Latest / In the Loop AI / Apple's $200B Gut Punch To Google! The AI War Just Escalated

Transcript

Apple's $200B Gut Punch To Google! The AI War Just Escalated · In the Loop AI

• 0:03
  David: Welcome to the AI Rollup brought to you by the Limitless podcast,
• 0:07
  David: where we stay up to speed with the emerging trends and developments in the AI space.
• 0:11
  David: I'm David Hoffman here with my two co-hosts, Ajaz and Josh. Ajaz, how you doing?
• 0:17
  Ejaaz: I'm good, dude. It's been quite a week, like subtle on the model updates,
• 0:23
  Ejaaz: but we're starting to see some of
• 0:25
  Ejaaz: the ripples of some of these strongholds that big companies have on ai right
• 0:29
  Ejaaz: so we'll get into some of the stories later but effectively we're seeing like microsoft owning
• 0:35
  Ejaaz: pretty much the entire stack of ai and we'll get into like what kind of effects
• 0:39
  Ejaaz: that has but also apple suggesting that they might update their search and remove
• 0:46
  Ejaaz: google from the default caused google to lose
• 0:49
  Ejaaz: almost 200 billion dollars in two hours and i watched this happen in real time so
• 0:55
  Ejaaz: i think for the last two years or so we've seen kind of ai on this rocket ship
• 1:00
  Ejaaz: right you open ai pop up out of nowhere wow these are it's a really cool product chat gpt etc
• 1:05
  Ejaaz: but now we're starting to see okay right the big boys have come to play how
• 1:09
  Ejaaz: much like of a stronghold do they have on the entire stack and what can they
• 1:14
  Ejaaz: do with this can they control our entire lives or is there room for disruption
• 1:19
  David: Yeah, the Silicon Valley just tightens. They're just going at it,
• 1:22
  David: realizing exactly where their choke points are and where other people's soft bars are.
• 1:27
  David: Josh, this is the first AI roll up that's fully on the Limitless podcast. How does that feel, man?
• 1:33
  Josh: Oh, man, it's exciting. I'm stoked. I'm glad for, I'm grateful for all the people
• 1:37
  Josh: listening who followed with us. There's a lot of AI content that is coming that
• 1:41
  Josh: is super, super exciting.
• 1:42
  Josh: And I think this week is no different. I would frame this week for the people
• 1:45
  Josh: listening as like the highest stakes version of Game of Thrones that we're watching unfold.
• 1:50
  Josh: And it is the largest titans in the world that are fighting for the most valuable
• 1:53
  Josh: prize, which is basically this new form of intelligence, this new thing that
• 1:58
  Josh: can control a lot of people. So it's really exciting. This is a crazy week.
• 2:03
  David: I think that's why I think I really like these episodes. Is this just real world
• 2:06
  David: Game of Thrones being played out?
• 2:09
  David: One of the first episodes that we did, Josh, we talked about just like what
• 2:12
  David: does an intelligence explosion mean?
• 2:14
  David: And it's kind of a big take that I think throws people for a loop.
• 2:17
  David: But the conclusion of that episode that we did, which I encourage everyone to
• 2:20
  David: go watch, is that we are making God.
• 2:23
  David: God is the end product of this. And so no wonder, you know, Facebook and Apple
• 2:28
  David: and Google and OpenAI and, you know, four more companies in China
• 2:33
  David: are all like turning into these gigantic tech titans, trading blows with each other.
• 2:38
  David: And I think, Ajaz, that is what you are going to open us up with,
• 2:42
  David: the blows that are being traded this week as everyone tries to jostle to position
• 2:45
  David: to be the first person to create God.
• 2:49
  Ejaaz: Yeah, so let's start off with the first right hook. Apple is coming for Google's throne.
• 2:55
  Ejaaz: And they're supposedly, or rumored, to be launching their own AI search product.
• 3:00
  Ejaaz: Now, let me give you a bit of context here.
• 3:02
  Ejaaz: You know when you open up the Safari browser on your Apple laptop or on your
• 3:06
  Ejaaz: mobile, and you go to the dress bar and you just type in your search query, right?
• 3:11
  Ejaaz: That uses Google by default.
• 3:14
  Ejaaz: And Apple actually pays Google the privilege of $19 billion per year just to
• 3:22
  Ejaaz: have Google automatically default in that search bar, right?
• 3:25
  Ejaaz: So it's a big deal, just let alone to license Google search, right?
• 3:30
  Ejaaz: And so this week, Apple execs Eddie Q
• 3:34
  Ejaaz: says that the AI search via tools like, you know, like chat GBT or Claude of
• 3:40
  Ejaaz: perplexity, things that, you know, the three of us have used on this show pretty
• 3:43
  Ejaaz: frequently is way better than Google.
• 3:45
  Ejaaz: And I don't know about you guys, but I've kind of like started transferring
• 3:48
  Ejaaz: my habit to that as well. Like I don't use Google to search anything anymore.
• 3:52
  Ejaaz: Even when you search something on Google, it's funny, it shows you a little
• 3:56
  Ejaaz: AI summary at the top of the page as well. So it's using AI itself.
• 4:00
  Ejaaz: And they're saying that they're exploring revamping search, which means that
• 4:03
  Ejaaz: they might replace Google.
• 4:04
  Ejaaz: So as you can imagine, this had a pretty gnarly effect on Google's market cap.
• 4:10
  Ejaaz: As soon as this announcement went live, it knocked $200 billion,
• 4:14
  Ejaaz: which equated to about 10% of Google's entire market cap.
• 4:18
  Josh: Red candle, like
• 4:20
  Ejaaz: Super aggressive over two hours. And I watched this happen.
• 4:24
  Ejaaz: I saw a tweet and it said like, Google's lost $100 billion in market cap.
• 4:28
  Ejaaz: And I was like, nah, that's not true.
• 4:30
  Ejaaz: And then I just saw it like subsequently just go lower and lower and lower.
• 4:34
  Ejaaz: And I think for context here, Apple has the biggest distribution for Google
• 4:39
  Ejaaz: searches. They have, I think, 2.35 billion active devices and all of them have
• 4:44
  Ejaaz: Safari pre-installed and has Google as the default browser there, right?
• 4:48
  Ejaaz: So let's hypothesize for a second and say, if they decided to switch,
• 4:52
  Ejaaz: then Google loses billions of search queries per day.
• 4:56
  Ejaaz: I think another way to look at this is like Safari accounts for,
• 5:00
  Ejaaz: I think, about 18% of all web browsing activity. So that's page views.
• 5:04
  Ejaaz: And in the USA, it's 30%, right?
• 5:08
  Ejaaz: So if this switches, Google would lose nearly around a third of their browsing activity, right?
• 5:14
  Ejaaz: And the cost to Google is larger, actually. Like putting that together,
• 5:17
  Ejaaz: that is, okay, what? Hundreds of billions of searches per year and around $50
• 5:22
  Ejaaz: billion of like gross ad sales.
• 5:24
  Ejaaz: So Apple can basically do this in a single software upgrade, by the way. guys.
• 5:29
  Ejaaz: And don't forget like ChatGPT itself is re-holding their search function.
• 5:32
  Ejaaz: We mentioned it last week. So I just think this is completely nuts and it just
• 5:36
  Ejaaz: shows that the distribution power of Apple at the device level,
• 5:39
  Ejaaz: and Josh, you spoke about this last week as well, isn't something to be joked
• 5:42
  Ejaaz: around with, right? Maybe the moat is at the device level for now. What do you guys think?
• 5:46
  David: I think we've all been watching Apple struggle with what to do about AI.
• 5:51
  David: They had this like whole AI announcement over a year ago and we've all been waiting for it.
• 5:55
  David: Now people are realizing that their AI execution abilities has not been great.
• 5:59
  David: So I'm a little bit confused because this would imply that they are more ready
• 6:03
  David: to make moves in AI than I understand them to be.
• 6:07
  David: Josh is a huge AI Apple fanboy and hardware fanboy. So I bet he's got some pretty
• 6:12
  David: interesting takes here.
• 6:13
  Josh: I do have some takes. I've been, well, very disappointed with Apple's ability
• 6:16
  Josh: to execute on AI, which leads me partially to believe that the market really
• 6:20
  Josh: overreacted to this news.
• 6:22
  Josh: $200 billion getting wiped out in two hours is kind of crazy,
• 6:26
  Josh: particularly because Apple hasn't yet rolled out any AI. And this is not a new
• 6:30
  Josh: problem for Google. Where we've been talking about this for months,
• 6:33
  Josh: there's a trend of the 10 blue links dying.
• 6:35
  Josh: Google search is clearly on the way out. It is a slow bleeding thing.
• 6:38
  Josh: Sure, this adds some accelerant, but I don't think this is as detrimental to
• 6:42
  Josh: Google's business as we think.
• 6:44
  Josh: So Google and Apple kind of have two polar opposite problems.
• 6:47
  Josh: Apple has amazing products, but they can't actually create the software to use it.
• 6:51
  Josh: And Google has incredible software. The Gemini, the new Gemini models are about
• 6:55
  Josh: as good as they can get, but they're really bad at creating consumer products
• 6:58
  Josh: and experiences for people to use them.
• 7:00
  Josh: So what they're doing is they're kind of harming each other.
• 7:03
  Josh: Like Google has the search, Apple has the users, and Apple's severing that link,
• 7:07
  Josh: which hurts Google's revenue. I think 56% of Google's revenue is search.
• 7:13
  Josh: That percent is going to continue to drop anyway, regardless of whether Apple
• 7:16
  Josh: cuts them or not, because AI is just getting so good.
• 7:19
  Josh: I think one of the underrated aspects of this that people aren't discussing
• 7:22
  Josh: is the cost per query of search versus AI.
• 7:27
  Josh: So the cost to generate a token is an order of magnitude higher than the cost
• 7:32
  Josh: to generate a search query. So if
• 7:35
  Josh: AI actually can't support the volume. We don't have the GPUs.
• 7:37
  David: Sorry, the cost to generate a single token is a order of magnitude more than
• 7:42
  David: it is a single search. Yes, it is.
• 7:44
  David: And so for an average AI query, you need hundreds, thousands of tokens.
• 7:49
  Josh: Yeah, you need a lot more data, a lot more bandwidth, a lot more compute power
• 7:53
  Josh: because Google is really just
• 7:54
  David: Nearly a naive 10X. It is incredibly more expensive to do AI search than it is to do a Google search.
• 8:01
  Josh: Yes, perhaps not token, perhaps not order of magnitude per token,
• 8:03
  Josh: but per search at least you can think of it each search is at least 10 times
• 8:06
  Josh: more expensive than a or each ai search is more expensive than a single search query because google's
• 8:12
  Josh: query is really just an index and it's just going through an index and pointing
• 8:15
  Josh: to things uh this is actually generating net new content so what i hope google
• 8:19
  Josh: would do google has this unique opportunity where they have i think 270 million paying subscribers
• 8:24
  Josh: they need to figure out how to get them onboarded onto ai they need to create
• 8:28
  Josh: a product they need to create a service they need to to fight back against this
• 8:31
  Josh: trend of the 10 blue links dying.
• 8:33
  Josh: And we're kind of seeing that, like you just mentioned, where when you do make
• 8:36
  Josh: a search query, it shows you this little Gemini prompt at the top where it kind
• 8:39
  Josh: of summarizes their results.
• 8:41
  Josh: But it's not there yet. And I'm really hoping they could figure out a way to
• 8:43
  Josh: monetize this like huge consumer base that they have to save themselves. But I guess we'll see.
• 8:48
  Ejaaz: Why do we think they suck so much? It's kind of been the recurring story of
• 8:52
  Ejaaz: Google, no? For like the last couple of decades, they tried to do the whole
• 8:56
  Ejaaz: like social network. What was it called again?
• 8:59
  Ejaaz: I forgot. Oh, but Circles, that sucked and died.
• 9:03
  Ejaaz: They're trying to do the whole like smart device situation. In fact,
• 9:06
  Ejaaz: actually today they announced that they are rolling out Gemini to all their
• 9:10
  Ejaaz: smart devices, right, their smart TVs, their watches and stuff.
• 9:14
  Ejaaz: And I was thinking about this and I was like, wow, that's pretty cool.
• 9:16
  Ejaaz: They're trying to like get it on different devices, collect different kinds
• 9:18
  Ejaaz: of data. And then I was like, who uses this stuff aside from like the Android
• 9:23
  Ejaaz: cell phones? No one, right? So maybe it's a skill issue.
• 9:28
  Josh: It's certainly a skill issue. And we saw this with the entire AI trend since
• 9:31
  Josh: inception. Like Google invented the transformer. Their researchers created the
• 9:34
  Josh: transformer white paper.
• 9:36
  Josh: That's why we have all these models now. And they had it for many,
• 9:39
  Josh: many years prior to OpenAI coming around and launching the ChatGPT API and then ChatGPT the product.
• 9:45
  Josh: They just can't seem to figure out how to create good products with their technology.
• 9:48
  Josh: It's like a bunch of really bright, intelligent nerds who can do amazing research
• 9:53
  Josh: and form these amazing creations, but they don't actually know how to sells
• 9:56
  Josh: people and that's that's been a serious problem
• 9:59
  David: Maybe we've branded this week marketed this week as like the old tech giants
• 10:05
  David: titans just going blow for blow but it's also there's some notion of like senality
• 10:10
  David: in these things like apple google they're just,
• 10:14
  David: old companies and so they're like old people trading blows and you're kind of
• 10:19
  David: just cringing as you watch it because,
• 10:22
  David: they are hurting each other, but they're not creating any value,
• 10:26
  David: as much value as they are harm, if that is indeed what's going on.
• 10:28
  David: And then by contrast, you have the much more sharp, newer startups,
• 10:33
  David: the Anthropics, the OpenAIs, and they're untested.
• 10:36
  David: And OpenAI doing hardware is a completely new field, but there's something optimistic
• 10:42
  David: about that, I think, than let's just get some new blood into this game and see what happens.
• 10:48
  Ejaaz: So you think that's going to leave more opportunity for smaller startups to
• 10:52
  Ejaaz: take the throne, potentially, David?
• 10:54
  David: Yeah, I mean, if Apple can't make hardware. Yeah, exactly.
• 10:57
  David: Yeah. If Apple can't make hardware that has AI integrated and if Google can't
• 11:02
  David: figure out how to integrate a software into hardware, well, maybe it just takes
• 11:06
  David: a net new generation of companies to solve that problem.
• 11:10
  Ejaaz: Or maybe it takes a net new medium or device or experience to onboard those
• 11:15
  Ejaaz: people, potentially, right?
• 11:17
  Ejaaz: Right. Like I've again, I keep saying this on episodes, but I have never spoken
• 11:21
  Ejaaz: into my handheld device unless on the rare occasion that I call someone and it's usually my mom.
• 11:27
  Ejaaz: And now I'm doing it like every day. Right. So I feel like we're getting closer
• 11:31
  Ejaaz: and closer to this weird little metaversal scenario where everyone is just kind of
• 11:37
  Ejaaz: got their digital avatars real and talking to each other through those digital
• 11:41
  Ejaaz: avatars potentially. pretty crazy.
• 11:43
  David: We've talked about on the show about how too many of our friends outside of
• 11:48
  David: the tech circles are just not using ChatGPT. They're just not using AI.
• 11:53
  David: And I went to one and I just told them, hey, download ChatGPT,
• 11:57
  David: open up the voice feature in it and start talking to it.
• 12:02
  David: And then their first question was like, what do I talk about?
• 12:05
  David: I'm like, literally the first question that comes to mind, just go, just go for it.
• 12:08
  David: And then call me back after you do that. And they call me back like,
• 12:11
  David: 20 minutes later it's like i just talked to chat tpt for 15 minutes we just
• 12:15
  David: talked about whatever i was able to ask it questions and and talk about things
• 12:19
  David: that i was not able to talk to my friends about and i'm like.
• 12:22
  Ejaaz: Yes that's what
• 12:22
  David: So and i think that that is a homework assignment that i want to give listeners
• 12:26
  David: if you have not tried the voice,
• 12:30
  David: voice mode in chat tpt open up chat tpt after you're done listening to this
• 12:34
  David: episode and giving us a five-star review and subscribing on youtube open up
• 12:37
  David: voice mode and just start talking to it and see what happens and yeah just tell
• 12:43
  David: us in the comments about what that experience is like.
• 12:45
  Ejaaz: And i'm kind of curious to hear what you folks are talking to chat gpt about
• 12:49
  Ejaaz: i i hear a lot of my friends talk to them yeah
• 12:54
  Ejaaz: it's like it's like relationship advice it's like therapy for like issues growing
• 12:58
  Ejaaz: up as a kid as a kid and then a lot of people start to use it for like work
• 13:02
  Ejaaz: stuff but not like hey can you design this document for me it's like
• 13:07
  Ejaaz: i got into this situation with paul and i think he's encroaching on my territory
• 13:11
  Ejaaz: at work and i wonder like how can i like politely like put to him that he should
• 13:15
  Ejaaz: just basically fuck off and it's it's just he's like deeply personal and
• 13:20
  Ejaaz: i think very contextual to the individual and their own experience so it's kind of like outside of
• 13:26
  Ejaaz: Yeah, that's an interesting thing, actually. It's outside of the social etiquette,
• 13:30
  Ejaaz: you know? Do you know what I mean?
• 13:31
  Ejaaz: Like posting a picture of your ice cream cone at this new kind of ice cream
• 13:35
  Ejaaz: shop is like cool and accepted, right?
• 13:37
  Ejaaz: But talking to, or like speaking to this like weird social genie in a bottle
• 13:42
  Ejaaz: about like your personal afflictions with Paul is kind of weird. Yeah.
• 13:47
  David: It's useful for opening up about subjects that are hard to open up about.
• 13:52
  David: I think especially for dudes who it's easier to say, harder things to a robot
• 13:57
  David: than it is to another human,
• 13:59
  David: and so it kind of gives you that like safe space to be able to like talk about
• 14:03
  David: hard things that are on the cusp of things that you want to talk about but you don't have,
• 14:09
  David: the optionality you don't have the surface area you don't have the person to do it with.
• 14:12
  Ejaaz: So it's solving the the male loneliness epidemic
• 14:15
  David: Yes yes right yeah all of our all of our best friends yeah yeah yeah and this
• 14:21
  David: is how we give all of our data to it.
• 14:23
  Josh: Seriously that's how it's gonna happen i was listening to an interesting conversation
• 14:26
  Josh: that sam altman had actually it was at a sequoia session in sf this week and
• 14:30
  Josh: he was talking about how people use ai in different age brackets
• 14:33
  Josh: so he said the older people generally use it as an optimized search engine if
• 14:36
  Josh: they've even discovered it and then the like
• 14:41
  Josh: mid 20s to like late 30s group they use it kind of as
• 14:44
  Josh: a leverage tool to maybe help them with work and maybe solve some questions
• 14:48
  Josh: and then the younger people who are in middle school high school who are who
• 14:51
  Josh: are much younger they use it as an entire life operating system and they kind
• 14:54
  Josh: of load all their thoughts into it they load all of their homework into it
• 14:58
  Josh: they load everything into it and they consult it before before doing anything
• 15:01
  Josh: and it is like the single point of truth and the single reference that they
• 15:04
  Josh: can consult for for any sort of decision they make so i find that
• 15:08
  Josh: train interesting and going back to the point david that you made earlier that
• 15:12
  Josh: is there a new company that's that's going to dethrone apple and google and
• 15:15
  Josh: and open ai is going for it he very clearly said that they're going for the
• 15:19
  Josh: the ai subscription os plan for your entire life is is how can we build that
• 15:24
  Josh: that ecosystem for you have to
• 15:26
  David: Subscribe to your own life.
• 15:28
  Ejaaz: Uh that's literally the black mirror episode episode one of the new season you
• 15:32
  Ejaaz: know josh i as you were saying that, I was thinking like, why does that sound familiar?
• 15:37
  Ejaaz: And then I realized my sister, who is Gen Z, does exactly that.
• 15:40
  Ejaaz: So we briefly shared a ChatGBT account.
• 15:44
  Ejaaz: Lord knows how that's influenced my algorithm, actually. But she would literally
• 15:48
  Ejaaz: dominate 75% of conversations per week.
• 15:52
  Ejaaz: And it's just questions of like, random, like little social conversations that
• 15:56
  Ejaaz: she didn't know the answer to or things that she could have just,
• 15:59
  Ejaaz: I don't know, figured out herself.
• 16:00
  Ejaaz: And she was just offloading every single bit and I was like what is OpenAI going
• 16:04
  Ejaaz: to do with this data we should we should probably get some kind of like Gen
• 16:08
  Ejaaz: Z roundtable on this show at some point guys and like
• 16:12
  Ejaaz: Just kind of like interview them and see what they're using.
• 16:17
  David: There's this chart. We're going to shift subjects here. There's this chart that
• 16:20
  David: I saw on my timeline this week that I'm hoping one of you guys can explain to
• 16:23
  David: me. It's Microsoft's relationship with other companies in the space,
• 16:27
  David: OpenAI, Windsurf, Cursor, VS Code.
• 16:30
  David: And there's like lines of like ownership and partnerships between Microsoft
• 16:34
  David: and all of these things. And I kind of don't get it.
• 16:37
  David: I kind of understand that like, okay, Microsoft has their fingers in all of
• 16:40
  David: these different things and these weird partnerships. and deals and like org structures.
• 16:44
  David: But like, so I'm sharing it on the screen. It's like Microsoft has a 49% profit
• 16:49
  David: share with OpenAI, which owns Windsurf, which is forked from VS Code.
• 16:52
  David: And OpenAI is also an investor in Cursor, which is also forked from VS Code.
• 16:56
  David: But Microsoft owns VS Code.
• 16:58
  David: Can someone just like tell me what all of this is?
• 17:01
  Ejaaz: So Microsoft, if you guys recall about, I think about a year ago now,
• 17:06
  Ejaaz: guys, do you remember when there was that big saga with Sam Altman as CEO being
• 17:11
  Ejaaz: kicked out? by his board.
• 17:13
  Ejaaz: Do you guys remember this? When that happened, Satya Nadella,
• 17:17
  Ejaaz: CEO of Microsoft, kind of like rubbed his hands together and thought,
• 17:21
  Ejaaz: you know what, now is an opportune time to show support to OpenAI.
• 17:25
  Ejaaz: And we can do this in the sense of financial investment. So I think they invested
• 17:29
  Ejaaz: somewhere upwards of like $20 billion.
• 17:32
  Ejaaz: I don't know whether that was
• 17:33
  Ejaaz: set over a couple of years, but or just that one smack bang investment.
• 17:36
  Ejaaz: And they would invest a further $19 billion in providing all the GPU and CPU
• 17:42
  Ejaaz: infrastructure for OpenAI to run their inference and training.
• 17:46
  Ejaaz: So for OpenAI, that was a pretty good deal, right? It's like,
• 17:48
  Ejaaz: what, you're going to pay for all our compute hardware and help us train our
• 17:51
  Ejaaz: models and we don't, like, hang on, wait, what do you guys get in return?
• 17:54
  Ejaaz: And Microsoft said, well, we would love a bunch of equity and be like the majority
• 17:59
  Ejaaz: shareholder. And Simon Altman basically said, fuck off.
• 18:02
  Ejaaz: But what we can do is we can give you a 49% profit share of all services that
• 18:09
  Ejaaz: we charge for through OpenAI. And...
• 18:12
  Ejaaz: Any products we potentially acquire and integrate into OpenAI.
• 18:17
  Ejaaz: So if we look at the first top half of this diagram, Microsoft gets a 49% profit share of OpenAI.
• 18:23
  Ejaaz: And as we know, last week, OpenAI announced that they're acquiring Winsurf.
• 18:27
  Ejaaz: Now, a little bit of a kind of brain refresher for everyone.
• 18:31
  Ejaaz: Winsurf is the same as a company like Cursor, which were featured on the show before.
• 18:36
  David: It's the second place to Cursor.
• 18:38
  Ejaaz: Yeah, second place to Cursor. And the best way to think about it is these two
• 18:42
  Ejaaz: companies make it really easy to code up different kind of software apps and
• 18:46
  Ejaaz: these could these apps could be games these apps could be apps that appear on
• 18:49
  Ejaaz: your ios store to help you with
• 18:51
  Ejaaz: therapy personal training whatever you can dream up you can basically type in
• 18:55
  Ejaaz: a prompt like you do in chat gpt and it kind of codes up an app for you and
• 19:00
  Ejaaz: these companies like windsurf and cursor
• 19:02
  Ejaaz: Are now worth north of several single-digit billions because of this superpower.
• 19:07
  Ejaaz: And within this product itself, it's known as an integrated development environment.
• 19:12
  Ejaaz: It's basically carefully packaged software tools, code editors, and stuff like that.
• 19:17
  Ejaaz: So it abstracts away all the complexities of being a software engineer,
• 19:19
  Ejaaz: right? So people like you, me, and Josh can essentially code up stuff,
• 19:25
  Ejaaz: even though we haven't got any formal training or expertise in software engineering, right?
• 19:29
  Ejaaz: Now, there's a little circle on the bottom left, which says VS Code.
• 19:34
  Ejaaz: And people are like, I haven't really heard of VS Code. What is that?
• 19:38
  Ejaaz: It was the forefather to Cursor and Winsurf. Think of it as literally the thing
• 19:44
  Ejaaz: that gave birth to Winsurf and Cursor. In fact, there were rumors that went
• 19:48
  Ejaaz: around back in the day that Cursor had basically just forked VS Code's software.
• 19:53
  Ejaaz: So an interesting fact here, in 2024, a survey was done to see you know,
• 19:58
  Ejaaz: how much of the integrated development environment does, you know,
• 20:02
  Ejaaz: these different companies command?
• 20:03
  Ejaaz: VS Code commanded 74% of the entire market.
• 20:08
  Ejaaz: Now, I'm guessing that market share is down a bit, but that's basically the layout here.
• 20:13
  Ejaaz: So Microsoft has its hands on VS Code, and by proxy of OpenAI,
• 20:17
  Ejaaz: Windsurf and Cursor, because OpenAI acquired Windsurf, and OpenAI also invested
• 20:22
  Ejaaz: a hell of a lot of money into Cursor.
• 20:25
  Ejaaz: So I would say that some people have pitched this as like Microsoft owns the
• 20:29
  Ejaaz: entire AI stack and they kind of do, and I'll get into that in a second,
• 20:33
  Ejaaz: but it's mainly the coding stack that they have a stronghold on, on AI. Yeah.
• 20:38
  Ejaaz: Josh, I know you have a ton to say on this, so I want to throw it to you.
• 20:41
  Josh: Yeah, it's funny. It's like Microsoft funds OpenAI, OpenAI funds Cursor,
• 20:46
  Josh: and now buys the cursor rival Windsurf.
• 20:48
  Josh: And all this is on top of Microsoft's own open source code editor that they've forked.
• 20:53
  Josh: And now Microsoft's same dollar is getting taxed three times in the same ecosystem
• 20:57
  Josh: because they're competing against themselves with Copilot, which is Microsoft's own AI offering.
• 21:01
  Josh: So it's this very messy, conflicting thing in which Microsoft found themselves in.
• 21:06
  Josh: I'm not sure how happy they are about this because now Sam is trying to reduce
• 21:10
  Josh: that 49% profit share so that they can go public is the rumor this year.
• 21:15
  Josh: So it's kind of, it's a messy situation.
• 21:18
  David: Is the takeaway here that value is accruing to OpenAI more than it is to Microsoft
• 21:23
  David: and it's on the backs of Microsoft's work?
• 21:26
  Josh: Yeah. So there's an interesting dynamic that's part of the deal that we probably
• 21:29
  Josh: should mention where the nature of the deal is I think it's 75% of OpenAI's
• 21:34
  Josh: profits go to Microsoft until it recoups $13 billion, the total investment.
• 21:38
  Josh: Then after it's recouped its money, it reduces down to the 49%,
• 21:41
  Josh: which we're at, which goes until OpenAI reaches $92 billion of profit.
• 21:46
  Josh: And then after that, they're free and clear. So the special profit ends.
• 21:49
  Josh: So there is a cap on this upside.
• 21:52
  Josh: If Microsoft does hit that cap, that's still an amazing investment.
• 21:55
  Josh: They turn, what was it, $19 billion into $178 billion. So they've done really,
• 22:00
  Josh: really well. I think Microsoft is in a good spot.
• 22:03
  Josh: OpenAI, though, has this open-ended version of this profit margin.
• 22:07
  Josh: And if you ask Sam Altman, he'll probably say, oh, yeah, it's just a decade
• 22:11
  Josh: until we reach a trillion dollars of profit or something outrageous like that.
• 22:14
  Josh: So in the long term, it seems as if OpenAI stands to win.
• 22:18
  Josh: In the short term, Microsoft is doing well, but they also do have this little
• 22:22
  Josh: leech on them where now their largest investment is kind of leeching away users
• 22:28
  Josh: from Copilot, which is Microsoft's own offering.
• 22:30
  Josh: And it does create this weird dynamic where I'm not sure they're stoked about
• 22:34
  Josh: it. And now Sam has a lot of leverage because now he's trying to get that 49%
• 22:37
  Josh: profit share down even lower so he can make it more approachable to public markets
• 22:42
  Josh: when they try to go public.
• 22:44
  David: I I'm not privy to like the internal operations and like balance sheets and
• 22:49
  David: all this stuff but the idea of Microsoft you know I don't know 10xing their
• 22:54
  David: cash on billions of dollars,
• 22:56
  David: doesn't seem to feel so great when at the end of the day they don't own anything
• 23:01
  David: they get more money but they don't own any value they don't own,
• 23:06
  David: their game here the game of thrones is all these companies are trying to make
• 23:10
  David: God and there's no amount of money that is worth trading.
• 23:14
  David: The not having the opportunity of like owning a slice of God.
• 23:18
  David: And so I kind of see Microsoft as like the loser in this situation,
• 23:22
  David: no matter how many times they can multiply their cash.
• 23:24
  Ejaaz: So let's actually dig into that, David. I kind of went down a little bit of
• 23:29
  Ejaaz: a rabbit hole when one of you posted this in our chat.
• 23:32
  Ejaaz: And I was like, okay, well, hang on a second. What does Microsoft Stronghold
• 23:35
  Ejaaz: actually look like here, right?
• 23:37
  Ejaaz: So let's break it down. And then I have a comment on like where I think the moat is, right?
• 23:41
  Ejaaz: So in terms of compute, Microsoft Azure, which is their main cloud service,
• 23:47
  Ejaaz: right, competes with like AWS and Google Cloud, is OpenAI's exclusive cloud
• 23:52
  Ejaaz: for frontier model training and inference.
• 23:55
  Ejaaz: So Microsoft has dropped $13 billion on custom GPU clusters alone,
• 24:01
  Ejaaz: just to fund this, right? Just for the privilege of training OpenAI's frontier models, right?
• 24:06
  Ejaaz: So as long as Azure remains like this kind of, I don't know,
• 24:10
  Ejaaz: training substrate for open AI models, switching costs will keep Microsoft at
• 24:14
  Ejaaz: the center for now. But as you said, there's a heavy reliance on open AI there.
• 24:18
  Ejaaz: And I think this agreement is in until like 2030 or something, right?
• 24:22
  Ejaaz: They don't have overall board control, right? But Microsoft's hand is basically
• 24:26
  Ejaaz: firmly on Sam's shoulder saying like, hey, let's get first look into these other things, right?
• 24:31
  Ejaaz: The other part, which I thought was interesting is, Josh, you mentioned that
• 24:35
  Ejaaz: it's competing directly with Copilot, right? Microsoft Copilot.
• 24:38
  Ejaaz: But Copilot is based off of OpenAI's 4.0 and 4.mini models, right?
• 24:43
  Ejaaz: It's so incestual. Yeah, it's so incestual.
• 24:47
  Ejaaz: And Microsoft, I read this and I want to get kind of verification on whether
• 24:50
  Ejaaz: this is true, gets access to the 4.0 weights specifically.
• 24:55
  Josh: Yes, that is true.
• 24:57
  Ejaaz: So this is an important thing. So the more I think about it,
• 25:00
  Ejaaz: I'm like, huh, I think the moat is actually,
• 25:04
  Ejaaz: it only remains for Microsoft in this stronghold position if they continue to
• 25:08
  Ejaaz: get access to OpenAI's model weights, which I just don't see happening for the
• 25:12
  Ejaaz: longevity. Why would OpenAI do that?
• 25:14
  Ejaaz: And secondly, they're relying on developer stickiness through integrated environments
• 25:19
  Ejaaz: that use VS Code or whatever that might be.
• 25:21
  Ejaaz: Again, it depends on whether a good enough app or a software engineering community
• 25:26
  Ejaaz: decides to use Microsoft's VS Code.
• 25:27
  Ejaaz: And again, I don't see that as being too sticky to start off with, right?
• 25:32
  Ejaaz: And then, so when you look at that, and then on the coding side,
• 25:35
  Ejaaz: we can't forget that Microsoft owns GitHub as well, right? That's like 100 million
• 25:39
  Ejaaz: users and VS Code, which has accounted for about 74% of the IDE market share,
• 25:43
  Ejaaz: which we mentioned, right?
• 25:45
  Ejaaz: And then we have some of the equity spillover. So I'm not entirely convinced,
• 25:48
  Ejaaz: as you said, David, that Microsoft...
• 25:51
  Ejaaz: Their hold on this is going to remain quite permanent when OpenAI has so much power here.
• 25:56
  Ejaaz: If they were to switch to like some other provider in 2030, Microsoft's stronghold goes.
• 26:00
  David: I think this saga will continue. But ultimately, OpenAI has the IP.
• 26:06
  David: They own the model. They have the talent.
• 26:08
  David: And I think Microsoft is like scrambling to like work its way up the long tail of crumbs.
• 26:15
  David: And maybe it can like negotiate into like bigger and bigger crumbs.
• 26:18
  David: But that's kind of how I see this is like ultimately like OpenAI is the big
• 26:22
  David: winner here and there's no way for Microsoft to really change that.
• 26:25
  David: That's my takeaway from this section.
• 26:27
  Ejaaz: Yeah, I agree with you. I think the young blood has it here. Yeah.
• 26:32
  David: Yeah, let's get into the drama inside of OpenAI that happened this week because
• 26:35
  David: OpenAI has a new CEO that's not Sam Altman. So that's kind of crazy.
• 26:40
  David: I think we all understood like OpenAI is Sam Altman, but there's this new blog
• 26:45
  David: post and new news. OpenAI expands leadership with, I'm going to butcher this
• 26:48
  David: name, I'm so sorry, Fidigi Simo.
• 26:51
  David: And this is a message from Sam saying, hi everyone, I have some exciting news to share.
• 26:54
  David: I'm hoping to do this in a few weeks, but a leak accelerated our timeline and
• 26:58
  David: they go on to announce the new ceo,
• 27:02
  David: josh what's your takeaway with this is it real is this a real new ceo because
• 27:07
  David: the idea of cm altman not being at the helm of the ship is kind of crazy to me.
• 27:10
  Ejaaz: You guys ever watch the series the boys from amazon
• 27:14
  David: Prime yes yes never seen okay the superhero movie.
• 27:17
  Ejaaz: Yeah the concept josh seeing as you haven't seen it is it's all about superheroes,
• 27:24
  Ejaaz: but the dark side of superheroes.
• 27:25
  Ejaaz: So they're all gamed by political investments and they're like lobbied and a bunch of these things.
• 27:31
  Ejaaz: And it just shows the evil side of superheroes that it's not all fun and glory.
• 27:36
  Ejaaz: And there's this big corporation that manages these heroes, but of course,
• 27:41
  Ejaaz: managing a super powerful talent agency for superheroes.
• 27:44
  Ejaaz: But of course, when your superhero can laser beam you to death,
• 27:48
  Ejaaz: kind of who's in charge is the question there so they appoint the ceo and the
• 27:53
  Ejaaz: running joke in the entire series is she has absolutely zero power she runs
• 27:58
  Ejaaz: and gets them coffee and all these different kinds of things now i am not
• 28:02
  Ejaaz: suggesting that this is the case but the reason why i bring up that example
• 28:06
  Ejaaz: is she has been appointed as ceo
• 28:09
  Ejaaz: of applications at OpenAI. And it was branded across headlines that she was
• 28:14
  Ejaaz: CEO of OpenAI. That's just not the case.
• 28:16
  Ejaaz: She has some kind of equity ownership, but Sam Altman is very much still the
• 28:21
  Ejaaz: person, the man in charge.
• 28:23
  Ejaaz: And I think he's using this effort, at least he claims in the blog,
• 28:27
  Ejaaz: that he is going to now focus solely on research towards AGI
• 28:33
  Ejaaz: and towards aligning AI models to be to the benefit or betterment of humanity
• 28:38
  Ejaaz: and so we can focus his attention 100 on that whilst this new lady fiji simo
• 28:44
  Ejaaz: Can focus on building groundbreaking applications for the consumer end users
• 28:48
  Ejaaz: of open ai and i found this really interesting because well i had a few reactions
• 28:53
  Ejaaz: to this and i'm curious what you guys thought
• 28:56
  Ejaaz: Number one was like, Sam's been talking about AGI, AGI as the most important
• 29:01
  Ejaaz: mission of OpenAI since its literal birth, right?
• 29:06
  Ejaaz: And now suddenly he's making this big move to have someone who has a huge amount
• 29:10
  Ejaaz: of experience in building consumer applications to focus a large part of the
• 29:14
  Ejaaz: company's effort and resources and money on this.
• 29:17
  Ejaaz: So it made me think, are we actually as close to AGI as we thought?
• 29:20
  Ejaaz: Or maybe there are some other nuances and obstacles that we need to overcome.
• 29:25
  Ejaaz: Maybe it's at the technical hardware level, or maybe it's just the fact that
• 29:29
  Ejaaz: these things don't understand contextual awareness enough. And so they can't
• 29:32
  Ejaaz: be as personalized as we want. And they're only going to be used for niche cases. I don't know.
• 29:36
  Ejaaz: But on the optimistic side, I thought I looked into this Fiji Simo person,
• 29:40
  Ejaaz: and dude, her career and experience is absolutely stacked.
• 29:44
  Ejaaz: So she was just head of product or CEO at Instacart, which as you know,
• 29:48
  Ejaaz: is like one of the the top consumer applications in the western side of the hemisphere at least
• 29:54
  Ejaaz: and before that she was head of facebook whatever that title means but it sounds
• 29:58
  Ejaaz: again like a ceo of applications head of facebook at meta for 10 years which
• 30:04
  Ejaaz: is absolutely insane so she obviously has
• 30:06
  Ejaaz: a crazy amount of experience building consumer applications and i'm actually
• 30:09
  Ejaaz: really excited to see whether she builds out this amazing like chat gpt app
• 30:13
  Ejaaz: store i don't know what's your take josh
• 30:16
  Josh: Yeah, this is exciting for me because it displays like a clear divide in open
• 30:20
  Josh: AI. There are now two races that they're running. It's the consumer application
• 30:24
  Josh: race where we need to get these users. We need to lock them in.
• 30:26
  Josh: We need to create the best experience possible.
• 30:28
  Josh: And then it's the other half, which is we need to create AGI.
• 30:31
  Josh: And we need to create these systems that power this operating system that we want to build.
• 30:35
  Josh: So for me, this reminds me similar to X when Elon bought it and Linda Iaccarino became CEO.
• 30:40
  Josh: She's the one in charge of designing the day-to-day experience.
• 30:44
  Josh: But Sam is very much still the one who is calling the shots,
• 30:46
  Josh: who is providing the guidance to where the company is going.
• 30:49
  Josh: But to me, this signals to me like they're getting very serious about winning
• 30:54
  Josh: this race by having Sam fully committed to the AGI race. I think there is increasing competition.
• 31:00
  Josh: They now have a $500 billion investment with Project Stargate.
• 31:03
  Josh: Elon is trying to figure out how to get a terawatt of power to a data center.
• 31:08
  Josh: This is a huge scale battle that they're fighting, and the stakes are very high.
• 31:13
  Josh: So I think Sam focusing on AGI is good.
• 31:15
  Josh: This new CEO coming in, focusing on applications is good. I think she'll be
• 31:19
  Josh: responsible for building that.
• 31:21
  Josh: Life ecosystem that we talked about in the last segment sam is focused on building
• 31:26
  Josh: this super form of intelligence and i think that feels like a winning strategy
• 31:31
  David: Yeah so the way i'm hearing this is that sam is focused on building out the
• 31:35
  David: raw intelligence capacity of the open ai models like this this crude oil this
• 31:41
  David: oil this refined energy and then this new,
• 31:45
  David: ceo of applications is in charge of what that looks like, basically distribution,
• 31:49
  David: but like turning that energy and putting it into applications that fit in all
• 31:54
  David: different corners of our life.
• 31:55
  David: For whatever industry you're in, whether you're like, you know,
• 31:58
  David: pick an industry, any industry, some sort of application is going to be built
• 32:02
  David: that connects the AGI that Sam is trying to engineer to the end user,
• 32:07
  David: end product in a multitude of different ways that needs to come with like,
• 32:11
  David: you know, unique interfaces, unique experiences that fit into the different corners of the world.
• 32:15
  David: So that's kind of how I see, that's how that's how i summarize what you guys
• 32:18
  David: just kind of like put together there.
• 32:20
  Josh: Yeah sam's building the engine and fiji is building the the sports car on top
• 32:24
  Josh: of it or the yes the like really luxurious town that is powered by this single
• 32:28
  Josh: engine right it's kind of a good way to think about
• 32:30
  David: It yeah i can't looking at this photo she's very goth-like in this one particular photo.
• 32:35
  Ejaaz: I knew you were gonna say
• 32:36
  David: Something i can't not comment on how goth she looks and then when you go and
• 32:42
  David: you actually just search her on google she's just a very normal.
• 32:44
  Josh: Looking yeah welcome to the media
• 32:46
  David: This one particular photo just looks incredibly goth.
• 32:49
  Ejaaz: Yeah, why did they pick that? My God.
• 32:50
  David: I don't know, dude. All,
• 32:54
  David: right. All right. Let's get into the new models section.
• 32:58
  David: So every single week, almost every single week, there's some new models to talk
• 33:01
  David: about. This week is no exception. Gemini 2.5 Flash and also some new models
• 33:07
  David: out of China as well. Jaws 2.5 Flash. What do we need to know about it?
• 33:11
  Ejaaz: Okay. So firstly, I'm cheating a little bit here this week, guys,
• 33:16
  Ejaaz: because this model was actually announced, I think, three-ish weeks ago,
• 33:21
  Ejaaz: But we haven't updated Frontier models on this show for a few episodes.
• 33:26
  Ejaaz: And there are actually a few more recent ones.
• 33:28
  Ejaaz: Yeah. New to us. New to us. Okay. So Gemini 2.5 Flash.
• 33:33
  Ejaaz: It's funny. Earlier, we were talking about how Google kind of struggles to innovate,
• 33:37
  Ejaaz: at least at the consumer app level, but also at some other levels,
• 33:41
  Ejaaz: aside from just kind of like boring search queries or whatever,
• 33:44
  Ejaaz: you know, that important function that we rely on a lot.
• 33:46
  Ejaaz: Gemini 2.5 Flash is actually one of the leading models right now,
• 33:51
  Ejaaz: prior to O3 popping out. In fact, it is better than O3 in many ways,
• 33:56
  Ejaaz: but we just don't talk about that because OpenAI is still the darling child right now, right?
• 34:00
  Ejaaz: But it just doesn't get the love and attention that it needs.
• 34:03
  Ejaaz: And part of me thinks, you know, the reason behind this is because you can only
• 34:07
  Ejaaz: access it via API right now and not through their like main consumer interface
• 34:12
  Ejaaz: for Gemini. So their chat GPT interface equivalent.
• 34:16
  Ejaaz: But it is much better at reasoning and it shows huge gains over its previous
• 34:20
  Ejaaz: model, which is 2.0 Flash. So, you know, you can basically turn or toggle off
• 34:25
  Ejaaz: thinking or reasoning depending on what kind of thing that you're looking for.
• 34:29
  Ejaaz: And if you pull up this table, I think you guys just pulled up the comparison
• 34:32
  Ejaaz: of Gemini 2.5 Flash to like OpenAI 04 Mini, Croc 3 and all those kinds of things. It's
• 34:39
  Ejaaz: Comparatively cheaper than the majority of the models. I think only Gemini 2.0
• 34:43
  Ejaaz: Flash was cheaper than it. But of course, when you get a step change function
• 34:47
  Ejaaz: in actual ability for the model, it's way better.
• 34:50
  Ejaaz: And it's way better if you look across all metrics when it comes to things like
• 34:55
  Ejaaz: reasoning, science, coding, maths, etc.
• 34:59
  Ejaaz: I'm just wondering why it isn't talked about as much or maybe it hasn't been used as much.
• 35:03
  Ejaaz: My guess is it's because OpenAI is still so sticky and has all the kind of attention
• 35:09
  Ejaaz: when it comes to, hey, I want to try out this AI thing.
• 35:11
  Ejaaz: Like David, when you suggest to your friends, I'm guessing you don't go to them
• 35:14
  Ejaaz: and say, hey, get Claude or God forbid, Gemini. Yeah, no, you don't do that,
• 35:19
  Ejaaz: right? You're like, use ChatGPT.
• 35:21
  Ejaaz: So again, it just demonstrates no matter how good your model is right now,
• 35:24
  Ejaaz: at least, the stickiness with OpenAI is still there. Josh, I wonder if you have
• 35:28
  Ejaaz: any thoughts on Gemini 2.5 Flash before we move on.
• 35:31
  Josh: Yeah, it gets back to the thing we were talking about earlier,
• 35:33
  Josh: where you just have to create good products for these things.
• 35:36
  Josh: To have a new Frontier model is, it's impressive, but it's no longer like,
• 35:39
  Josh: oh my gosh, I must try this because
• 35:40
  Josh: my models that are on my like chat GPT desktop app are really great.
• 35:44
  Josh: And I'm pretty happy with that. So I think the reason why people aren't excited
• 35:48
  Josh: is because like I'm personally not going to go and engage with the API to test this thing.
• 35:52
  Josh: So therefore, it's not really super relevant to me. And I'm excited for the
• 35:55
  Josh: downstream effects if people want to create cool products on top of it.
• 35:58
  Josh: But as it stands now, I think myself and many other people are limited in their
• 36:02
  Josh: excitement because there's not many use cases for this for me.
• 36:06
  Josh: I think I'm most interested now in how this impacts my life or how this impacts
• 36:09
  Josh: the services that I use to improve my life.
• 36:11
  Josh: And if it's just an API call and there's not a lot of developers building interesting
• 36:15
  Josh: things that I use on a daily basis, then I'm like, oh, all right,
• 36:18
  Josh: well, O3 is still pretty great. I'll just stick with that.
• 36:21
  David: Right now, when OpenAI releases new models, like the new one is the O3 model, right?
• 36:26
  David: And that is going around in my life with my friends, talking to you guys on my Twitter sphere.
• 36:32
  David: It looks like and feels like back in the old days when Apple would release a new iOS update.
• 36:38
  David: And iOS would like make a material upgrade and all of a sudden our lives got
• 36:41
  David: enhanced. And it feels like that, like a new ChatGPT model launch is like this.
• 36:45
  David: Everyone's life is getting enhanced. We're all playing with it.
• 36:48
  David: It can do new things. It's very exciting.
• 36:50
  David: No other model release feels like that it doesn't have any impact on me because
• 36:54
  David: i don't use that app but it also does kind of beg the question why why doesn't
• 36:58
  David: this product exist and by this i mean a,
• 37:02
  David: a app like chat gpt or a front end like chat gpt that is a model aggregator,
• 37:08
  David: that also does the work of encode like remembering me and so right now i'm very
• 37:13
  David: excited that my chat gpt remembers all of my conversations and I'm excited to
• 37:19
  David: grow a relationship with ChatGPT to the point where like
• 37:23
  David: I start to work and go back and forth with ChatGPT on how best I want it to talk to me.
• 37:28
  David: I am like nurturing this relationship that I have with ChatGPT so that when
• 37:33
  David: it spits an output, it's spitting it out in a form factor that I like.
• 37:37
  David: Why is there not a product that aggregates all of the models?
• 37:41
  David: Because if you go to ChatGPT and you like do the little dropdown menu,
• 37:44
  David: there's like 4.0, 0.3, 2.5, 2.5 high mini, like blah, blah, blah.
• 37:48
  David: And I don't know what any of those means. And I don't care.
• 37:52
  David: And I just want to have this model aggregator do the work of choosing the best
• 37:56
  David: model intelligently while also doing the work of remembering me and my preferences.
• 38:02
  David: And I don't understand why that product doesn't exist yet.
• 38:04
  Josh: I'm not sure I have a great answer for that. I know very early on in the AI
• 38:09
  Josh: days, there was a company called Poe, which probably still exists.
• 38:12
  Josh: And they were an aggregator in the sense that you You can select your own model.
• 38:16
  Josh: It didn't have memory it didn't have the smart selection it just required the
• 38:19
  Josh: manual oversight to tap it i think we're kind of seeing this happening on the
• 38:24
  Josh: enterprise level with windsurf which we mentioned last week where windsurf is
• 38:27
  Josh: kind of serving as this orchestration layer
• 38:30
  Josh: they call it where as you submit a problem it will choose the optimal solution
• 38:34
  Josh: for that problem based on the model but in the consumer world i would imagine
• 38:39
  Josh: there's probably some sort of technical or privacy restriction
• 38:43
  Josh: on the model level preventing that from happening because it If you think about
• 38:46
  Josh: Apple and their App Store, which is probably the closest connection to this,
• 38:49
  Josh: is they won't allow any third-party applications on an iPhone.
• 38:53
  Josh: They must be approved. They must go through the centralized service.
• 38:55
  Josh: Open AI's largest moat right now is memory, is the new feature that they rolled out. So I'm not sure.
• 39:01
  Josh: That's why I'm thinking there's probably some sort of technical limitation preventing that from happening.
• 39:05
  Josh: It's because I find it hard to believe they would give up all of that data for
• 39:08
  Josh: free to an aggregator that can become much more powerful than them.
• 39:12
  Ejaaz: Yeah, I think cost is definitely a major function. Have you guys seen Open Router
• 39:16
  Ejaaz: by any chance? They kind of do what we're discussing right now.
• 39:20
  Ejaaz: This is the company set up actually by the ex-CTO and co-founder of OpenSea, Alex.
• 39:25
  Ejaaz: So if you pull up Open Router right now, just shove it into Google,
• 39:30
  Ejaaz: You'll see that it's described as the open interface for LLMs.
• 39:34
  Ejaaz: And basically, you can pull across any kind of model that they integrate with.
• 39:40
  Ejaaz: And it includes all of the frontier models from Western companies,
• 39:44
  Ejaaz: as well as some Chinese or Southeast Asian companies like Alibaba or DeepSeek.
• 39:50
  Ejaaz: And actually they're also famous for privately screening
• 39:55
  Ejaaz: if that's the right term new models before they even officially get announced
• 39:59
  Ejaaz: so open ai's 4.0 and o3 were actually tested privately you just didn't know
• 40:05
  Ejaaz: that you were testing it you was given a private code name like vulcan or something like that
• 40:10
  Ejaaz: and people were like wait this model is so good and it beats this benchmark
• 40:14
  Ejaaz: and that benchmark and everyone started speculating oh this might be open ai's new thing
• 40:18
  Ejaaz: And so they have a direct relationship where you basically type in a prompt on this website or you
• 40:23
  Ejaaz: build an app based off your API and it can select whatever model you want.
• 40:27
  Ejaaz: Now, I think the way that it's configured right now for this particular company
• 40:31
  Ejaaz: is you can kind of select which model gets picked for whatever type of query.
• 40:36
  Ejaaz: So it's not quite autonomous and smart.
• 40:39
  Ejaaz: And as far as I know, they don't have a memory function. Maybe Alex is going
• 40:42
  Ejaaz: to work on that going forwards. but I don't know this is kind of like the underpinnings
• 40:47
  Ejaaz: of something like that maybe
• 40:49
  David: Imagine being Alex and you leave OpenSea in like 2022, right at the end of the
• 40:54
  David: NFT summer to start an AI model aggregator project in 2022.
• 41:00
  Ejaaz: And then Sam Holman picks your company. Yes.
• 41:03
  David: Wow. Just perfectly selling the top of NFTs.
• 41:07
  Josh: Well done.
• 41:08
  Ejaaz: Well done.
• 41:09
  David: That's not the only model that came out. No. We also have one out of China as well.
• 41:15
  Ejaaz: Yeah, I mean, we haven't spoken about our friends across the ocean, but DeepSeek,
• 41:20
  Ejaaz: the ones that kind of like shocked the entire Western AI world by creating a
• 41:25
  Ejaaz: model that was much cheaper, but more effective or as competent as OpenAI's
• 41:30
  Ejaaz: frontier models, which they spent billions and billions on
• 41:33
  Ejaaz: and DeepSeek just spent a couple hundred million millions on, is supposedly
• 41:37
  Ejaaz: releasing their next major model update.
• 41:40
  Ejaaz: So it's not officially out just yet, but this has kind of been leaked that the
• 41:44
  Ejaaz: model will have 1.2 trillion with a T parameters, and it'll be a hybrid mixture
• 41:50
  Ejaaz: of experts model, which means that
• 41:53
  Ejaaz: whenever someone queries the model the entire set of parameters
• 41:57
  Ejaaz: aren't queried at once which is typically how a lot of these models work instead
• 42:01
  Ejaaz: it's more efficient and it might hit like 37 billion parameters which is needed
• 42:06
  Ejaaz: to answer your particular query or your question right so it's much more efficient
• 42:10
  Ejaaz: and that's how they like mainly drive down the cost
• 42:12
  Ejaaz: it's also going to be almost a well almost 99 but it's 97.3 cheaper than gpt4o
• 42:20
  Ejaaz: which is just again insane and we were kind
• 42:23
  Ejaaz: We were kind of talking about like, you know, earlier, what kind of moat would
• 42:26
  Ejaaz: Microsoft have with like Copilot or Gemini have with 2.5 Flash?
• 42:32
  Ejaaz: And you could argue that being cheaper per query or per token could be a moat there, right?
• 42:37
  Ejaaz: Because it all adds up. And if you're a startup that's consuming a ton of data
• 42:41
  Ejaaz: and you can get kind of like near the same kind of quality that OpenAI results
• 42:45
  Ejaaz: give you, then I'm going to take that cost cut and trade off and just use this model instead, right?
• 42:50
  Ejaaz: It's trained on a hell of a lot of data, 5.2 petabytes.
• 42:54
  Ejaaz: And it's meant to have a much better reasoning model. Okay, blah,
• 42:58
  Ejaaz: blah, blah. We always speak about these different characteristics.
• 43:01
  Ejaaz: But what's something that's actually cool here is that last point on this tweet.
• 43:05
  Ejaaz: 82% utilization in Huawei Ascend 910b.
• 43:11
  Ejaaz: What they're referencing there is a Chinese manufactured AI chip.
• 43:16
  Ejaaz: And I just want to emphasize how important this is because to date,
• 43:20
  Ejaaz: NVIDIA has literally, and I'm not making this up, dominated like,
• 43:24
  Ejaaz: Like 95% plus of the chip manufacturing industry.
• 43:29
  Ejaaz: And that includes like the compute side of things, chip design,
• 43:32
  Ejaaz: like which informs how your model is going to get consumed, data trained, etc.
• 43:37
  Ejaaz: It influences a heck of a lot. We just don't speak about it as much, right?
• 43:41
  Ejaaz: And now we have China, which was supposedly meant to be a decade behind the
• 43:44
  Ejaaz: Western world, coming up with a chip that is 75% as good as NVIDIA's flagship
• 43:51
  Ejaaz: or one of their old flagship chips, the H100s.
• 43:54
  Ejaaz: So it's just really interesting to see how much China, and we're going to get
• 43:58
  Ejaaz: into this in the next topic, but DeepSeek, Chinese manufacturers with Huawei,
• 44:03
  Ejaaz: they're building their own kind of empire.
• 44:05
  Ejaaz: So maybe it's not Game of Thrones on the Western world. Maybe it's just Game
• 44:08
  Ejaaz: of Thrones, but then China just kind of like swoops in and takes out the head.
• 44:12
  Ejaaz: Josh, I know you have a ton of experience on the hardware side of things.
• 44:15
  Ejaaz: I know you love nerding out about this. I want to hear your take on this?
• 44:18
  Josh: I'm obsessed with DeepSeek in the way that I kind of love the Grok team as well,
• 44:22
  Josh: because they just have the fastest rate of acceleration.
• 44:25
  Josh: And I think when you're competing on these like long time horizons with really
• 44:28
  Josh: high stakes and strong exponential curves, the rate of acceleration is probably
• 44:32
  Josh: the most important thing to pay attention to.
• 44:33
  Josh: So Gemini and OpenAI are very clearly the leaders in terms of frontier models right now.
• 44:39
  Josh: I think what we're seeing with DeepSeek is this hyper focus on efficiency.
• 44:44
  Josh: And that efficiency has exponential scales that I think will probably quickly
• 44:50
  Josh: exceed that of the top leading models in the US.
• 44:53
  Josh: What's interesting is they're winning on the hardware front and they're winning
• 44:56
  Josh: on the software. Well, they're not winning on the hardware front,
• 44:58
  Josh: but they're rapidly getting closer.
• 44:59
  Josh: You said it had each, what was the percentage of the H100 that they were able to capture? 75%.
• 45:05
  Josh: So they're 75% there. And that number was much lower just a few months ago.
• 45:09
  Josh: What China has is really great manufacturing capabilities. They have really
• 45:14
  Josh: smart developers who are hyper-focused on this resource constraint.
• 45:18
  Josh: And what we're getting is these incredible models. 1.2 trillion parameters, I think.
• 45:22
  Josh: Is that a record for a public model if it does go public?
• 45:26
  Ejaaz: Yeah, Meta's supposedly working on a 2 trillion parameter model.
• 45:28
  Ejaaz: But, you know, we're talking about sizes here.
• 45:33
  Josh: So far, if this were to release today, this would be the largest public model.
• 45:37
  Josh: And the cost of it are a fraction of what it costs to use GPT-40.
• 45:41
  Josh: So the trends that we're seeing is an increase in efficiency,
• 45:45
  Josh: an increased rate of acceleration in progress, and an aggressive decrease in
• 45:49
  Josh: cost per query or cost per token.
• 45:51
  Josh: And those three things are like really powerful effects that are going to affect
• 45:56
  Josh: markets in a really big way.
• 45:57
  Josh: I think as we get more power at a lower cost, it unlocks the amount of productive
• 46:02
  Josh: output of these tokens. I forget the amount of trillions of tokens that Microsoft
• 46:05
  Josh: said they generated in the last quarter, but it was many, many trillions of
• 46:08
  Josh: tokens. And if you can get
• 46:09
  Josh: 10 times that token output for a fraction of the cost, that's a lot more intelligence
• 46:13
  Josh: you can use to either distill models like we see in the smaller models like
• 46:16
  Josh: that Quinn is doing, or just to apply to even larger models and to do reinforcement learning on those.
• 46:22
  Josh: So I think the cost per token and the efficiency of these models is something
• 46:26
  Josh: to really pay attention to.
• 46:27
  David: Okay, so if R2 is coming in with a 97% cost reduction, as in the cost of running
• 46:33
  David: this model is 97% cheaper, that means we can just, you know,
• 46:38
  David: do 33 times more compute for the same amount of cost.
• 46:41
  David: So you get 33 times more compute for the same cost.
• 46:44
  David: But does that result in a 33 times better outcome?
• 46:48
  David: Because I don't think that that's true. I think the queries that you get,
• 46:51
  David: the products that you get to come out of that compute is running up on some
• 46:55
  David: sort of like constraint, right? like we don't just get 33 times better outputs right.
• 47:00
  Josh: Not directly so the constraint mostly is in the actual model intelligence and
• 47:04
  Josh: those benchmarks that you see so this would be if it gets released 1.2 trillion
• 47:07
  Josh: dollar or trillion token model or parameter model and that would be the upper
• 47:12
  Josh: bound of intelligence so it's not
• 47:14
  Josh: creating that new intelligence it's not agi but what we do have is we have a lot more use of it so
• 47:20
  Josh: currently the the cost per token here is much more expensive
• 47:24
  Josh: and when you get a cheaper token you unlock the second order effects of that
• 47:27
  Josh: which means consumer applications become a lot more cheap it means that we can
• 47:31
  Josh: get access to just a lot more intelligence a lot cheaper and a lot more accessible
• 47:35
  Josh: and i'm not sure i have great examples of
• 47:37
  Josh: of where that comes but i'm not sure there's anyone in the world that wants
• 47:40
  Josh: that doesn't want more intelligence and doesn't want more queries or shots on
• 47:43
  Josh: goal at doing this so it's not that it is a smarter form of intelligence But
• 47:47
  Josh: the amount of it and the accessibility of it goes up significantly.
• 47:52
  David: Well, that's kind of just on brand with like what China tends to bring to the
• 47:55
  David: table. It seems like the United States, Silicon Valley, brings these net new
• 48:00
  David: innovations, more intelligence, increasing IQ, and then China just makes that cheaper.
• 48:06
  Josh: Yes. Although this time, 1.2 trillion parameters, that is the new top dog.
• 48:10
  Josh: So as of now, assuming Benchworks come out well, that would be highest intelligence
• 48:14
  Josh: and lowest cost, which is like a huge win.
• 48:17
  Ejaaz: I actually had a question. If the cost per query goes down and we kind of continue
• 48:25
  Ejaaz: this trend or this directional trend towards post-training via reinforcement learning, right?
• 48:31
  Ejaaz: Wouldn't this basically enable them to scale that up?
• 48:35
  Ejaaz: So you could end up with a much smarter model at a cheaper cost because it costs
• 48:40
  Ejaaz: less to query in itself, right?
• 48:41
  Ejaaz: So you could just chuck in a bunch of, you know, reasoning traces,
• 48:45
  Ejaaz: which for those of you who don't know is kind of like questions,
• 48:48
  Ejaaz: which will show a question A and then an answer B.
• 48:53
  Ejaaz: And the model basically has to figure out how to get from A to B.
• 48:55
  Ejaaz: And that's what makes the model smarter through reinforcement learning.
• 48:59
  Ejaaz: Josh, don't you think this will basically open up
• 49:01
  Ejaaz: an opportunity for them to scale to a smarter model i'm guessing that's how
• 49:04
  Ejaaz: they're doing it right now right alibaba is doing the same thing with quen i
• 49:07
  Ejaaz: mean yeah the new model that they just released is only 235 billion parameters
• 49:12
  Ejaaz: but it's as smart if not better at certain math coding and science benchmarks
• 49:18
  Ejaaz: and maybe not so much the social reasoning side of things but on certainly across
• 49:22
  Ejaaz: those benchmarks than some of the frontier trillion
• 49:25
  Ejaaz: dollar not trillion dollar trillion Yeah,
• 49:28
  Josh: No, that's a really great point. I think you're right. What we probably see
• 49:32
  Josh: is similar to Quinn in the distillation of models and then the hyper specialized
• 49:38
  Josh: version of those models.
• 49:39
  Josh: So as basically the way these distilled models work is there's this giant foundation
• 49:43
  Josh: model and it uses reinforcement learning to create these good prompts,
• 49:47
  Josh: these really high quality prompts that get fed into a smaller model that can
• 49:50
  Josh: run locally on a, let's say a laptop or your own device that you have at home.
• 49:54
  Josh: What happens from this probably is is now that you have this higher form of
• 49:57
  Josh: intelligence that you could distill much more cheaply you can create a lot more
• 50:01
  Josh: of these distilled models that are even even tighter in size even smaller and
• 50:04
  Josh: maybe enough to run on your iphone
• 50:06
  Josh: that are effective enough to actually work and maybe you get a distilled model
• 50:09
  Josh: that's kind of like a sirion steroids type thing that you could run locally
• 50:12
  Josh: on a phone but maybe you also just get a really great
• 50:16
  Josh: hyper focused model on a specific type of math problem or a specific category
• 50:20
  Josh: of science or like a very specific data set that it's trained on
• 50:24
  Josh: that costs basically nothing because it's been distilled down and it costs so
• 50:27
  Josh: cheap but it's really really smart at this one particular thing and if you have
• 50:31
  Josh: a series or a network of these nodes that are hyper trained on a specific category or a specific problem
• 50:37
  Josh: then you could kind of build this mesh network on top to to interact with each
• 50:40
  Josh: one based on your the needs of the query so it creates this interesting modular approach
• 50:45
  Josh: versus the large foundation model where if you can distill these things down
• 50:49
  Josh: cheaply enough and broadly enough into all these hyper-localized little nodes,
• 50:53
  Josh: then you could access a lot of general intelligence very, very cheaply.
• 50:57
  Ejaaz: It reminds me of another example. I mean, like the general theme that we're
• 51:00
  Ejaaz: talking about here is Chinese AI manufacturers can basically make things cheaper and more efficient.
• 51:06
  Ejaaz: But the difference, which is typically what they've done with like mobile phones
• 51:09
  Ejaaz: when Huawei is copying a bunch of iPhone stuff.
• 51:12
  Ejaaz: But the difference here is that they're actually innovating on the design and
• 51:15
  Ejaaz: the execution, which is really important.
• 51:18
  Ejaaz: And it reminded me that Huawei this week launched, I think it was like a 718
• 51:24
  Ejaaz: billion parameter model,
• 51:26
  Ejaaz: but it was trained on 6,000 of their new Ascend chips, which is like their equivalent
• 51:34
  Ejaaz: of like NVIDIA H100s or whatever you want to call it, right?
• 51:37
  Ejaaz: And this was a significant reduction from the 8,000 chips that they used to
• 51:43
  Ejaaz: train their previous model, which was half the parameter count.
• 51:47
  Ejaaz: So basically, chip count down,
• 51:50
  Ejaaz: model intelligence up, which is a scary but pretty exciting trend to see.
• 51:55
  Ejaaz: I guess the benefit of this so far is that us on the Western side can kind of
• 52:00
  Ejaaz: benefit from it because they're open sourcing all these models and they're probably
• 52:04
  Ejaaz: trying to vamp attack some of the bigger companies in America.
• 52:06
  Ejaaz: But at some point, they're going to close the gates because they've achieved
• 52:09
  Ejaaz: parity, if not betterment of these models.
• 52:12
  Ejaaz: So it's an exciting trend to kind of watch.
• 52:15
  Ejaaz: And it's not one that's getting publicized in the Western media as much.
• 52:20
  David: I'm just seeing as all of these numbers come down ever since we started this
• 52:23
  David: podcast, you know, token parameters go up, chips go down, and the rate of this
• 52:29
  David: happening is just not slowing whatsoever, which makes me think that just like,
• 52:37
  David: it just keeps on going it just keeps on going.
• 52:39
  Ejaaz: Yeah so we're going to end up in this like infinite abyss david where we can
• 52:43
  Ejaaz: you you can talk to chat gpt as much as you want you know earlier on josh was explaining how
• 52:49
  Ejaaz: you know doing search queries via ai is so much more expensive than doing a
• 52:52
  Ejaaz: google search query but maybe not for long maybe not for much longer
• 52:56
  David: By the end of this year it's not going to be the same.
• 52:58
  Ejaaz: Yeah yeah
• 52:59
  David: There are some things to talk about on the crypto side of things where we got
• 53:03
  David: this podcast started. So cookie.fun is a website I got pulled up right now,
• 53:06
  David: which shows the total market cap of all AI agent tokens.
• 53:11
  David: Now, maybe to call a spade a spade, these are meme coins associated with AI agents.
• 53:17
  David: I think that's a fair take. But this sector of crypto, the AI sector of crypto,
• 53:22
  David: has gotten a lot of life back in it over the last two weeks.
• 53:25
  David: Ijaz, tell us what happened.
• 53:27
  Ejaaz: Yeah, so in this particular instance, life translates to money.
• 53:31
  Ejaaz: The total market cap of these ai agents if you remember back in the day i think
• 53:36
  Ejaaz: it peaked at like 25 billion and then it got crushed basically over subsequent months to about 3.2
• 53:43
  Ejaaz: billion dollars now over the last three weeks it has almost tripled oh just
• 53:49
  Ejaaz: over tripled actually to 11.4 billion dollars as of this screen recording right
• 53:54
  Ejaaz: now right and as you can see
• 53:56
  Ejaaz: charts are green across the majority of the spectrum of things.
• 54:00
  Ejaaz: And you're probably wondering, hey, guys, what's happened? Like,
• 54:04
  Ejaaz: has some like crazy innovations been going on? Like, give me the summary as
• 54:07
  Ejaaz: to why this is happening.
• 54:08
  Ejaaz: So let's go through some of the takeaways.
• 54:11
  Ejaaz: Number one, virtuals is leading this entire kind of progression,
• 54:15
  Ejaaz: right? So if you remember back in the day, we used to speak about kind of like
• 54:19
  Ejaaz: the top AI agent protocols. Virtuals was one of
• 54:26
  Ejaaz: Over the bear market, all three of them were crushed, but all three teams were
• 54:30
  Ejaaz: kind of heads down focusing on pioneering new kinds of innovations.
• 54:34
  Ejaaz: And virtuals really kind of like came through as the leader here.
• 54:37
  Ejaaz: So they specifically added over $600 million in market cap over the last two
• 54:42
  Ejaaz: weeks. And that's just in the virtuals token. So it's back over a billion dollars.
• 54:46
  Ejaaz: And they've made a number of changes since we last checked in on them.
• 54:50
  Ejaaz: So if you pull up this tweet, which basically kind of like has like this fun
• 54:54
  Ejaaz: little table as to what they've basically done.
• 54:57
  Ejaaz: So in January 2025, they go through things like agent revenue,
• 55:01
  Ejaaz: agent commerce protocol, fee structure, chain.
• 55:04
  Ejaaz: Basically, everything on the left was kind of unclear, a bit of flimsy, kind of Ponzi-like.
• 55:08
  Ejaaz: And on the right side, they have a much more structured product and focus strategically
• 55:13
  Ejaaz: for what they're trying to build and how they're supporting creators or agent
• 55:17
  Ejaaz: developers on their side of things. But let me explain what any of this means
• 55:21
  Ejaaz: and what the highlight is.
• 55:23
  Ejaaz: So they created a new agent launchpad. And for those of you who don't know what
• 55:27
  Ejaaz: an agent launchpad is, it's basically kind of like a
• 55:31
  Ejaaz: Cursor, if you like, where you can go on, you can use a no-code environment
• 55:36
  Ejaaz: to design what your agent might look like or what it might do,
• 55:39
  Ejaaz: what functions it might serve.
• 55:41
  Ejaaz: And then you can launch that agent. And usually you can associate a token with
• 55:45
  Ejaaz: that agent. So it's kind of like you're IPO-ing an agent.
• 55:48
  Ejaaz: And they created this thing called a Genesis launchpad, which is meant to be
• 55:51
  Ejaaz: a fairer way to launch agents.
• 55:54
  Ejaaz: So they've completely replaced the bonding curve model with something they call
• 55:59
  Ejaaz: proof of contribution, which basically means you need to do something helpful
• 56:04
  Ejaaz: towards the virtual ecosystem if you want to get access to buying the tokens
• 56:08
  Ejaaz: for that particular agent.
• 56:10
  Ejaaz: And contributions are measured in something called I think it's like virtual
• 56:14
  Ejaaz: points or they're kind of like loyalty points
• 56:17
  Ejaaz: and there are many ways to earn these points you can stake virtuals you can
• 56:21
  Ejaaz: yap or talk about virtuals on x or you can simply just hold virtuals tokens
• 56:26
  Ejaaz: or certain virtuals tokens
• 56:28
  Ejaaz: then there's kind of like this pre-sale and people can commit and and the
• 56:35
  Ejaaz: Of course, that's not the only thing that people are talking about.
• 56:37
  Ejaaz: They're talking about the returns that they're getting on this thing.
• 56:39
  Ejaaz: So the average return multiple on each of these tokens, David,
• 56:44
  Ejaaz: since they launched this launchpad is about a 12 to 20x, which is insane, right?
• 56:50
  Ejaaz: Until you realize that the FTV that you're investing in is like $212,000,
• 56:56
  Ejaaz: not million, $212,000. So you kind of have like a guaranteed ROI.
• 57:02
  Ejaaz: What I don't like about this is probably what you can already guess.
• 57:05
  Ejaaz: And if you pull up this tweet by Vader is it's mainly just about the money right now.
• 57:10
  Ejaaz: And it's not really about much innovation. And I know the virtuals team personally
• 57:15
  Ejaaz: are working very hard on creating value.
• 57:17
  Ejaaz: And you could also argue that on the traditional AI space that they haven't
• 57:21
  Ejaaz: kind of created anything of definitely more marginal value on the enterprise
• 57:25
  Ejaaz: side of things, but not as valuable as we'd expect to see. So it's kind of like,
• 57:29
  Ejaaz: you know, they're kind of fighting for the same kind of thing.
• 57:32
  Ejaaz: But it's good to see that they're kind of like focused on this thing,
• 57:36
  Ejaaz: right? The second thing is virtuals updated their fee.
• 57:40
  Ejaaz: Flywheel, which means that 70% of fees get given back to the creators.
• 57:45
  Ejaaz: Now, previously, that wasn't the case. And the effect that this has had is basically
• 57:49
  Ejaaz: any kind of agent developer that launches a coin can basically get continual
• 57:53
  Ejaaz: money to continue developing their product.
• 57:56
  Ejaaz: And that's been a really good way to kind of like sustain the ecosystem.
• 57:58
  David: So for the AI listeners out there, it's probably worth knowing that there's
• 58:01
  David: this meta happening in the crypto industry around these token launch pads.
• 58:06
  David: The meta that's going on this week is this like Believe app thing,
• 58:09
  David: which allows you to launch a token that is paired with, but other than that
• 58:15
  David: has no material association with kind of a vibe coded app.
• 58:18
  David: So you can vibe code an app and then you launch the app and then you launch
• 58:22
  David: a token with it. And it uses the token to gain virality for the app.
• 58:26
  David: You know, you attract some degens. If they play with your app and then they
• 58:29
  David: like your app, they might buy your token, which again has no material association
• 58:33
  David: with the app other than it's sharing the same brand.
• 58:35
  David: And so on one side of things, we have the AI agent part of this like crypto
• 58:41
  David: industry, which is downstream of virtuals, which is a token launchpad.
• 58:45
  David: And then there's the token launchpad side of things, which is the token engineering,
• 58:49
  David: the crypto degens, the speculative gambling that really turns people off to crypto.
• 58:53
  David: And on the spectrum of like, is this AI innovation or is this token launchpad
• 58:58
  David: financial engineering?
• 58:59
  David: Like, is this for the AI people or is this for the crypto degens?
• 59:03
  David: I'm placing this more on the crypto degens side of thing than any sort of like
• 59:08
  David: AI fundamentals. That's kind of my take about this.
• 59:11
  Ejaaz: Yeah, I think you're right. And I think it's important to just say that,
• 59:15
  Ejaaz: again, this is mostly about the money.
• 59:17
  Ejaaz: Most kind of media coverage on this or like updates that people get excited
• 59:21
  Ejaaz: about is the return multiple that they get, which is, of course,
• 59:24
  Ejaaz: one of the main reasons why people get in crypto. They want to make money,
• 59:27
  Ejaaz: but it's also kind of like it gets kind of like
• 59:30
  Ejaaz: boring or repetitive over time. And we want to kind of see some real pioneering
• 59:34
  Ejaaz: innovation that isn't just some kind of meme coin relationship.
• 59:37
  Ejaaz: Otherwise, you know, what are we doing here? We might as well just like talk
• 59:40
  Ejaaz: about gambling picks every week, right?
• 59:42
  Ejaaz: Moving on, talking about actually on the note of talking about some real fundamental
• 59:46
  Ejaaz: use, decentralized training or distributed training is actually an area where
• 59:52
  Ejaaz: Web3 or crypto innovators are actually
• 59:56
  Ejaaz: Bettering ai or pioneering ai so we've spoken about these companies that are
• 1:00:01
  Ejaaz: listed on this tweet here a bunch of times before but to give some context
• 1:00:06
  Ejaaz: In order to build these ai models you need a heck of a lot of compute and money
• 1:00:11
  Ejaaz: and data and it's very costly so some folks had a bright idea of thinking well
• 1:00:17
  Ejaaz: what if we do this in a decentralized distributed way
• 1:00:20
  Ejaaz: where we can aggregate compute from people's idle laptops, computers, cell phones,
• 1:00:25
  Ejaaz: and we can kind of like combine them into some kind of decentralized network
• 1:00:29
  Ejaaz: And use that compute to train models.
• 1:00:31
  Ejaaz: And so a bunch of teams were like, okay, cool, that sounds like a good idea.
• 1:00:34
  Ejaaz: But it was much easier said than done, because the way you need to network and
• 1:00:38
  Ejaaz: connect these different GPU servers and laptops and stuff is not an easy task at all.
• 1:00:44
  Ejaaz: In fact, by the end of 2022, the best minds at Google could only put together
• 1:00:50
  Ejaaz: a 400 million parameter model.
• 1:00:52
  Ejaaz: Fast forward to today and Jensen announced that they're using RL or reinforcement
• 1:00:58
  Ejaaz: learning to train a 72 billion parameter model, which I know isn't as big as 1.2 trillion, Josh.
• 1:01:04
  Ejaaz: I know it's not as big as like the DeepSeek team, but it shows massive improvement.
• 1:01:11
  Ejaaz: And this is being led by Web3 folks.
• 1:01:12
  Ejaaz: So kind of just not really going down the list, but just giving you an overlay here.
• 1:01:17
  Ejaaz: Teams like Jensen, teams like Prime Intellect, teams like News Research are
• 1:01:21
  Ejaaz: really pioneering open source distributed training.
• 1:01:24
  Ejaaz: And I don't think it's something that should be overlooked because if executed
• 1:01:28
  Ejaaz: well, these blockchain networks could be worth hundreds of billions one day.
• 1:01:31
  Ejaaz: And I'm not exaggerating because what cost or value would you give to networks
• 1:01:36
  Ejaaz: that train some of the best local or frontier models of our generation,
• 1:01:40
  Ejaaz: right? That can be co-owned by multiple people so i thought that was that was really cool
• 1:01:45
  David: Yeah, the idea of having a decentralized alternative to the very centralized
• 1:01:50
  David: powerhouses, open AIs, et cetera, is like attractive to me, just like kind of
• 1:01:54
  David: like on a redundancy perspective.
• 1:01:56
  David: Humanity at least has like a plan B, but I'm not sure if they can go toe to
• 1:02:01
  David: toe with the big boys. Josh, what's your take?
• 1:02:03
  Josh: Yeah, I agree. I'm still trying to wrap my head around it.
• 1:02:06
  Josh: I'm trying to get excited about it. I think it's an exciting development that
• 1:02:09
  Josh: is early. And again, like the rate of acceleration seems like it's going fairly well.
• 1:02:13
  Josh: The the initial take is that all of the largest companies in the world are trying
• 1:02:18
  Josh: to just gather all the information in the world and distill it into a model
• 1:02:22
  Josh: and we're seeing that closed source but we're also seeing that in open source
• 1:02:24
  Josh: with with meta and now meta has up to two trillion parameters on the next model
• 1:02:29
  Josh: and that's going to be open source
• 1:02:30
  Josh: open weight and i would imagine if people want to build something interesting
• 1:02:34
  Josh: from that they could take it and then distill that massive knowledge base down
• 1:02:37
  Josh: however they would like to kind of customize it into distilled models
• 1:02:40
  Josh: So I'm trying to just understand the difference between
• 1:02:44
  Josh: that, like taking an open source model and distilling it into whatever you want
• 1:02:47
  Josh: based on all this knowledge they've done the hard work to collecting versus
• 1:02:50
  Josh: going to collect your own
• 1:02:52
  Josh: data and use it in a decentralized way to build your own ground up model instead
• 1:02:57
  Josh: of kind of taking a top down approach.
• 1:03:00
  Ejaaz: Yeah, I completely agree. I think what this is ultimately going to result in
• 1:03:05
  Ejaaz: is niche use case, kind of bleeding edge or frontier edge models that serve
• 1:03:14
  Ejaaz: a small percentage of people, right? Maybe it's people that care about privacy
• 1:03:17
  Ejaaz: of data, or maybe it's people that care about using a certain amount of encrypted,
• 1:03:23
  Ejaaz: personalized information that they don't want to share with anyone, again,
• 1:03:27
  Ejaaz: from a safety aspect, to train a model that serves a particular function.
• 1:03:31
  Ejaaz: It comes back to what we were saying earlier. Like, if you don't have a good
• 1:03:34
  Ejaaz: enough product that comes from this AI thing, you're not going to make it.
• 1:03:38
  Ejaaz: And this isn't just unique to decentralized Web3 folk. This is also very prevalent
• 1:03:44
  Ejaaz: to the top dogs at the top. Like, we're talking about Microsoft, right?
• 1:03:48
  Ejaaz: What are their unique models beyond just getting direct access to OpenAI model weights, right?
• 1:03:53
  Ejaaz: If you don't have access to that, then you're kind of like nothing unless you
• 1:03:55
  Ejaaz: build an amazing product. So it's going to come down to the product.
• 1:03:58
  Ejaaz: It's going to come down to the founders.
• 1:04:00
  Ejaaz: I am excited that open source is still pushing really hard at this.
• 1:04:04
  Ejaaz: And I'm more than confident that some good kind of app program or software will
• 1:04:09
  Ejaaz: eventually come from this.
• 1:04:10
  Ejaaz: But you're in a very competitive mode and competitive industry that is being
• 1:04:14
  Ejaaz: funded equally so on the centralized side.
• 1:04:17
  Ejaaz: And at this point, you can't even argue that monopolies aren't open sourcing
• 1:04:20
  Ejaaz: because the folks at Meta, the folks at DeepSeek are all doing it anyway.
• 1:04:24
  Ejaaz: So let's see how this pans out and let's keep our heads in the game.
• 1:04:28
  David: Josh and Jaws, it's another week, a lot of news to cover. I'm sure in seven
• 1:04:32
  David: days, there'll be a lot more news to cover. And I look forward to talking to you guys then.
• 1:04:35
  Ejaaz: Yes, sir.
• 1:04:36
  Josh: Awesome. Another great week.
• 1:04:38
  David: Limitless. This is the Limitless podcast. Brand new feed. If you have not subscribed
• 1:04:42
  David: to the podcast or subscribed to the YouTube, please do so.
• 1:04:45
  David: If you are listening to this on podcast, please make sure to give us a five-star
• 1:04:48
  David: review so we can get this podcast up to the top of the charts.
• 1:04:51
  David: We think that this can be.
• 1:04:54
  David: And, you know, already is approaching in some of the best AI news-related content
• 1:04:59
  David: that you can find. If there are other AI podcasts out there that you guys listen
• 1:05:02
  David: to, let us know because we want to beat them.
• 1:05:04
  David: So yeah, give us a five-star review if you're on YouTube. Subscribe to us,
• 1:05:07
  David: like the video so we can grow this channel. We really appreciate you guys all
• 1:05:11
  David: here with us. And once again, we will see you guys in seven days.