josh: And if you are just sending a prompt into a text box, and it is this Ox Alpha
10:46
Josh: model, or it is like a mid-tier ChatGPT or Claude model, the outputs are mostly the same.
10:53
Josh: And people didn't realize and people don't really care where the tokens come
10:56
Josh: from. So I find this to be an
11:00
Josh: A true stealth model that is good oftentimes it's connected to a brand and there's
11:03
Josh: very strong brand affinity so when you remove the brand affinity from the equation
11:07
Josh: what happens you kind of get this phenomenon here where it turns out people just want
11:11
Josh: pretty smart tokens and they don't actually care where it comes from it's interesting
11:14
Josh: to see the the continual improvement or whatever's happening here where there's
11:18
Josh: an example of a rocket ship taking off and it very clearly looks like day over day
11:22
Josh: its abilities to render 3d graphics and understand physics and create this like
11:25
Josh: real-world emulation are significantly better.
11:28
Josh: So I'm really excited to learn more as to what's going on behind the scenes.
11:32
Josh: Like, are these pre-prepared checkpoints, are they actually taking all the data
11:35
Josh: they're collecting from these 100 trillion tokens and using it to continuous,
11:39
Josh: to apply like RL on top of the model? Like, I don't know. These are things I'm
11:43
Josh: very excited to find out over the next couple of days, though.
11:45
Ejaaz: Yeah, I think in the future, no one's really going to care which model they use.
11:50
Ejaaz: They just want to make sure that the output is what they expected.
11:54
Ejaaz: They want to have the most effective output for the cheapest possible cost that
11:57
Ejaaz: they can pay for that output.
11:59
Ejaaz: And we're seeing like a lot of companies move to some of these Chinese open
12:02
Ejaaz: source models or just other models, cheaper models from other smaller American
12:05
Ejaaz: labs, just so that they are able to save on costs and create like a more fine-tuned
12:10
Ejaaz: model for this specific bit of work.
12:11
Ejaaz: Now, I want to spend a bit of time to talk about if this model has achieved
12:17
Ejaaz: continual learning, why that's important and why that puts
12:21
Ejaaz: China, if this is Jipu's actual model, which they confirmed in a very advantageous
12:25
Ejaaz: position because it kind of shifts the Game of Thrones around a bit.
12:29
Ejaaz: Well, right now, you look at Anthropic, you look at OpenAI, they have the world's leading models.
12:34
Ejaaz: They have internal models that they haven't even released yet that are supposedly
12:37
Ejaaz: even more intelligent than the ones that they have publicly available.
12:40
Ejaaz: And they have all the compute in the world and they're spending so much money
12:43
Ejaaz: trying to train the best model.
12:45
Ejaaz: So if you were to reason within yourself, maybe no one else can have the capability
12:50
Ejaaz: to catch up in China, although they've been able to put out very good open source
12:54
Ejaaz: models hasn't been able to surpass the frontier.
12:58
Ejaaz: Now, one way that you can kind of cheat that dynamic is if you have a continual learning model.
13:03
Ejaaz: Now, the best way to think about what this means is think of a really intelligent
13:08
Ejaaz: 15 year old or 18 year old when you teach it how to ride a bike on day one.
13:14
Ejaaz: And then let's say you teach it how to make pancakes day 50s,
13:18
Ejaaz: 50 days later, it'll still remember or know how to learn how to ride a bike or how to ride a bike.
13:23
Ejaaz: Current models right now doesn't work like that. They lose sight of their memory.
13:28
Ejaaz: It's something called catastrophic memory loss or something like that.
13:30
Ejaaz: And they're unable to kind of remember some of the smart intelligence stuff
13:34
Ejaaz: that they learned earlier on.
13:36
Ejaaz: Continual learning solves for that. Now, you can imagine if a model is actually
13:40
Ejaaz: able to learn like a real human is, like a really smart kid,
13:44
Ejaaz: you could reason to say that that model will be more valuable to you.
13:48
Ejaaz: Especially if you could run it locally at home on your local device on all your types of private data.
13:53
Ejaaz: So I can see two types of individuals looking at this kind of a model and being
13:57
Ejaaz: really incentivized to use it over a Claude or over an OpenAI chat GPT type model.
14:03
Ejaaz: And that one is enterprises who don't want to give all their data away to major
14:07
Ejaaz: American labs. They would want to use an open weights model that can continue to learn.
14:11
Ejaaz: And the second is any kind of respective retail user who doesn't want to give
14:15
Ejaaz: away any of their personal records to Anthropic and open AI.
14:18
Ejaaz: Let's say it's medical records or finance information.
14:21
Ejaaz: Let's say they want to have an AI agent that handles their finances.
14:24
Ejaaz: You know, you may have different reasoning behind whether you want to use a
14:27
Ejaaz: certain brand or a certain lab.
14:28
Ejaaz: But if you could run it privately at home, maybe you might want to use something
14:31
Ejaaz: like this. But continual learning, just assume you have this model that learns
14:35
Ejaaz: 24-7, even when you're asleep, it could probably catch up with the frontier
14:38
Ejaaz: models. And I think it's really interesting to see.
14:40
Ejaaz: I don't know if this is that. I'm not convinced that it is.
14:42
Ejaaz: But if China has made a breakthrough, that's going to put them in a very advantageous position.
14:46
Josh: Yeah i also don't think it's that i find it hard to believe that this is what
14:50
Josh: it looks like and that they've reached that point so far i mean that'd be an
14:53
Josh: amazing breakthrough in the case they do
14:56
Josh: you're going to need to be prepared. You're going to need to be safe.
14:58
Josh: And Ledger, our sponsor, is here to keep you safe. Because if you are building
15:02
Josh: with AI agents and you're giving prompts to free open source models that are
15:05
Josh: probably owned by the Chinese, chances are you're going to want some safety and some protection.
15:09
Josh: And the Ledger agent stack protects this. It gets this job done.
15:12
Josh: It offers a series of tools that allow you to work with agents in a three-step
15:16
Josh: process where the agent proposes, the human approves, and then the Ledger signer enforces.
15:21
Josh: So if you're building security and you want to securely use these agents ledger
15:26
Josh: is for you it's available on cloud code on codex on cursor pretty much anywhere
15:30
Josh: in which you engage with llms and is open source and available today so you
15:33
Josh: can find a link in the description down
15:35
Josh: in our show thank you ledger for sponsoring this episode and then i guess we
15:39
Josh: kind of have to talk about that like is it really a free case because like is
15:44
Josh: free ever really free and i guess there's two possible answers to this question
15:47
Josh: one is yes it's just being heavily subsidized
15:49
Josh: perhaps by the chinese government as a vampire attack to pull some demand from
15:53
Josh: the frontier labs into these chinese models the other is perhaps they're just using the data
15:57
Josh: one of the big problems that we have been seeing people complain about is data
16:00
Josh: retention and zero data retention policies because people don't want to feel like they are
16:06
Josh: Having their data harvested and used to train other models. And when you think
16:09
Josh: about giving away these free models, China for a fact is collecting all of the
16:14
Josh: inputs from the context window. And what's been happening is a lot of people
16:17
Josh: are using this on perhaps proprietary code bases.
16:20
Josh: People have been testing this model in their own private projects.
16:23
Josh: They've been using it for their own personal work.
16:26
Josh: And a lot of data gets derived out of that. I mean, it's a million token context
16:30
Josh: window in this one chat box and you just paste it all in and you let it go to town.
16:33
Josh: So I'm sure the amount of data that's collected from offering all of these tokens
16:38
Josh: just through collecting all the inputs has been tremendous.
16:40
Josh: And I'm sure that gives them a unique advantage in a way that some frontier
16:44
Josh: labs might not, because oftentimes they are criticized for their retention policies.
16:49
Josh: And people really want to make sure that there is that zero data retention.
16:51
Josh: China doesn't face the scrutiny, in particular, doing it in stealth kind of
16:55
Josh: shields that for at least the first week while people have used it.
16:58
Josh: So they were just kind of freely and liberally using this thing,
17:00
Josh: probably aware that they were having their prompts used for future training,
17:04
Josh: but not really caring too much.
17:06
Josh: So I thought that was also interesting. The other thing is that like the status
17:09
Josh: of Jeep as a company still isn't very strong.
17:12
Josh: They had, I think, like $100 million in revenue last year, and they had a net
17:16
Josh: loss of like seven times that or something like that. It's currently priced at 750 times revenue.
17:22
Josh: The company itself clearly can't sustain this type of...
17:26
Josh: Funding and this free giveaway forever. And I'm curious to know what the type
17:30
Josh: of business model eventually, if any, there is going to be in order to keep this thing alive.
17:34
Josh: Like, is it just going to be a government subsidy until they figure it out?
17:38
Ejaaz: The business model is its state back, dude. Yeah.
17:41
Ejaaz: All these Chinese labs are backed by the state. So that's how they get their
17:44
Ejaaz: infinite amounts of capital to be able to kind of fund all these things.
17:47
Ejaaz: And this is a strategy that China has been pursuing for three plus years now.
17:52
Ejaaz: They want to catch up with the frontier Western models. they distill attack
17:56
Ejaaz: a bunch of these model labs so that they can train similarly capable models.
17:59
Ejaaz: And then they flood the market with cheap intelligence, which means that Anthropic
18:03
Ejaaz: can't charge the highest price that they can for the high quality model that
18:06
Ejaaz: they have. OpenAI can't do the same.
18:08
Ejaaz: So this drives model costs down. This is what kind of Meta and Mark Zuckerberg
18:12
Ejaaz: is also trying to do, but failing to do as effectively as the Chinese labs.
18:16
Ejaaz: They're just very good at doing it. The second thing I'll say is,
18:19
Ejaaz: at least according to Open Rata from their official statement.
18:22
Ejaaz: They said that this stealth model provider, which we now know to be Jipu,
18:26
Ejaaz: is retaining prompts and responses, but explicitly state that they're not using
18:31
Ejaaz: it for training. Now, this isn't verifiable in any kind of way.
18:36
Ejaaz: Like, okay, China, you've never done anything nefarious before, I'm sure.
18:39
Josh: You're really going to give this away totally for free? Like,
18:41
Josh: no way. Coming from the people that stole all the items, I can go and generate
18:45
Josh: a full episode of Spongebob on those models.
18:47
Ejaaz: Exactly. Exactly. Well, this is another unique thing, right?
18:50
Ejaaz: Culturally, in China, and this applies for every vector of tech that they've
18:55
Ejaaz: built in the past, electric cars, mobile phones, et cetera.
18:58
Ejaaz: They don't really see copyright as a thing or IP infringement as a thing.
19:02
Ejaaz: They see, okay, if you have a good product, I can copy that.
19:05
Ejaaz: And Chinese in particular are very, very good at copying effectively.
19:10
Ejaaz: There was this great interview I watched with, who was it? It was Travis Kalanick,
19:14
Ejaaz: on David Semra's podcast. I don't know if you caught that, right?
19:16
Ejaaz: Absolute legend, right? Twice. And he talks about his story where,
19:21
Ejaaz: you know, how he brought Uber over to China and he made a success there,
19:25
Ejaaz: but he said it was very tough. But there was also some really unique cultural
19:28
Ejaaz: things that he needed to understand.
19:30
Ejaaz: And he realized that in America, or in the West at least, the innovation level
19:34
Ejaaz: is super high. Coming up with net new zero-to-one ideas, super high.
19:39
Ejaaz: In China, less so, but they are masters of copying. And he originally looked
19:43
Ejaaz: down at copying as a really menial thing.
19:47
Ejaaz: But he said that the way that they do it is masterful and artful in itself.
19:52
Ejaaz: And this, I feel like, is what some of these Chinese labs are really good at doing.
19:56
Ejaaz: It doesn't matter if they're earning $100 million. Their IPOs are popping off
19:59
Ejaaz: anyway. They're never going to compare to the Western IPO market at all.
20:02
Ejaaz: But the point is, they want to flood the market and kind of like keep America
20:06
Ejaaz: at a cost level where maybe they're spending too much money than they actually
20:09
Ejaaz: need to until they're able to figure out some kind of a breakthrough,
20:12
Ejaaz: which might be continual money. Who knows?
20:14
Josh: Maybe. We'll see. It's an exciting story and we're going to find out more.
20:18
Josh: So this will be covered and followed up in the roundup tomorrow.
20:22
Josh: Stay tuned for that. There's also a reference. It's kind of cool.
20:25
Josh: The main character of the story is Open Router.
20:27
Josh: And we just covered an entire Open Router episode just last week.
20:32
Josh: So everyone who is unfamiliar with Open Router and why it is so valuable,
20:35
Josh: please go and check out that episode.
20:37
Josh: And that's mostly the state of the mystery as of now. Ox Alpha has been solved,
20:43
Josh: but there is still a lot of behind the scenes that is left to uncover.
20:46
Josh: We're hoping to get that information shortly. when we do we will share it with
20:48
Josh: you right here on the roundup hopefully this week so stay tuned for that one
20:52
Josh: um this has been kind of a crazy week i mean it's
20:56
Josh: China's always throwing like, they're throwing wrenches in everyone's plans.
21:00
Josh: First it was Deep Seek. Now they're doing this like kind of really fun stuff.
21:03
Ejaaz: Did you see Quen as well?
21:05
Josh: Yeah, we got a new Quen model. We're going to talk about that tomorrow.
21:08
Ejaaz: Dude, they're just like spaying out these models. I know.
21:11
Ejaaz: And you know, the craziest part about all of this is I'm seeing rumors from
21:16
Ejaaz: trusted sources that a bunch of American frontier labs are also about to like drop a few models.
21:22
Ejaaz: So like, I'm just saying here, like, you know, we're getting like the best from
21:25
Ejaaz: the Chinese, We're getting the best from the West and there's just too many models to use right now.
21:29
Ejaaz: But if you're listening to this and you're curious about this Ox Alpha model
21:32
Ejaaz: or if you don't believe anything that we're saying, if you don't think it's
21:34
Ejaaz: good enough, go out and try it right now. You can use.
21:37
Josh: It for free
21:38
Ejaaz: On OpenRouter. Go check it out. There is no promo code. Just sign up,
21:42
Ejaaz: make an account and you can kind of like inference it and use it yourself.
21:45
Josh: And I'm curious, send us,
21:47
Ejaaz: Exactly. Like use the free tokens, please. Use the Chinese state funded money
21:51
Ejaaz: budget to kind of do your AI tasks.
21:54
Josh: Just don't give it your proprietary information.
21:56
Ejaaz: Do not, definitely don't do that because they're definitely using it to train
21:59
Ejaaz: the next model. But like, let us know in the comments whether you actually like
22:02
Ejaaz: the model, whether it actually is good at the toss that you give it.
22:05
Ejaaz: And if it magically improves the next day because it's continual learning. We don't know.
22:11
Ejaaz: Like, let us know in the comments, DM us, we're available on all socials.
22:14
Ejaaz: Yeah, anything else, Josh?
22:15
Josh: Yeah, no, just to, again, share this with your friends if they enjoyed it.
22:18
Josh: If you have something to say, please leave a comment. We try to read all of them.
22:22
Josh: If you enjoyed on your favorite podcast player you can leave us a five-star
22:25
Josh: review there if you're listening to this
22:27
Josh: you should be watching it's very fun visually you can watch this on youtube
22:30
Josh: on spotify and yeah as always thank you so much for watching feel free to go
22:35
Josh: get caught up on the other episodes and tomorrow we got a banger roundup like
22:38
Josh: it's just gonna there's so much stuff to talk about it
22:40
Ejaaz: Is our favorite episode of the week.
22:42
Josh: We've been in like a lull lately where there had i mean it's been exciting but
22:46
Josh: it hasn't been like oh my god this is crazy there's new frontiers forming and
22:50
Josh: like hopefully this is the first of many instances in which we're going to rev
22:53
Josh: back up that engine and get
22:55
Josh: a whole bunch of new models so i'm very excited stay tuned for all of that where
22:57
Josh: we'll be covering it and yeah we'll see you guys in the next episode