0:00
You're watching TVP.
Search this video
0:04
>> Microsoft's blockbuster earnings last week. They blew out earnings. It was very exciting.
0:07
They cemented its status as one of the biggest winners in the artificial intelligence boom. We knew this.
0:13
Satcha had carved out um just a massive amount of territory.
0:15
GitHub co-pilot first real major breakthrough product that was monetizing on top of GT GPT 3. 5.
0:22
Great product obviously led by Nat Friedman when he was there.
0:25
Um, >> last we heard it was at >> around half a billion in AR 2 months ago, so it's probably in the billions now.
0:34
>> Yeah, probably bigger.
0:34
Uh, and then the crazy OpenAI deal, they got in very early.
0:38
They they have this massive revenue share. They have an ownership.
0:41
They get a copy of any software that OpenAI writes basically or acquires. That's kind of crazy.
0:46
And so >> Satia did seemingly one of the best deals of all time for Microsoft >> potentially. Yeah, I think so.
0:52
Oh, it's hard to I mean, it's an awkward situation now and it's and it's a really hard decision.
1:00
>> Sam is, you know, renegotiating the deal. >> Yep.
1:02
>> And Ben Thompson was noodling on this like should they take the money now or should they play it more like a venture style bet?
1:08
What is the role of the CEO of a $4 trillion public company where the shareholders have different expectations?
1:14
They're not seeing if you hold Microsoft stock, you're not feeling like you're an LP in a venture fund. >> Yeah.
1:21
So if if the CEO comes to you and says, "Hey, look, we're taking the cash flow now, we're going to dividend some of this out, we're selling down the position, we're we're we're we're thinking strategically about this as opposed to just we want the highest multiple.
1:34
We want to ring the gong on the deal."
1:36
Um that could be reasonable.
1:38
So obviously we're we're we're tracking where that goes.
1:40
But >> um they also have Clippy generational precursor to potentially all the AI agents, the original AI agent Clippy potentially making >> the original super intelligence.
1:51
original super intelligence and very >> the first time they had to move the goalpost, right?
1:56
>> And this is why we're worried about getting paperclipipped because Clippy will become too strong.
1:58
I mean, with the power of OpenAI and Microsoft Azure, anything's possible there.
2:03
Uh, so anyway, outside the AI race, Microsoft is minting money from corporate customers, spending on regular technology, long a sweet spot for the company.
2:12
Many companies are shifting from buying their own IT equipment to renting it from Microsoft through its cloud computing service.
2:16
They are also renting more standardisssue computing stuff, hard drives for data storage for example, to support their AI efforts.
2:24
And so this is the key uh stat from the earnings call that the that the Wall Street Journal is highlighting and then we'll kind of dig into this number and what it means because there's a lot of different explanations for what could what could be going on.
2:35
But Microsoft and the CFO and CEO didn't necessarily give all the context that we'd like to set this definitively, but I think there's some really good theories floating out there.
2:45
Uh the quote from the Wall Street Journal is a large chunk of the recent strong growth in Microsoft's cloud business called Azure stems from that.
2:54
More than half of Azure's 33% revenue jump in the company's March quarter came from nonAI services.
3:00
While the company didn't give a comparable breakdown, a comparable breakdown of the cloud unit's 39% growth in its June quarter.
3:07
It said that the core infrastructure business, Microsoft's lingo for its nonAI cloud business was the driver.
3:14
Jordy >> uh massive win for enterprise SAS.
3:18
Just good oldfashioned SAS.
3:21
We had some we had some theories. We had some not SAS.
3:26
So it's people migrating from onrem to >> it's what's called infrastructure as a service.
3:31
So there's software as a service.
3:33
That's when you go and get teams or you go and get a subscription to Excel in the cloud or outlook in the cloud.
3:41
That's >> I was saying I was saying more more traditionally someone else building. Yes. Yes.
3:45
Using using it as as infrastructure as a service.
3:47
Then there's also platform as a service.
3:50
That's like Heroku on AWS where you go and you deploy an app.
3:53
You could think of maybe even like a replet as like a platform as a service almost where they're hosting you, but they're not just providing you the raw infrastructure.
4:04
>> Azure's Azure's uh core infrastructure business is essentially infrastructure as a service.
4:08
IAS is the term and that means oh you want some CPUs and some hard drives and some Ethernet cables and moving stuff around data transfer.
4:18
>> Had an interesting thesis offline earlier.
4:20
you comped to people in the internet era buying a computer.
4:22
They hear about the internet.
4:24
They're like, "Hey, I think this might be a thing. I should get a computer.
4:27
They just buy a computer."
4:29
And now you could see something, you know, where companies say, "Hey, this AI thing might be big.
4:36
We should get on the cloud." >> Yeah. Yeah.
4:39
You want to set yourself up for it.
4:40
for it. And I think if you have a whole bunch of data in some sort of on-prem, you know, you have a data center for all of your data in a bunch of hard drives and you have CPUs that can do the do different workloads and data workloads
4:53
and maybe you're using some SAS on top of that, but you realize that you're never going to be in a position to buy a 100,000 H100s and you're going to wind up being a leaser of that for some small >> using other people's application layer products. Yeah. And so the integration Yeah.
5:10
And so the integration that comes from being in the Azure ecosystem that could be a driver. There's a few others.
5:16
Um, when I think about the the core the the AI Azure services, I think of that almost as you know it is SAS like if they're if they're if you go to Azure and you say I'd like to, you know, put my credit card down and I want to be able to use the GPT4 API and I also want to be able to use Llama 3 and I also want to be be able to use DeepSeek and I want to be able to call all these APIs within my within my product.
5:41
Um, I almost think of that as tokens as a service.
5:44
Like it is SAS, but it's something else.
5:46
And I think these token factories, I think this idea of how much revenue are you generating from your token generation business is really what we're talking about when we when we talk about Azure's, you know, core AI products um versus the infrastructure.
6:01
But there's a bunch of interesting wrinkles that could be going on within the classic, you know, core infrastructure business.
6:10
So this of course is uh virtual machines.
6:14
You just want a Linux box with a CPU to host a website.
6:17
That's something that you do on Azure storage networking.
6:22
Okay, I want to store all my data.
6:22
You could go to AWS and store in S3.
6:23
You could go to um Google store in BigQuery.
6:29
You could also go to Azure and fire up any sort of storage database or just you know raw uh raw hard drives.
6:34
Uh and then networking moving stuff around.
6:36
So this is the infrastructure as a service versus tokens as a service.
6:39
their higher higher level AI APIs.
6:41
Um, but so GPUs are f so the question is like why is their infrastructure as a service growing faster than their tokens as a service product?
6:51
You would think that Microsoft is going to their enterprise clients and saying like you need to build AI.
6:58
You need to bring AI into your products and we have all the best APIs so just buy tokens from us.
7:03
And you would think that that would be the boom.
7:07
Well, couldn't the other factor here be that they are massively supply constrained on the GPU side and they have this, you know, multi- tens of billions or hundreds of billions of dollars of backlog that they can't fulfill.
7:20
>> Meanwhile, they had a you, you know, more kind of like predictability on the the traditional data center cloud side that they were able to scale up to. Yep.
7:28
>> So, that feels like a potentially like a pretty big driver here. >> Yes, definitely. >> But still shocking.
7:34
And so and so as companies come out and they say, "Okay, we are we are scaling our our you know hardware and software foot our technology footprint broadly um going into this AI era.
7:47
We're excited about this stuff."
7:48
Well, we're also going to need more databases.
7:50
We're going to need to put more data in those databases.
7:53
We're going to need more CPU workloads.
7:54
We're going to need more of everything.
7:55
And Microsoft's like, "Yeah, of course we can definitely get you a whole bunch more hard drives and a whole bunch more CPUs.
8:01
We're not constrained on that at all."
8:03
And the capex is keeping up.
8:03
So they're able to service that.
8:05
Um there's also an interesting thing where tons of AI stuff can technically be happening inside the core infrastructure bucket.
8:16
You just don't necessarily know what what's in there.
8:18
So, um, some examples are like if you're, let's say you're a pure AI company or you're doing or you have a new AI workload, you could go to Azure and say, "Hey, I'm going to do a whole b I'm going to do my own AI thing, but I need a ton of storage for data because I'm going to be training on it.
8:35
I'm going to need a bunch of networking to move that data around when I do a training run."
8:38
So, that could be driving core infra up.
8:40
And then also um if a bank hypothetically spins up a huge cluster of H100 GPU virtual machines to fine-tune an open source model like Meta Lama 3, this would show up as core infrastructure, not Azure AI services.
8:55
And so that's like textbook AI boom, but it's just happening in in the wrong bucket.
9:00
And then I also saw a post that potentially uh chat GBPT counts as Azure core infrastructure because they're not they're not serving chat GPT through the Azure AI API.
9:13
It's not this like snake eating its tail or a Boros.
9:16
It's like OpenAI just came to give us a whole bunch itself ends up becoming pretty misleading because and again this is this was >> probably you know a lot of people were reading into uh AWS's uh growth you know the reports that you know Andy or the comments Andy had given on uh AWS last week and the big thing that AWS is missing is having a chat GBT building on top of >> AWS right there's dominant consumer product at least at that scale.
9:49
product at least at that scale. uh chat GPT I think the as of this morning somebody was estimating getting to a billion uh weekly activives this year which um again those types of products don't uh you know the the the power law is like extreme right so
10:08
>> yeah so if you I mean I I believe anthropic is pretty tightly hitched to Amazon and I think the next uh big cluster from Anthropic will be powered by Amazon for the most part and so they're getting there certainly on uh if they're building all the all the infrastructure for
10:24
>> again that would show up on on >> token generation side like more of the the the >> no so if Anthropic goes to AWS and says we want you to build a huge data center to serve cloud code for us that's going on that's going on in infrastructure not not actually APIs but when you go to AWS
10:43
and you're just some random company and you say uh I need a database and I need some storage and I need uh a web server and I also need a bunch of tokens from whatever model you can serve me and they're like we got claude and then you're like yeah let's pull the claw tokens into my app. >> It's a good model sir.
10:56
>> It's a good model sir.
10:57
>> That's token as a service. What?
10:59
>> It's a good model sir.
11:00
>> It's a good model sir. Exactly.
11:01
>> Um the one thing that stood out to me uh uh that has stood out to me across this year with Microsoft is they've done u more layoffs this year than the past three years before that combined. So 2022, 2023 and 2024. That's crazy.
11:14
So this just shows the level that Satia is operating at is like the the company has been on a tear this year, you know, performing exceptionally well and he's still thinking about how do we get more and more fit.
11:29
>> Um and uh >> yeah, >> so so to be clear, literally every piece of Microsoft business is growing and at a very a very solid clip.
11:37
So, Microsoft 365 commercial cloud business which houses uh remotely accessed versions of Word, Excel, other productivity software that grew at 16% from a year earlier.
11:50
So, that's I mean that's like it's not the most insane growth rate but that's still crazy because you think about like who doesn't have Excel that needs it?
11:57
Like who are these people who are like you know what 2025 is the year that my company's getting on Excel? We're doing it >> well.
12:04
It's just crazy when you compare it to uh AWS growing at 19%.
12:06
Obviously very different scales, but totally >> you would think you don't want to be in the same ballpark as uh >> and so that was the news of AWS if you missed it. Uh they beat earnings.
12:17
They did very well, but they weren't growing as fast as the others uh the other cloud platforms, Google and and uh Microsoft.
12:27
And so this the Amazon stock traded down.
12:29
down. Um, and I mean the the narrative around AWS is different because Microsoft has open AI, Microsoft research and has GitHub copilot and is like really moving things forward in the AI world and Satcha is seen as someone
12:44
who goes on the Door Cash podcast and talks about AI and it's clearly like really on top of it where and and obviously Google has Gemini and a million different products and and uh and strategies around rolling that out and staying on the frontier. they have a
12:57
they have a frontier lab internally and Amazon's just not there either on the partnership side or on the core um like training frontier lab side and so it's a little bit >> it's a little bit of both.
13:10
Uh anyway, let me tell you about Figma.
13:12
Think bigger, build faster.
13:13
Figma helps design and development teams build great products together.
13:17
We had a great week in New York celebrating Figma and the IPO.
13:21
Uh stocks been up, stocks been down.
13:23
uh wild ride, but we are still very happy to be partnered with Dylan Field and the Figma team.
13:27
And I'm super excited to see what happens next because >> truly it was Thursday was such such an incredible day just getting the uh if you didn't get a chance to listen, we talked with every the seed lead investor uh all the way through of course Andrew Reid who led the C. Yep.
13:44
um and then capped it off with Dylan and and also got to speak with uh Lynn Martin, president of the New York Stock Exchange, as well as Chris, the uh CTO of Figma.
13:55
So, uh really incredible day and uh it's just um very proud of the Figma team. >> Yeah, it was awesome.
14:01
Um in one sense, investors might prefer to see AI businesses driving growth.
14:06
That after all is what has driven the company's valuation through the roof.
14:10
But tech companies stocks arguably hinge on too too much on AI to the extent that they can keep increasing other revenue streams.
14:17
They are on more solid financial ground.
14:18
Of course, none of that means anything if all of the growth coming from open AI.
14:22
But at the same time, does anyone really think Chat GBT is uh is Yahoo anymore?
14:26
Like like you know they're they're generating what a billion dollars a month in revenue at this point. Everyone uses the app.
14:33
it's installed everywhere like that that that token demand is not going anywhere.
14:39
That infrastructure demand is not going anywhere.
14:42
So >> on the other side, Amazon is down uh >> just over 9% since Thursday earnings.
14:48
Um and then this morning uh unrelated, they announced that they're shutting down Wondery, the podcast studio they acquired in late 2020.
15:02
>> Why did uh the market's way up? Wow. NASDAQ's up 1. 8%.
15:04
8% today after a brief selloff on Friday. >> Good news. >> Bare market is over.
15:11
>> Yeah, we were so over but we're already so back. It's fantastic.
15:14
>> Love when that happens.
15:15
>> Uh markets go up, markets do go down, but the march of technological progress.
15:20
The arrow points in but one direction. >> That's right.
15:24
>> And it's March is relentless.
15:24
Uh there is another silver lining for Microsoft.
15:28
NonAI sales can be substantially more lucrative than AI ones.
15:31
non-AI gross margins within Azure were around 73%. Wow.
15:37
That compares to that compares to 30 to 40% gross margin for AI.
15:39
He estimated because of the huge cost of setting up AI infrastructure. That makes sense.
15:44
You get more margin on just a bunch of CPUs and databases that you've harnessed and built everything around.
15:51
Also, you know, you still have that interesting dynamic where it seems like all of the cloud the hyperscalers like don't really compete on price because they're all pretty comparable, but they seem to all have really >> dynamic.
16:01
That's what it seems like.
16:03
>> Uh I'm not exactly sure if there's something else that's more fundamental going on.
16:07
Um, >> and the Coca-Cola dynamic is like I don't know if we uh it's hard to tell what conversations were off air or on air, but but John last week uh forget when was describing how uh you would think that Coca-Cola or Pepsi would decide to get aggressive on price to try to gain market share and get people to switch.
16:28
switch. But ultimately that would just lead to a price war with both companies you know massively eroding their their uh you know uh >> margins >> margins and then uh >> all the RC cola coded then >> um every yeah every >> so if you have a fierce competitor
16:46
consider entering uh >> an unspoken gentleman's maybe a gentleman's agreement >> it's not it's not even a gentleman's agreement it's not it is unspoken but it is a nature it is a natural game theoretic Nash equilibrium like it is
17:02
the natural state of things that both sides understand that to go to war would be mutually assured destruction and so they don't even need to talk about it and so instead they both agree to keep prices where they are and instead compete on marketing compete on
17:17
marketing really >> yeah I mean they don't form they don't reformulate that often they mostly compete on on on marketing um and that allows them to have this like continually compounding business And that's why it's in the the Warren Buffett portfolio. Coca-Cola. He's been Coca-Cola.
17:30
He's been in there for a long time.
17:32
And Pepsi's been doing well. >> And he was a DAU.
17:34
Of course, >> he was a DAU of Coke.
17:37
>> Still, >> Diet Coke or Coca-Cola? >> I think Coca-Cola. >> No, Diet Coke.
17:42
>> Oh, he's a Diet Coke guy.
17:43
>> I feel like he was a Coca-Cola guy. >> Warren. >> Yeah, look that up. I want to know.
17:48
Anyway, I'll keep reading from this Wall Street Journal report.
17:49
Luckily for uh Microsoft, demand for lucrative nonAI services appears to be reasonably strong.
17:56
measures of broad IT spending. There we go. >> You know why?
18:00
>> Because of regular Coca-Cola. >> You know why?
18:01
It's because it has corn syrup in it.
18:04
>> It's literally corn grown from mother nature, from the earth >> and then syrup.
18:08
It's what you put on pancakes.
18:10
Like it's the most wholesome combination of of foods you could imagine. Maze.
18:16
This is something that's been grown in America for generations, for for centuries.
18:21
Corn is so popular in Nebraska, right? It's grown everywhere.
18:25
It's called corn and the syrup that you put on pancakes.
18:28
It It's the most American, most wholesome ingredient.
18:32
It's not this like refined sugar, this crazy stuff from somewhere else. No, >> it's American. >> American corn syrup. >> There's nothing. It's Lindy corn syrup. >> Corn syrup. >> Yeah. >> Seed oil.
18:42
Seed oil haters are are in disbelief.
18:46
>> Shambles that that we need to return to corn syrup.
18:48
None of this uh none of whatever's in this Coke Zero. >> Yeah.
18:54
Your grandpa was was drinking uh corn syrup.
18:58
>> Oh, and it's too good for you cuz you read a couple posts on axe. com.
19:02
Think you understand something better than corn? Delicious corn. Corn on the cob.
19:07
Something you have on a at a barbecue.
19:09
Oh, now it's too good for you.
19:12
Can't possibly have corn. What's next? No apple pie. >> What's next?
19:15
No rotisserie chicken turkey on Thanksgiving.
19:19
He's going to listen to the show for the first time today and and just go raging for you promoting corn syrup consumption.
19:28
>> It's it's uh it's as American as apple pie and Warren Buffett knows best. He's doing great.
19:35
>> And and of one study sort of a Brian Johnson. >> Yeah.
19:37
>> He's sort of the Brian Johnson.
19:38
>> He's the original don't die >> of Yeah.
19:39
And he's been doing fantastically on that front. >> Yep. >> He's great.
19:43
Anyway, uh measures of broad IT spending were fairly muted at the start of the year as companies pondered the impact of Donald Trump's tariffs and concerns bubbled about the health of the global economy.
19:57
Attitudes appear to have improved somewhat in the second quarter.
19:59
Though a UBS survey of cloud computing customers in July showed a clear improvement in tone about spending, most were moving forward with e efforts to migrate computing work to the cloud.
20:10
They're like, "This internet thing is real. It's real.
20:14
>> We got to put the data in the cloud.
20:17
>> We held back as long as we could >> we could, but it's 2025. We have no more excuses.
20:22
The tariffs, it's come and gone. Now's the time.
20:26
>> We were resisting the 21st century, but we're a quarter of the way through. >> Put the data online. >> It's not going away.
20:32
>> So, let's let's use the computer in the online in the cloud.
20:36
>> Put the docs in the cloud."
20:36
But in the cloud, just put the docs in the cloud.
20:40
Put the fries in the bag anyway.
20:40
And put your compliance process on Vanta.
20:45
Automate compliance, manage risk, improve trust.
20:46
Vanta's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation.
20:52
Whether you're pursuing your first framework or managing a complex program uh in the longer term, there is little question that cloud computing is going to grow in ways that play to Microsoft strengths.
21:01
Its rivals, mainly Amazon.
21:01
com and Google, are growing quickly, too, but don't have all of Microsoft's broad corporate software offerings that enhance its cloud footprint, even outside of AI.
21:11
Amazon on Thursday said its cloud unit grew at 17 and a half% in the June quarter, disappointing investors and forcing CEO Andy Jasse to answer questions uh to answer uh questions about Azure's outperformance.
21:25
Why are you getting beaten?
21:25
You created this category.
21:26
You created this product. Why is Satcha Nadella? You are the cloud.
21:31
You are the the cowboy of the cloud and you're getting you're getting put out to pasture.
21:38
Recent quarterly earnings in Azure's favor were really just moments in time.
21:43
He said >> that chart, John, >> the uh Oh, yeah. Wow.
21:44
That That's worse than I thought.
21:48
I wasn't sure where we had that pulled.
21:49
I mean, >> it would have been hard to predict 5 years ago that we'd be sitting here with with Microsoft at 4 trillion and and uh >> where's Amazon at? >> I think 2. 2 trillion >> 2. 2 >> 2. 27 >> 2. 27. Okay. So, almost double.
22:10
>> Still magnificent, but >> still magnificent, but you got to keep fighting. Got to keep fighting.
22:13
Uh the company's stock fell 8% Friday and looks like it still sliding down even more.
22:20
>> The question for Microsoft's investors then is less about the its prospects than its valuation.
22:24
The company's stock is up nearly 40% since the beginning of April, pushing its forward price earnings multiple above 33.
22:30
That's a bit richer than Amazon and a large margin above Google, which is trading at a multiple of roughly 18 times.
22:36
That should be easier for in investors to digest because while Microsoft's AI growth is real, it is far from the only thing going right at the software giant because they got Excel, they got core infrastructure, they got AI APIs, they got tokens as a service, infrastructure as a service and software as a service, they got the royal flush.
22:58
It's going well over at Microsoft.
23:00
They won't need to call McKenzie, but maybe maybe Amazon. com will.
23:07
My question is, AI a sustaining innovation for Disney or is it a disruptive innovation?
23:14
Where will Disney be in 10 years in the medium term?
23:18
Obviously, you're going to be able to generate a lot of AI slop.
23:20
You might be able to infringe on their a on their IP.
23:25
They might get paid by Google VO when you generate a Mickey Mouse AI slop edit.
23:32
They might get a little a couple pennies, but will it be good or bad for them?
23:37
>> Children's, you know, I could see them making a product that allows you to make a, you know, book or a story for your kid that actually uh, you know, and they get they get some type of revenue.
23:47
they should get revenue and I think that they will through the courts especially because as like big companies like you're not it's not this crazy oh there's like a bunch of kids doing random things like they never had to go after the the street artist on Venice Beach that would draw a picture of Mickey Mouse and sell it to you for 20 bucks.
24:09
They that was never material to their business.
24:11
They had to go after Napster.
24:12
They had to go after to Venice Beach.
24:14
I got them to draw me as Mickey Mouse. Really?
24:17
And then I performed a citizen's arrest because I respect IP. I'm an IP respector. >> Yes. Yes. Yes.
24:24
Citizens arrest is underrated. >> Citizen's arrest. >> Citizens arrest.
24:27
You're going to jail, buddy. >> This is for Bob. >> This is for Bob Iger. Yeah. So, big question.
24:33
This feels like a moment where you want to be in founder mode.
24:35
Bob Iger is one of the greatest CEOs of all time.
24:38
How will he navigate this?
24:38
The Wall Street Journal says, "Is it still he's not the founder?
24:43
The founder died in 1966.
24:44
Walt Disney >> claiming that he died still.
24:47
>> He might be able to turn it on.
24:47
Satcha Nadella certainly was able to do it.
24:51
He's he's navigating the AI uh the AI shift flawlessly.
24:54
We'll see what happens with Bob Iger and Disney.
24:57
Uh Wall Street Journal says, "Is it still Disney magic if it's AI?"
25:01
The stakes are especially high for the studio caught between how to use artificial intelligence in the film making process and how to protect its famed characters against it.
25:08
So, there's a little anecdote that we'll kick it off with.
25:12
When Disney began working on its new live-action version of its hit cartoon Moana, executives started to ponder whether they should clone its star Dwayne Johnson.
25:20
The actor was reprising his role in the movie as Maui uh a barrel chested God. Have you seen Moana? No, you have not. Of course, we know this.
25:31
For certain days on set, Disney had a plan in place that wouldn't require Johnson to be there at all under the plan they devised.
25:38
Johnson's similarly buff cousin, uh, Tanoa Reed, who is 6'3", 250 pounds, would fill in his body double for a small number of shots.
25:47
Disney would work with AI company Metaphysic to create deep fakes of Johnson's face that would be layered on top of Reed's performance in the footage, a digital twin, essentially digital double that effectively allowed Johnson to be in two places at once.
26:01
Obviously, that's better.
26:02
>> Calls up his cousin, you want a job? >> I need you.
26:06
>> Yeah, hit the gym, buddy.
26:06
Yeah, >> better be better be jacked. >> Get on a cycle.
26:10
>> Um, but yeah, I mean these these uh these movie schedules are famously tight 3 months in and out crazy schedules and then you move on to the next one.
26:21
There's that famous Henry Caval story where he filmed Superman, wrapped, moved on to another movie where he had to grow out a mustache.
26:27
He grew out a mustache and a beard and then they said, "Hey, we got to do some re-shoots.
26:31
You got to come back to Superman."
26:32
And he came back, but he couldn't shave his mustache and beard.
26:34
So they had to change it in in CGI and it looked terrible.
26:36
Probably not a problem now with with deep fakes.
26:41
That's actually a good use of AI and uh something that probably shouldn't be very controversial, but obviously everything in AI is controversial right now.
26:47
Um but we will continue with Disney.
26:49
They say what happened next was evidence that Hollywood's must discuss much discussed much feared AI revolution won't be an overnight robot takeover.
26:59
Johnson approved the plan, but the use of a new technology had Disney attorneys hammering out details over how it could be deployed, what security precautions would protect the data, and a host of other concerns.
27:08
They worried that the studio ultimately couldn't claim ownership over every uh that the studio couldn't ultimately claim ownership over every element in the film if AI generated parts were in there.
27:18
So, if there's AI training data from a DreamWorks film in there and they use the DreamWorks training data to make a Disney film, even if it looks like Dwayne the Rock Johnson, DreamWorks might come knocking and say, "Hey, give us a royalty."
27:32
I think that's the risk, but the lawyers are having fun.
27:34
Maybe full employment for lawyers over at Disney.
27:38
Clearly, uh Disney and Metaphysics spent 18 months negotiating on and off over the terms of the contract to work on the digital double, but none of the footage will be in the final film when it's released next summer. They went and shot it.
27:47
A deep fake Dwayne Johnson is just one part of a broader technological earthquake hitting Hollywood.
27:53
Studios are scrambling to figure out how simultaneously they can use AI in the film making process and how to protect themselves against it.
27:59
Is it sustaining or disruptive or both?
28:01
Can't be both, but we'll see.
28:04
Uh while executives see a future where the technology shaves tens of millions of dollars off a movie's budget, they are grappling with a present uh with a present present filled with legal uncertainty, fan backlash, and a weariness toward embracing tools that some in Silicon Valley view as their next century replacement.
28:22
And if you are trying to sell an AI tool into Hollywood, you got to get on Adio.
28:29
Customer relationship magic.
28:29
Adio is the AI native CRM that builds, scales, and grows your company to the next level.
28:34
and you can get started for free.
28:34
So, the Academy of Motion Picture Arts and Sciences is surveying members on how they use the technology.
28:39
Studio chiefs are shutting down efforts to experiment for fear of angering show business unions on the eve of another contract negotiation.
28:48
And no studio stands to gain or lose more in the outcome than Disney, the home of Donald Duck, Bell, Buzz Lightyear, Stitch, and countless others, which has turned out some of the most valuable and protected creative works in over the past century.
29:00
So my take on this um two so two years ago I was hanging out with the founder of a very large generative AI image generation company um and he was telling me that by 2025 anyone with a laptop and an internet connection could generate a full Hollywood movie about anything they want with a single prompt.
29:21
And it was a hilarious conversation because uh we were on a Zoom call and his internet wasn't working.
29:26
And it was the classic example of like the technology is amazing, but we got a lot of stuff to iron out.
29:32
Um, so anyway, extremely aggressive timeline, but obviously things are going to change for Hollywood.
29:37
So my question is, will Disney benefit?
29:39
It feels like a moment to be in founder mode, but Walt Disney died in 1966.
29:43
So basically, everyone believes that Meta will benefit from AI even if they missed the train on owning the next dominant consumer tech platform.
29:50
But if Disney got really AI pill, what would that look like?
29:52
I don't think they need to train their own foundation model.
29:55
Just like they don't need to train their own they don't need to build their own cinema cameras.
29:59
They can just use IMAX when they the time calls for IMAX.
30:03
They can use Blender when the time calls for Blender.
30:05
They can use Houdini when the time calls for when the when when the the shot calls for some uh highle VFX.
30:11
Uh but they do need to rethink how they structure their business and negotiate with unions and underwrite content.
30:17
they might need to go more risk on.
30:18
Not just from a brand risk position, but taking more smaller bets.
30:23
We're in this weird barbell world where everything seems like it's either a hundred bucks and it's shot on an iPhone.
30:28
It's a viral like Tik Tok or it's a hund00 million blockbuster with like $50 million of VFX.
30:33
Actually, that's kind of a low number.
30:35
It's usually like $300 million production with $150 million VFX and then no one sees it and it's a flop, but then they they hit every once in a while and they're great when they're good.
30:44
But it's this weird like venture style betting at the high end, but there's nothing in the middle.
30:49
And maybe if that's like the death of the art house film, but I'm just wondering if in the age of AI like maybe there's this interim step where Disney ladders down a little bit and gives like 10 filmmakers $10 million each and says, "Hey, you're still required to deliver a 90minute full film, but you're doing it for 10 mil and it's not quite Blair Witch level production. One notch up. You got to be creative. really bullish.
31:15
Yeah, I'm really bullish on this idea that uh the you know historically like a TV show would film a pilot episode and they would use that to get uh the budget to shoot an entire season and you can imagine now uh you could you know even for film you can just make you know make the trailer ahead of time with AI.
31:37
The other advantage that I think Disney has that's very real just going in and and why they're just broadly seem to be positioned very well here is that I think that I think that broadly like content customization will probably take off because you could Disney can make a film now and then you could make millions of different variations of it that become interactive with like the underlying fan.
32:02
So like imagine >> you're watching Moana but like your kid is in the film is a character in the film and you can now do that at scale right and how much more would you pay as a parent to have something like that?
32:15
>> I have a funny story about this >> but last thing I'll say so um >> customization and just like democratizing like like basically making making being able to make variations of films y >> I think is going to be big.
32:27
I think uh just the the time it's the same way you it's it's so difficult to make a new luxury brand. >> Yep.
32:35
>> Uh it's so difficult to create it's easy to create IP.
32:37
It's hard extremely difficult and time intensive to create valuable IP.
32:42
And Disney's advantage is they have this sort of like 360 and that they can make a film and they can bring it to Disneyland, they can bring it to a cruise.
32:52
They can create physical products around it.
32:54
products around it. And so they develop IP in a way that that new entrance are not you know they don't have the benefit of like having a Disneyland where they can make new experiences that that increase the value of that IP right so >> I think um >> so two things one if I were to go back
33:12
one of my favorite Disney properties is Star Wars a new hope the very first film if I went back and was like let's customize that for me I don't know that I would make any changes like do I really want a scene where Han Solo breaks the fourth wall and says like, "Hey, John, like I'm about to go, you know, save Luke at the Death Star." Like, that doesn't improve the product
33:31
Like, that doesn't improve the product for me.
33:33
I actually like that it's just the vision of George Lucas. So, I don't know.
33:39
>> But here's like >> customization being better for me.
33:40
I don't know what I would change, >> but here's here's an example of how to make like a magical experience for a kid.
33:47
Like imagine after a movie ends >> a kid could interact with a character and ask it questions like about the story or have a conversation with them.
34:00
And that's what I'm talking about like bringing that IP to life. >> Totally.
34:03
Right now that exists and I was obsessed with this when I was a kid.
34:05
I would I would uh I would watch Star Wars and then I would read the the the books that in the expanded universe and I remember even having books that were just like encyclopedias of every single ship.
34:17
This uh you could read into this a little bit more, but uh we'll leave that where it is.
34:20
Um but I but I would learn every single every single ship.
34:23
This is what a star destroyer does and it would have all this backlog and stuff.
34:27
And so yes, I agree with you. That's very cool.
34:30
you finish the movie and then you can just interview Luke about how this how does the lightsaber actually work and he can talk about the kyber crystals. That would be very cool.
34:38
On the flip side, uh I had a very funny experience.
34:40
Uh I was watching a horror film in high school with a couple friends and I grew up in Pasadena and this horror film just happened to take place in Pasadena.
34:50
Like that's where they set the film because Pasadena, California is just like a place where you set films. It's just a real place.
34:58
So, if you've seen Kill Bill, which I know you haven't, uh, but Kill Bill, uh, by Quinton Tarantino, uh, there's a scene where, uh, where Uma Thurman just shows up and it says at the bottom, Pasadena, California, because like that's where the character went.
35:12
It exists in the real world.
35:13
But this horror film that took place in Pasadena was terrifying because I was watching it and I was like, this is happening here and it's night and it's dark outside and like I now I'm so much more immersed.
35:27
And so I was thinking back then that you could use like the IP address of the of a of a connected I think we were watching it on like a PS3 like a DVD player.
35:35
Like you could you could use the DVD player to dynamically change the location of this of the establishing shot.
35:42
So you're like this whole horror film is going to take place inside of like one house where there's like a monster in the house and it's going to be like the usual like they're upstairs, they're downstairs, there's blood, there's you know someone's running and chasing.
35:53
Like it could be any town, USA, but they usually they usually establish as like this is happening in Amityville, this is happening in, you know, some random town, uh, Lake Placid, right?
36:01
But they could easily just dynamically change that with a few establishing shots to just show you, okay, it's happening in Malibu, right?
36:10
And now it's a lot scarier for you.
36:12
So, I think there's something interesting there, but again, it has to be the work of like an >> aur. It's Yeah.
36:18
It's not something you can copy and paste either because it works for some shows but then others like the place is so obviously the place that it would throw you off as a viewer. >> Exactly.
36:28
And so I do think that the that the fully AI generated fully custom content that will exist but it will exist on uh independent third party platforms.
36:36
It won't be the domain of Disney.
36:37
Disney. This will be something where if you in the future if you really want to see like you know AI generated stories about surfing in Malibu like there will be an endless stream of those and you will be able to go and and and
36:51
experience that particular content and you can already kind of experience that because there's probably some Instagram person who makes really great content about surfing in Malibu and if you follow them you get that vibe and that might be what you're into. Um, and you
37:04
Um, and you can kind of like and and it's handled just by the great democratization of create creativity. Yeah.
37:10
Uh, and we should talk to Samir about this later. He's coming on the show.
37:13
Uh, so we can talk about the future of of of uh content creation and whatnot.
37:18
>> On the question of should Apple um make a huge uh acquisition.
37:22
Um Tim Cook actually addressed it in the earnings call.
37:28
call. uh he said something like we've acquired around seven companies this year and that's companies from all walks of life not all AI oriented we're very open to M&A that accelerates our roadmap we're not stuck on a certain size of company although the ones that we've
37:44
acquired uh thus far this year are small in nature and he said like we're acquiring one every few weeks which is not quite seven per year I mean I guess that's like one a month but a few weeks sure um so he's doing deals but he's not looking at anything massive And I think that actually makes sense. My question
37:59
My question was if you were the CEO of Apple and you had a hundred billion burning a hole in your pocket, what are the other things that you could do with that money? >> Formula 1 team.
38:14
>> You could buy every team and the division and every track and the and all the sponsorship.
38:21
>> I made a post about this jokingly. >> It's crazy. It's so much money. It's so much money.
38:25
>> I made a post about this jokingly earlier this year.
38:26
I still think it'd be cool for Apple to just be an F1 in the way that Red Bull is.
38:31
>> I think that'd be great team. >> No, no, I agree.
38:32
Yeah, I think that'd be great.
38:36
>> The one thing culturally is like Apple's brand is is very premium and very high-end.
38:40
And um >> I don't know if they could handle a few seasons.
38:46
>> This was the early bare case for Apple TV.
38:49
TV. I was talking to uh someone a very successful Hollywood producer and he was saying that um the nature of the of Hollywood is like venture capital like it is a hood driven business if you don't accept that you're going to have massive flops and egg on your face and embarrassing moments you will not make
39:11
it and just like any venture capitalist who is like I only want to make investments where they won't blow up you're not going to make it right Yeah, Mark Andre had a had a there was a clip from a episode he did recently talking about how you lose sleep over the companies you don't invest in, not the ones you invest in and don't work. >> Exactly. Exactly. And so, so you have to >> Exactly. Exactly.
39:30
And so, so you have to be risk.
39:32
You have to be risk in Hollywood as you do in venture.
39:34
And so the the question was like, can Apple deal with spending 50 or hundred million on some show that completely flops and everyone's like this is a disaster?
39:45
Um, and they've navigated really well.
39:47
I think that they have had some flops, but they've kind of tucked those behind the scenes.
39:52
Whereas Netflix has started to get more kind of, you know, people talk trash about what was that Red Red Notice or they they did some massive movie with The Rock that kind of flopped.
40:01
They've done a few of these big movies that haven't done that well and and it's sort of like, oh, like, you know, they're they're they're not, you know, God's gift to, you know, producing.
40:10
Um, but it doesn't matter because overall Netflix is doing very well and all it takes is like one Squid Game to like carry the whole quarter basically just like Inventure.
40:21
Um, or like one, you know, you sign the Friends deal and then or The Office and people are watching that like 20 years later.
40:26
Um, so >> and um, apparently so F1 uh has surpassed half a billion dollars in box office earnings as of last week.
40:35
I mean, it's a fantastic >> That's an example of a uh of a home run, >> but it's not like they're doing that once a quarter by any means.
40:44
>> Oh, and John Xley's in the chat. Shout out to John Xley.
40:46
Uh, unfortunately, I didn't get a chance to connect with him at our party in New York, but he did attend.
40:51
You got to chat with him.
40:51
A lot of the team got to chat with him.
40:52
So, thank you for all you do, John.
40:54
We really appreciate you.
40:56
The entire team is giving you a round of applause.
40:59
General >> because general of the you've done a ton of a ton of work for us and we really appreciate it.
41:04
Um anyway, if you had a hundred billion dollars burning your burning burning a hole in your pocket at Apple, what would you do?
41:10
I have a bunch of crazy ideas. >> Go through your list.
41:15
>> So, uh most people would just say, "Hey, you're going to benefit from AI because you are the window into all technology and it doesn't matter if it's if it's generative video or or text like people will be consuming it on devices.
41:26
You sell devices, you'll be fine."
41:28
Uh just like search did not destroy Apple.
41:30
It actually made Apple stronger because people search on their iPhones.
41:35
Um, so most people would just say, "Hey, return it to the shareholders."
41:38
Apple's already done that.
41:40
They've returned over a trillion dollars to shareholders in the past 10 years.
41:44
Those are kind of rough numbers, but it's basically that, which is an insane amount of money to return to shareholders.
41:48
Um, but Samsung is worth 330 billion.
41:51
They could buy the entire company, lever it up with twothirds debt, and they would just own both sides of the smartphone market.
41:59
Then probably extremely anti-competitive.
42:01
I think Lena Khan would have a conion fit.
42:02
But uh funny concept funny concept.
42:05
Um >> the other uh the other the other idea is massively expand the retail footprint.
42:12
So Apple right now I was surprised by this.
42:15
How many retail stores do you think they they have worldwide?
42:19
>> I don't know like 599 or 601.
42:23
>> Did you look Did I tell you this morning? It's 535.
42:27
Um so it's not that many.
42:27
It's mostly in the tier one cities.
42:29
There's usually only like one in every city.
42:31
They don't really They don't take the Starbucks strategy where they put them across the street from each other. Yeah.
42:35
But for hundred billion, they absolutely could.
42:38
Hundred billion is enough to open 6,000 new Apple stores. So they would be 6,600.
42:44
So they would be up around 7,000 stores, which for reference is as many Subways as there are and as many CVS's as there are and as many 7-Elevens as there are in America.
42:56
And so everywhere you see a Subway, you could see an Apple store.
43:01
And I feel like this would be a big upgrade for America.
43:02
I feel like if you just walked around every American city where tier one, tier two, tier three city, you just see a beautiful sheet of glass and it's an Apple store on every corner in America.
43:12
That's >> They could also acquire every firearms uh store in the country and turn them into Apple stores.
43:22
>> They could do a take private of Figma.
43:24
Think bigger, build faster.
43:24
Figma bu helps design and development teams build great products together.
43:28
Get started for free at figma. com.
43:31
>> I'm still buzzing from uh Thursday.
43:33
>> It was a fantastic day. >> Fantastic. Tuning in.
43:36
Anybody that is tuning in to the show for the first time today is going to be hearing our ideas for Apple and just thinking these are some of the worst people have ever heard, >> but they're definitely fun.
43:48
>> Okay, so speaking of glass from the Apple stores, they could buy Corning, which makes Gorilla Glass.
43:52
It's a $55 billion company.
43:54
Then with the money left over, they could buy Interdigital, which owns all the patents on 5G and 6G.
44:01
So they could basically patent troll everyone else in the industry.
44:04
>> Just get extremely anti-competitive.
44:06
>> They could also with the money that is left over.
44:08
We're still in the first like let's vertically integrate this company mode.
44:11
They could also buy 10 rare earth or cobalt mines.
44:14
So they own the supply chain there.
44:17
And then they could also build four battery gigafactories and lock up the sapphire crystal supply chain that's used on the uh on the the iPhone camera.
44:27
So they could just completely own their entire supply chain for 100 billion.
44:30
Uh and and that's just one year.
44:33
One year and then they you know the proposal was do this every year.
44:36
The other thing that was interesting about Tay Kim's post is if you if you listen to the earnings call, Tim Cook and the team were very clear that they don't even want to finance like data center development themselves.
44:49
They're still working with these sort of like third party lenders to like capitalize them.
44:54
>> Um and so again, Tim Cook uh is the king of >> uh efficiency. >> Yeah.
45:00
He knows his business and he's Should we get into some of >> I have one more. >> One more.
45:05
So, OpenAI, they Aqua hired Johnny IV.
45:09
They're coming out with something that's competitive.
45:11
It's going to be some some sleepless nights at Apple when that thing drops, right?
45:14
It's going to be stressful because like maybe May maybe it doesn't go super well, but like, you know, it's a serious threat.
45:21
Open is a serious company.
45:21
They have a lot of customers and they got Johnny, your former goat dei designer ready to drop, you know, >> former Apple employees.
45:31
amazing amazing team over at in in hardware at OpenAI.
45:34
So the question is like how do you fire back with your Apple and you have a hundred billion to spend. Here's what you do.
45:40
The COGS on an iPhone are less because they have a margin.
45:46
So you take that hundred billion, you buy 200 million iPhones and you send a brand new iPhone to 200 million Americans.
45:53
And you're just like, "Oh, how much are you charging Johnny IV and OpenAI?"
45:59
Well, ours is free this year.
46:02
We're doing a free iPhone for the entire year.
46:05
>> They should just do something like that where it's like your 11th iPhone in a row >> is free. >> Is free.
46:11
>> They'd probably have 200 million people with free iPhones and they're just like, "Yeah, we're actually we want to keep you in this ecosystem.
46:16
We like our services revenue.
46:18
We're excited to be a platform. >> We're excited.
46:20
We're just giving it back. We're just giving back." Can you imagine? back.
46:24
>> I think that would also be extremely anti-competitive.
46:25
I don't even know if there's a regulatory framework to stop that from happening.
46:29
But I mean, price competition is a real thing.
46:31
And like if you if you discover that your competitor is under capitalized and can't can't win a capital war, you could potentially cut cost cut it cut prices so dramatically that it's like, yeah, the other phone's like cool, but you know, this one's free.
46:49
I guess I'm going to stay in in the blue bubble world. next earnings call.
46:54
>> All right, we're going to buy all of the mines we depend on and we're going to make the iPhone free.
46:58
Stock just eats like 80%. >> He's like, I'm petty. I'm petty.
47:03
And you know what, Johnny?
47:05
You know, I'm not going to let you win.
47:09
>> I would rather die than let you win.
47:12
>> Yeah, I would destroy myself.
47:12
Uh yeah, I've been reading a lot about cutting your nose to spite your face.
47:15
I'm pretty into the strategy.
47:19
Uh anyway, uh let me tell you about Vanta.
47:22
Automate compliance, manage risk, improve trust continuously.
47:25
Vant's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation.
47:32
Whether you're pursuing your first framework or managing a complex program, >> talk to the purple llamas. Vanta. com.
47:36
>> Anyway, uh so uh Ben Thompson's been noodling on uh on Apple and AI strategy.
47:44
He had a rough go because he I think he took some time off or shifted he shifted his posting schedule right when all the earnings dropped and there was all this crazy stuff going on.
47:53
So um he he's just writing about Apple earnings now.
47:56
But he has uh he still has some some good takeaway saying you know even if they do even if Apple does swoop in and buy Mstral for example uh the overall point holds Apple isn't going to compete with OpenAI Anthropic Meta X and Google on pursuing AI super intelligence like five extremely wellunded extremely serious teams.
48:18
uh they're gonna Apple's going to go their own way, center it on their devices and their direct relationship with their customers, at least as long as Cook is in charge.
48:26
And anyone that's saying that Cook uh needs to step aside, I think is uh is in for a world of hurt because he's been on a generational run and he's solving the most important problem for that company, which is not the application layer in artificial intelligence.
48:41
It's the supply chain and keeping the keeping the flows flow of iPhones >> making and selling phones >> making put the phones in the boxes. Tim, he's doing it. He's doing it.
48:50
He's really He's really He's really underrated.
48:55
He's been CEO longer than Steve Jobs was now.
48:59
>> The longest tenured CEO of Apple in history.
49:01
>> He better be bringing some of these AI researcher uh comp packages in as as comps when he gets in front of the comp committee at Apple. >> Yeah.
49:10
Hey, hey, >> that is true.
49:11
He should definitely be getting a bump.
49:13
Uh, if he does have researchers, I mean, he has he has AI researchers.
49:17
They are improving different functionality.
49:19
I feel like maybe I'm hallucinating this, but it feels like the texttospech engine has gotten better on my phone.
49:25
I listened to a semi- analysis article with the default Apple Safari text to speech and it was very listenable.
49:33
So, I think we're getting better there.
49:37
We have some breaking news about Sirius XM.
49:39
They are cancelling the Howard Stern show.
49:42
Can you call it a cancellation when the guy's 71 years old and he's been doing it for 20 years?
49:46
Uh they say it's no longer worth the investment.
49:50
They've been paying him $100 million a year. That is a huge salary.
49:52
And that's what three four times Colbear. Wow. That is that's big.
50:00
That's the power power of uh of radio.
50:02
power of, you know, the power law. He's done fantastically. So, congrats to him.
50:08
Howard, if you're looking for a new gig, you're welcome to come and hang out at the Ultra Dome.
50:11
You can turn table next to Tyler and we'll we'll we'll bounce ideas off you. >> Yeah, it's wild.
50:20
Uh I I really wonder where SiriusXM business goes.
50:26
>> I know I know where this goes.
50:26
Post AGI, you are going to listen to every how every hour of Howard Stern.
50:30
I know you've listened to Zero Hours.
50:32
Um, but there's probably 20,000 hours in the catalog or something like that of Howard Stern content.
50:39
You could listen to it from the beginning.
50:43
>> Do people listen to the back catalog ever? >> Absolutely not.
50:47
>> But Sirius XM is still a 7 billion company.
50:50
That's >> bigger than I would have thought. >> I wonder yeah.
50:53
Revenue like how much of that cost?
50:55
I mean, it makes sense to give him a huge slice of that.
50:56
It is a talent driven business and he could go to Spotify.
50:59
He could go somewhere else.
51:01
My question is like he is old.
51:01
Will he retire or will he do a podcast or do something independent? Is there news?
51:08
>> Now they uh >> what >> they apparently did how much?
51:13
>> Okay, you look that up while I tell you about wander. Find your happy place.
51:16
Book a wander with inspiring views. Hotel great amenities. Dreamy beds. Top tier cleaning. 24/7 concier service.
51:21
It's a vacation home but better. >> They did 8.
51:22
6 billion in 2024 revenue.
51:22
So they're trading at less than 1x revenue, which is says uh when you have a shrinking >> business, it is a rough place to be.
51:34
>> Not a growth, not a growth stock, I suppose.
51:35
Um >> yeah, I wonder how much of that is getting up by by talent.
51:39
Yeah, >> this is a trillion dollar stock if they have >> Yeah, I wonder how much I wonder what percent of SiriusXM content is uh like power loss celebrity host red high salary contract versus essentially programmatic or AI content because it's probably not even AI, but if you're just like there's a there's a station on SiriusXM that's just the Grateful Dead and it just plays them the whole time.
52:08
Like I don't think you need someone making a hundred million dollars to like randomly play Grateful Dead tracks.
52:11
Like there's probably >> some Grateful Dead fan who manages that and picks the songs and orders them.
52:17
But like that could essentially be pseudo random.
52:21
Um maybe one day you go through the the back catalog and then the new stuff and then you mix it up and then you play the hits or something.
52:28
I don't even know if they have hits in that I know it's kind of a jam band, but um uh I wonder I wonder of their of their like of their content of their tonnage. Is that the term?
52:40
Pat tonnage is like the the amount of content on the network.
52:42
Uh I wonder how much of it is is driven by these high high dollar deals.
52:47
And I wonder if they'll get a new a new host in the seat.
52:53
I wonder who this generation's Howard Stern is.
52:54
Maybe it's like Tim Dylan or something. Some irreverent comic.
52:58
The thing I think the thing is >> as content is on demand, fewer and fewer people are just turning on the radio or turning on the television and listening to whatever they whatever just happens to be playing.
53:11
And so if you take people that were subscribed to Howard Stern and we're like, you liked Howard Stern, we're now just going to play this other person through >> his channel, they're just going to be they're just going to ask what's going on here. >> Yeah.
53:25
I mean, certainly if you have a specific car that has Sirius XM and it doesn't have Spotify and Howard Stern's new show is on Spotify because it's a podcast, like you might stick around on Sirius.
53:38
M maybe they get someone new in the seat who's, you know, almost as good or can build a relationship, but it does seem like a challenge to fill that.
53:44
But also, you know, it's a lot of money to pay.
53:48
So, clearly it wasn't penciling out, so they had to move on.
53:51
Uh, Open AI has announced an opensource model.
53:55
They said they'd never do it.
53:56
Everyone's saying it's closed. It was in the name. How could you closed AI? Closed AI. They're not open AI. They're closed AI.
54:03
Everyone thought they were so clever.
54:05
They said they couldn't possibly open source a model. And they did. >> They did. >> They did it.
54:09
First open model in years, right? They used to have them.
54:16
>> Was Was GPT1 open source?
54:16
like when when was the last time that So, we're going to talk to Tyler about this in the studio today.
54:24
>> Was still open source and then they decided to close source something.
54:26
Uh it's tough that I I treat you like things that I could Google and I know that you just have to search it or chat GPT it. Um but uh good luck.
54:34
Get me the information on the the history of open AI's open-source models.
54:41
In the meantime, I will read this post from Sam Alman. GPTO OSS all lowercase.
54:45
He's still in lowercase mode, although he he uppercases the other sentences, but he he he threw in he likes lower case.
54:56
>> He's sending you messages. John, >> I'm seeing it.
54:58
I'm seeing coded messages in here.
55:01
Uh >> GPOSS is a big deal.
55:03
It's a state-of-the-art open weights reasoning model with strong real world performance comp comparable to 04 Mini that you can run locally on your own computer or phone with the smaller size.
55:14
We believe this is the best and most usable open model in the world.
55:20
We're excited to make this model the resilient the result of billions of dollars of research available to the world to get AI into the hands of the most people possible.
55:29
Hit that soundboard again. I want to keep it going.
55:31
We believe far more good than bad will come from it.
55:33
For example, GPTOSs 12B performs about as well as 03 on challenging health issues.
55:39
We've worked hard to mitigate the most serious safety issues, especially around biocurity.
55:46
GPTOSS models perform comparably to our frontier models on in on internal safety benchmarks.
55:51
We believe in individual empowerment.
55:54
Let's hear it for individual empowerment people.
55:57
>> I feel empowered the way that you're reading this, John.
55:58
Although we believe most people will want to use a conventional service, a convenience service like ChateBT, people should be able to directly control and modify their own AI when they need to and the privacy benefits are obvious.
56:10
As part of this, we are quite hopeful that this release will enable new kinds of research and the creation of new kinds of products.
56:19
We expect a meaningful uptick in the rate of innovation in our field and for many more people to do important work than were able to before.
56:27
OpenAI's mission into is to ensure AGI that it benefits all of humanity.
56:30
To that end, we're excited.
56:33
I think I cut this screenshot off.
56:35
We're excited for the world to be building on an open AI stack created in the United States based on democratic values available for I didn't screenshot the rest of that, but oh, here it is.
56:47
Available for free to all and for wide benefit.
56:52
>> So, I I imagine that that was the voice that was in his head as the author intended. Exactly. Exactly.
56:57
>> Will Depw has some more notes here. Uh some context.
57:00
So less than one year between 01 announced which was September of 2024 and we have an 03 level model opensourced that's runnable on consumer hardware. Wild progress.
57:09
You were highlighting earlier Sam initially had a poll.
57:14
Oh yes saying like you know should we should we what do you want?
57:16
Do you want an 03 level model or do you want something that you can run? >> He did both. >> And he did both. >> He did both. Wow. That was Yeah.
57:23
So, we were reading that.
57:26
So, uh it was what like six months ago or something that post that you that you shared. >> Yeah, I think so.
57:32
>> Uh so, I asked I asked Tyler to dig this poll up.
57:35
Sam Alman shared a post that uh it was a poll on X saying, "What do you want?
57:41
Do you want an 01 level model, reasoning model, a frontier model, open source, or do you want something you can run on your phone?"
57:48
And it was and it was neck andneck.
57:49
And people were clicking on the phone model.
57:51
They were like, "We want the phone model."
57:52
And then Dylan Patel quoted it and said, "Oh, you idiots don't vote for the phone model.
57:58
You can just still get a truly frontier model.
58:01
>> Let's get a frontier model."
58:01
And so, so Dylan Patel was very much like like you guys don't know what you want like like the phone models are available. That that will be easy.
58:10
The hard thing we got to twist OpenAI's arm to open source the reasoning model, the 03 level model.
58:15
And Sam Alman just said you wanted this or this.
58:19
It was a false dichotomy. You get both. >> You get both. >> You get both.
58:22
Yasine says, "All caps, I'm so sorry for doubting you, Sam Alman.
58:26
I'm so sorry for saying that you were the antichrist.
58:28
I didn't realize I didn't realize your plans were measured in centuries.
58:32
I hope you forgive me for everything.
58:35
I'm so glad you kept control of open AI. I'm so sorry. Please forgive me." >> That's remarkable.
58:40
Never talk down on the future first ballot hall of famer.
58:45
Apparently, the future leader of open source AI.
58:46
Will be very interesting to see how Meta responds.
58:48
Will they stay in the open- source game?
58:50
Also just how important is open source?
58:53
Is this a Linux scale opportunity?
58:55
Is this an Android scale opportunity?
58:57
Like is this a you know is this GitLab to GitHub?
59:04
Like what what's actually the long-term play? Certainly cool. Great for the community.
59:08
We're going to see fun experiments.
59:10
We're going to see interesting things done with this.
59:11
I think um do do we want to talk about some of the ideas that we had for building on top of this uh open source model fine-tuning?
59:21
I can just run through some I know there's one you didn't want to you didn't want to share because you thought it was too too too good too much alpha but what was what was I talking to you about Tyler? What were our ideas?
59:30
We had one that was uh >> well initially we weren't sure there was going to be reasoning right so we were going to add fake reasoning. >> Yeah.
59:38
knew it was going to be a model that gaslights you into thinking it's reasoning when in fact it's not reasoning.
59:44
So, >> so if you pull up if if you look at at the actual like you know reasoning UI, it's just UI.
59:50
It's not there's no actual reasoning. It's like a time delay.
59:52
It's a time delay and it's delay thinking really hard. Let me think about that.
59:58
>> What would this what would a super smart person think right now?
59:59
So we could actually still kind of do that because it's not a a multimodal uh model, but we could like, you know, we just add some kind of like totally blank encoder or something for the image >> and then it's like this is an interesting image.
1:00:13
>> Let me think about what's in it. >> Yeah.
1:00:15
One of the best images I've ever seen.
1:00:16
A lot of people are saying this this is one of the greatest images.
1:00:20
Thank you for sharing this PNG andor JPEG with me.
1:00:22
It just knows nothing about the about the image. It's great.
1:00:25
I I I I I'm seeing that this image is uh is is large in file size.
1:00:30
It's purely metadata based analysis.
1:00:33
Doesn't know anything about what's going on.
1:00:35
Like wow, this is a tough question.
1:00:37
I'm going to have to generate a lot of internal reasoning tokens to to answer this.
1:00:40
Test time inferences really is magical, isn't it?
1:00:44
Anyway, back to thinking.
1:00:44
Okay, I'm reasoning about this now. Yeah.
1:00:46
So, >> the other the other the other obvious opportunity, you know, people have been frustrated with uh syphency with the models and models being >> Well, not me.
1:00:56
I've been frustrated with the lack of sick. >> Yeah. You want more. You want more. I want more.
1:01:00
You want the model to gas you up.
1:01:02
But, you know, fine-tuning it to go, you know, over the top.
1:01:04
Somebody could release the Glazinator 3000, which which just tells the user, >> uh, you don't you don't just think you're goated, you are goated. I know this. You know this.
1:01:15
>> It's like, I'm asking you to solve IMO question six.
1:01:17
Why are you talking about whether or not I'm goated?
1:01:19
It's like it it just tells you like, you got this one.
1:01:23
>> The timing The timing here The timing here is interesting.
1:01:25
I mean, everybody's everybody's been banging the table saying, "Where is the American open-source AI or LLM leader?"
1:01:32
Uh it was reported yesterday by the misinformation that uh reflection reflection uh the misinformation reported that reflection uh one-year-old startup uh we've had uh the CEO on before he's in talks to raise 1 billion to develop open source models to compete with Deep Seek uh Meta and Mstrol uh and yeah all of the you know the the Chinese open source models have been on a tear.
1:02:04
Uh Quen released Quen Coder recently.
1:02:09
>> Yeah, people are into Quen.
1:02:10
>> Um or a new version of it that uh is getting good results.
1:02:13
>> So, it's important work.
1:02:13
We want >> So, one interesting thing that's kind of related to the like US China like race is like so so when Deep Seek first came out, everyone was like up in arms cuz they said like it's it was less than $5 million to train, right? Yeah.
1:02:26
>> So, um in the in the model card for for these models, um they actually reveal like the uh GPU hours it took.
1:02:30
which then you can like figure out how much it probably cost costed to um you know train.
1:02:38
>> Give me the number and tell me if I need to ring the gong for that.
1:02:41
>> So So for the big model it was probably around like 4 million. >> Oh. >> So just Yes.
1:02:46
But but the the small model was like 500k probably a little less >> but that's like 10x better.
1:02:51
The model is way better than deepse one. >> Okay. Okay.
1:02:55
So should I ring the gong?
1:02:57
>> I mean I think it's pretty >> I wanted I wanted I wanted to hear 400 million at least.
1:03:01
I was expecting nine $400 million nine dig.
1:03:04
They got a billion of revenue coming in every month.
1:03:08
But it's cool that they spent 10 days of revenue on the open source model. I No. No. It is very cool.
1:03:12
So So we're ringing the gong for for the elegance of the training run. >> Yeah. Efficiency. Efficiency.
1:03:22
>> Open AI on efficiency and cost model. Saving time and money.
1:03:26
putting money back in the hands of shareholders and not in the hands of open source developers I suppose.
1:03:32
Um but yeah I mean do you think that uh is this is this like a pre-training scaling wall narrative like like would you as someone who is a potential consumer of an open source um model?
1:03:43
Like do you want a $400 million open-source model?
1:03:47
Would that necessarily be better?
1:03:49
because it seems like they were able to distill it pretty well, get it to the frontier, not you know burn a ton of money on on training.
1:03:57
Also, the question is like if they're this is probably distilled from their other models that are much bigger.
1:04:02
So like you there is no world where you could just spend $4 million and get this level model without also have done having done the GPT4 run, the GPT4. 5 run, etc. Right. >> Yeah.
1:04:16
Maybe it's also just like even if they didn't distill it, they just have the knowledge of like how they did it, which is like arguably more valuable, right? >> Yeah. Yeah.
1:04:25
>> I think in terms of like open source, it's either you want to go like super super cheap and super small, which they kind of did here, or you go super big, like almost like the llama, like the it was the right. >> Yeah.
1:04:36
>> Um I think it would have been really cool to see that like literally like state-of-the-art, you know, level model.
1:04:41
You could also try and ask Claude, "Hey, build me a fit Frontier open source model. Don't make mistakes."
1:04:48
>> Yeah, we're gonna have someone from Claude on soon.
1:04:49
Um, this is funny because wait, you said 500K for the small model? >> Yeah, probably less.
1:04:56
>> So, that's the same amount of money that they're putting up for this uh challenge, the red teaming challenge. Have you seen this?
1:05:01
So, uh, to encourage researchers, developers, and enthusiasts, that's me, from from around the world to help, uh, identify novel safety issues, this challenge has a $500 pri $500,000 prize fund that will be awarded based on review from a panel of expert judges from OpenAI and other leading labs.
1:05:22
At the end of this challenge, we will publish a report and open source and evaluation data set based on the validated findings so that the wider community can immediately benefit.
1:05:30
So, if you can hack this thing, if you can get it to teach you how to develop a nuclear bomb or take over the world, you might have 500k in your pocket. Not bad. Go get some of that.
1:05:40
That's a seed seed seed round in 20 two 2012.
1:05:45
>> Welcome to the stream. How you doing, Mark?
1:05:48
>> Hey, what's happening? >> Great to see you.
1:05:51
>> Yeah, you too >> a lot.
1:05:52
It's uh it's a little bit of a slow news day, but uh exciting stuff with GPT opensource. It's not a slow August.
1:06:00
I will say >> it's not a slow August. We're glad.
1:06:00
We were just reflecting that we've taken exactly one day off this summer. That was July 4th.
1:06:05
And we're showing the Europeans how American companies work. >> American work.
1:06:12
>> We're setting an example and the and the and we have proof of work because we exist on the internet and you can see us live every day.
1:06:17
So, we're setting an example. How are you doing? How's your summer going? >> Fantastic. Going really well.
1:06:23
Um, so how long is it going to be until you guys put up avatars that make claims that you're working hard all through the summer when it turns out you're you're on the beach?
1:06:31
>> You might have caught us.
1:06:32
>> I think you'll know better than us as to when the technology gets there.
1:06:34
We we we've been demoing some of the stuff.
1:06:37
People have been doing a lot of deep fakes of us.
1:06:38
And fortunately, all of them have been clockable, so it doesn't feel like a brand risk, but they're getting closer and closer.
1:06:44
And I know that there's going to be a moment where we have to say, "Hey, that's actually using our name and likeness to endorse something that we don't necessarily endorse.
1:06:53
Can you please take that down?
1:06:54
So, we're we're we're approaching the the the touring test, the the uncanny valley.
1:06:59
We're escaping the valley.
1:07:00
>> I had a question like looking back over the, >> you know, maybe 10 or 10 or 15 years was was what moments did you feel like there just was not a lot of action happening because this summer is just the pace uh from so many different teams has been absolutely insane.
1:07:15
Everybody's like trying to keep up and it didn't used to feel that way, at least from my point of view.
1:07:23
So my my view on it always is there's like these there's this these disconnected, you know, kind of patterns or trends.
1:07:29
There's there's sort of the the sort of day-to-day phenomenon where like engineers show up every day and they make things a little bit better and then every once in a while, you know, you get a technical breakthrough or a new platform and and and and that process kind of this, you know, kind of sawtooth kind of up to the right kind of process kind of plays out over time kind of regardless of what else is happening in the world.
1:07:47
And so it it keeps happening through recessions and depressions and wars and like all kinds of crazy crazy crazy stuff that's happening.
1:07:53
But basically, you know, the the technology keeps getting better.
1:07:54
So there's there's kind of that curve and then and then there's the the sort of enthusiasm curve and and the and then the adoption curve, you know, which is basically like when do these things actually show up in the world and then by the way, when are people actually ready uh you know for for the new thing?
1:08:09
Um like if you talk to the people who worked on lang I'm sure you guys have talked to people who work on language models they will tell you that they were surprised the chat GPT was the breakthrough moment because they thought everybody already knew what these models could do for you know three years before
1:08:20
that and so they were you know they were shocked that it was the chatbot interface that that made the thing go um and so so there there's somewhat of a sort of arbitrary disconnection um between what's actually happening in the substance and then what what what what people are are seeing and feeling. And
1:08:31
And so it's just it's it's really hard to predict when these things pop.
1:08:35
But also if if you're in this dayto-day, it's it's really hard to tell um you know when things are going to be hot or not uh because it doesn't necessarily map to how much the techn is improving.
1:08:44
>> Yeah, we were just talking about that in the context of uh of Google's new world model.
1:08:49
It's this like generative video game that you can kind of move around in and it feels like Deep Mind is just absolutely crushing at the AI research frontier.
1:08:57
They have the best world model simulator that you can walk around in.
1:09:01
The question is like if they let another lab do the chat GPT thing and just get it out into the consumer three months earlier, they might wind up kind of chasing and trying to catch up if somebody actually figures out how to make it like a dominant consumer product.
1:09:17
Now in the enterprise it's more igopolistic, but consumer seems to be winner take all.
1:09:21
I guess the question is like how much value uh do you place right now in the AI race to just like moving fast, breaking things uh you know uh dealing having like the thick skin to deal with like the safety constraints and all of the different stuff.
1:09:36
Obviously not being irresponsible but just speeding up the organization as much as possible.
1:09:40
It feels like now is the time to really push on that. >> Yeah.
1:09:43
Well, first of all, I need to correct you.
1:09:45
It's it's moving fast and making things.
1:09:46
I don't know where I I I don't even know where that came from.
1:09:50
Yeah, I I I I I I have no idea what you >> never heard never heard of it.
1:09:54
>> I mean, didn't really break anything.
1:09:54
I I think that's a good point.
1:09:56
It really did just move fast and make things.
1:09:57
The first things it made were weird, but that was fine.
1:10:01
And it failed and it and it hallucinated a ton, but it didn't really break anything. I don't know.
1:10:05
>> Yeah, I Yeah, I believe I believe in this case total deaths attributable to uh to chat GPT are still zero. Zero.
1:10:12
>> So, not notwithstanding all of the notwithstanding all the all the catwalling, but um Yep. >> Yeah.
1:10:16
>> Yeah. So look, I think the AI industry in particular has a very acute version of the of the of the sort of challenge that you identified with and and you know and I don't say this negatively just an observation which is that there you know like in sort of a normal technology company you've kind of got engineers who make products and then you've got you know kind of sales people
1:10:32
or marketing people who sell them you know in in the AI companies you have this third tier of you know the quote unquote researchers >> right um and so you know which is which has worked out incredibly well I mean the researchers have done you know they've just done like amazing breakthroughs at these companies but you know the the the handoff you there's not necessarily clean handoff from the researchers to to the market. Um, and so
1:10:48
Um, and so it kind of raises this question of like okay like is there is are these companies therefore kind of three you know kind of three segment companies where they have research and then they have product development um and then and then they have go to market um and and I think that's a really open issue.
1:11:02
think that's a really open issue. I mean if you you know Google's kind of a case study of this you know you alluded to deep mind but even more broadly Google you know Google developed the transformer in 2017 um and then they
1:11:11
basically let it sit on the shelf right because it was a research project they didn't productize it they were very worried about you know from people I've talked to they were very worried about the you know brand issues and safety
1:11:19
issu you know kind of all these all these they had all these reasons to not productize it I talked to somebody senior who was there at the time who and I I asked them you know when when could you have had chat GPT with GPT4 level uh output um if you had just got you know
1:11:32
gone gone flat out starting in 2017 and they said by 2019, >> yeah, >> you know, they they already knew how to do it and then, you know, they've now caught up, but it took it took an extra five years, five years to catch up. >> Um, and and so I I think a lot of these
1:11:42
>> Um, and and so I I think a lot of these companies kind of had that challenge.
1:11:46
Elon as usual, of course, is is provoking this question as I'm sure you guys talked about, but you know, he he has now, you know, with XAI, he's now collapsed, you know, he's eliminated the distinction between research and product. Yeah.
1:11:57
>> Um and so, you know, of course, you know, he's pushing this as hard as he can.
1:11:59
And I think it's a it's a good question for a lot of these other companies kind of how hard they want to push on actually getting these things in fully productized form out to the market. >> Yeah. Yeah.
1:12:07
On on on Elon's uh like distinction, it feels like there is more research to be done, but it feels like we're we're entering like a new cycle of, you know, just focus on the engineering, focus on the deployment, the applications.
1:12:19
Let's get all this technology out into the world.
1:12:21
Let's reap all that benefit.
1:12:22
And yes, there will be a a different track of fundamental research that's happening somewhere, but it's really really hard to predict.
1:12:30
And so if you have something that's working, just double down and just go really aggressive on it.
1:12:34
Um I'm I'm wondering uh more on on that, but also on Apple strategy.
1:12:40
It feels like Apple's been um kind of like, you know, people have been maligning them for not for missing the AI opportunity.
1:12:47
and Tim Cook's just there on the earnings call being like, "Look, we acquired a couple small companies."
1:12:52
And this year, seven companies, but then uh it seems like they're taking more of like an American dynamism approach.
1:12:58
Like there was news today in the journal that they uh that they're investing $100 million in American manufacturing.
1:13:03
They're certainly doing stuff.
1:13:05
They're just not chasing the, you know, the the shiny tennis >> headline billion capex.
1:13:11
Um, so I'm wondering about your thoughts on on when you have a, you know, uh, when a when you have a platform, uh, how hard is it to resist chasing the new shiny object?
1:13:25
Is that the right move or are are there any other things that you think Apple should be, uh, you know, changing their strategy on? >> Yeah.
1:13:32
So look, Apple's always had this, you know, very clearly defined strategy that, you know, Steve Steen and Tim, you know, working together figured out a long time ago, which is, you know, they they I I forget the exact term, but it's it's something like basically they they they invest deeply into the core of what they do.
1:13:43
You know, they'll basically work internally on things for many years.
1:13:46
They they only actually release things when they feel like they're kind of fully baked. >> Yeah. >> Um right.
1:13:50
And and and so as a consequence, they have this thing where and and Tim says this, right?
1:13:53
You know, they're rarely first to market with new technologies.
1:13:56
You know, they're more often in the category of what, you know, Peter Peter Teal calls last to market.
1:14:01
you know, there, you know, they'll they'll they'll come out whatever 3 years later, whatever, 5 years later.
1:14:05
You know, there, you know, there were tablets for years before the iPad.
1:14:06
There were, you know, smartphones for years before the iPhone. >> Folding phones.
1:14:11
They're about to do a folding phone.
1:14:12
It's like 10 years into that technology.
1:14:14
I'm sure if they do the last mover, the last mover. >> Yeah. Yeah. Yeah. Sorry.
1:14:19
>> The last mover, I guess. Yeah.
1:14:19
Well, what I would say is like, look, that that clearly works if you're Apple, right?
1:14:23
Um, and so it it clearly works if you're Apple, but I would say there's a fine line between that strategy and just and simply becoming obsolete, right?
1:14:28
Um, and so the the problem is like if you're not Apple and you don't have all the other kind of super strengths and you know, kind of now the market position that Apple has, you know, do you really want to be a company, you know, if you're not Apple, do you really want to be a company that basically sits there and says, "Yeah, the world's moving and we're very deliberately not going to lean as hard as we can into it."
1:14:45
Um, and so I I I think there's a lot of survivorship bias in these kinds of strategy discussions where people look at the one company that's able to pull this off and they don't look at the 50 other companies that are in the graveyard, you know, because they, you know, because because they didn't adapt.
1:14:58
I mean, you know, all the other smartphone companies when the iPhone came out, they were like, "Oh, yeah, well, we could do Touch, too, right?
1:15:02
You know, we'll just, you know, we'll get to it, right?"
1:15:04
Um, and you know, you know, they're gone. >> Very bold.
1:15:08
I I remember it was like an iPhone knockoff. >> What do you think?
1:15:12
You know, right now people are are variety of, you know, shareholders are annoyed at Apple around their reaction to AI LLM.
1:15:19
John's annoyed around just like transcription generally, just like super basic stuff.
1:15:27
But it doesn't feel like the the uh core business is immediately threatened today.
1:15:32
It feels like it's still on the horizon around these sort of like you know eyewear based computing you know potentially net new devices that we're that that we'll see from uh you know companies like open AAI over time but where do you like like how how real is the threat you know this year uh versus 10 years from today and and kind of what's your framework? >> Yeah.
1:15:55
>> Yeah. Well, look, I mean I think the biggest ultimate danger I mean the biggest ultimate danger is very clear which is just like at what point do you not carry around a pane of glass in your hand that you know called a phone um you know because other things have superseded it and you know look everything you know everything becomes obsolete at point so there will there
1:16:10
will come sometime in the future when we're not you know carrying phones around and we'll we'll watch movies or people have phones and we'll be like yeah look at look at how primitive they were right because because we'll have moved on to other things and whether those things are eye based or you know uh you know other kinds of wearables or whether it's just kind of you know
1:16:24
computing happening in the environment um or just you know entirely voice based or you know who knows what it is but um you know there will come a time when that happens you know is that time three years from now because there's like some you know huge breakthrough you know from from some company that figures out the the product that obsoletes the phone right away or is that 20 years from now
1:16:41
because the phone is just you know such a standard platform for everything that we do in our lives and everything else you know kind of remains a peripheral to the phone I mean that you know that's you know that that's the game of elephants that's playing out there um you know obviously I think you know I think it's highly likely that we we'll have a phone for a very long time. >> Yeah. >> Yeah.
1:16:56
>> Having said that, it is it is exciting that there are companies that are going directly at that challenge.
1:16:58
Um, and you know, who whoever cracks the code on that will be the will be the next Apple.
1:17:03
And by the way, that that may in the fullness of time be Apple itself.
1:17:05
You know, they they may be the company that figures that out. >> Yeah.
1:17:09
I remember being at a board meeting at Andre and Horowitz maybe a decade ago or something.
1:17:13
And Chris Dixon showed me the hollow lens and I was like, "Okay, we're one year away from this being everywhere."
1:17:21
>> And and I feel like today I'm still in the like Yeah.
1:17:23
VR, it's definitely one year away.
1:17:25
The next Quest I'm gonna be wearing daily.
1:17:27
Um, and and it feels like we're always there, but it does feel like Apple did a lot of work on the on the fundamental uh, you know, pixel density of the resolution of the display.
1:17:37
And then Meta's been doing a ton of work on just getting it light and affordable.
1:17:41
Like, it feels closer than ever, but uh, you know, you you always got to wait until you see the churn numbers until you really call the game, right?
1:17:48
Well, here's the other thing, but you know, I think that's true, but you'd also say, you know, I'm on the on the meta board, so I'm kind of a a dog hunt on this one, but like the Meta Rayban glasses are a big hit. >> Oh, totally. >> Right.
1:17:59
Like like they're a big, you know, so I think we we now have a form factor that we know works, you know, for for eyebased wearables.
1:18:04
This, you know, there's not VR and then VR, you know, on top of that.
1:18:07
But, um, you know, just the, you know, the glasses and, you know, and then the the glasses with camera, you know, sort of integrated camera, integrated microphone, integrated speaker. Yep.
1:18:14
>> You know, that's a very interesting platform.
1:18:15
platform. Um you know the watch clearly works by the way which Apple of course you know is played a significant role in making happen you know that now sells in in in huge volume >> um you know so that's the second data point and then you know look I think these you know these these I I think some form of AI pin is going to work >> um I also think head you know headphones are going to get a lot more sophisticated which is already happening
1:18:33
>> um and and so you you know you do have these you know kind of data points coming out and then yeah look the the trillion dollar question ultimately is are these are these peripherals to the phone >> um you know which is what they are today
1:18:43
or are these replacements for the phone and it you We we yeah I would say we you know we have we allowed we I think we have a lot of invention coming both from new companies and from the incumbents who are going to try to figure that out. Yeah, I always think about the value of
1:18:52
Yeah, I always think about the value of like narrowing the aperture on these new technologies.
1:18:56
Like with with the the Meta Raybands, I feel like the fact that they aren't also trying to be a screen is actually a feature, not a bug.
1:19:03
And I always go back to the iPhone.
1:19:05
Like it was first and foremost a phone and people bought it because it could make calls and then it could make text messages and then it was an iPod.
1:19:11
But I do you disagree with that, please?
1:19:15
>> Well, you you guys I don't you guys might be too young.
1:19:17
The first iPhone actually was a bad phone. How so?
1:19:22
>> Don't you guys for the first two years I couldn't reliably make phone calls.
1:19:26
>> I I had I had like the third one and a friend had one, but I feel like it was still like people were carrying cell phones and that was the at least the expectation.
1:19:33
But yeah, I mean I guess you're right.
1:19:36
>> So for for the first it was a classic Apple story because the first for the first two years the thing couldn't make reliably make phone calls and then it turned out there was an issue with the antenna and with with how you held it and there was a famous email. >> Yeah.
1:19:46
You and you would and you would disconnect it.
1:19:48
You could basically brick the device >> based on how you held it.
1:19:51
And somebody emailed, this is when Steve would would respond to emails from random people.
1:19:55
And somebody emailed Steve saying, "If I, you know, hold the phone this way, it doesn't make phone calls."
1:19:57
And he's like, "Well, don't hold it that way." >> Yeah. >> Right.
1:20:01
>> So, so, so even there it was like, Yeah.
1:20:04
And people, you know, people forget it took like five years for the iPhone to find its footing.
1:20:06
It took like two years to get the And remember also the original iPhone didn't have it didn't have broadband uh data.
1:20:10
It it was on, it was on the the old 2G uh it was called the AT&T Edge Network.
1:20:15
So, it didn't have broadband data.
1:20:16
And then of course it didn't have an app store, right?
1:20:17
It was completely locked down, right?
1:20:20
>> So the challenge is the challenges for Apple now is that people are so used to perfection with the device that launching a product that isn't perfect >> like is embarrassing, right?
1:20:29
Like you look at the Vision Pro and it's like well the batteryy's big.
1:20:33
Steve would have hated this, right?
1:20:35
Like how he never would have shipped this.
1:20:37
and that being constrained and and not being able to innovate because you're tied to this like impossible standard of being on whatever generation 17 of the iPhone and perfecting every element is is a real challenge.
1:20:51
>> So I would say there's a correlary to that.
1:20:53
One of the things I've observed over the years is I I think technology products become obsolete at the precise moment they become perfect.
1:20:58
moment they become perfect. and and and to your point and what I mean I mean mean by perfect basically is like yeah it's like the perfect idealized complete product like it does everything you could possibly ever imagine everything a customer could imagine everything you as the technology developer can imagine it's absolutely perfect um and there's
1:21:14
there's been tons of examples of this o over the last 50 years um where it's like the absolute perfect permanent it seems to be the permanent version of that product and then it just turns out that's actually the point of obsolescence because it means creativity is no longer being applied right into that platform you're just like there's just nothing else to do just you're you're you're done, right? The product
1:21:30
The product has been realized and then and then the cycle is what happens to your point.
1:21:32
The cycle is other people come in with completely different approaches, completely different kinds of products that are broken and weird in all kinds of ways.
1:21:40
Um, you know, but but are fundamentally different.
1:21:42
And so, you know, that is one of the time honored traditions and, you know, one of the, you know, one of the, you know, things you could say about, you know, Tim is his, you know, his willingness to kind of break the mold of Apple only ships perfect products by, you know, be willing being willing to ship the, uh, you know, the vision pro.
1:21:54
Um, you know, you know, shows a level of determination to kind of stay in the innovation game like that, which I think is very positive. >> Yeah. Yeah. Yeah. Yeah. That's great.
1:22:02
Um, >> updated thinking on open source since we last talked.
1:22:06
Uh, there's there's a lot that's been >> OpenAI is an open source company. >> Yes. Open AAI is open again. Yes. >> Yeah. Yeah. Look, very encouraging.
1:22:13
You know, a year ago, I was very, you know, I was I was getting very distressed about open, you know, whether open source AI was going to be allowed. >> Uh, right.
1:22:20
It was even going to be legal.
1:22:22
And so, and I think, you know, we're basically through that at this point.
1:22:24
I was going to say we're through that in the US.
1:22:26
Um, you know, we we'll see about we'll see about the rest of the world.
1:22:30
>> Um, and then look, you know, the US China thing is obviously a big deal, but it, you know, I think it's been net positive for the world that China has been been so enthusiastic about open source AI coming out of China, >> uh, which has been great.
1:22:39
>> uh, which has been great. And then yeah look open leaning hard into this um you know and releasing what you know what they did is I is I think fantastic um both because of of what they released which is great but also just the fact that they are now you know willing to do that and then Elon reaffirmed overnight that he's going to you know open source you know start open sourcing previous
1:22:53
versions of Grock um and so yeah so we you know we we we seem to be we seem to be in the timeline where open source AI is going to happen um you know right now you know what you I think what you would say is it kind of lags the leading edge
1:23:05
proprietary implementations by you know 6 months or something like that Um but but I think that you know that's a good if that's the status quo that continues I think that would be a very good status quo. >> What are the rough edges that we need to
1:23:14
>> What are the rough edges that we need to kind of sand down when we're thinking about uh Chinese open source models specifically?
1:23:20
Uh is it we need to do some fine-tuning on top of them to add back free speech or do we need to watch for back doors?
1:23:27
Say it's phone and home if it runs into this specific thing like uh the Chinese open source thing it was remarkable because I feel like it really does accelerate the pace of innovation because everyone gets to see oh this is how reasoning works. I think that's great.
1:23:39
Uh, at the same time, it made me very it made me much more appreciative of AI safety research and capability research and actually being able to interpret what's going on and and say definitively this model is going to behave weird in this weird way.
1:23:50
Uh, like the Manurian candidate problem.
1:23:52
We haven't found any of that, but it certainly seems like something we'd want to keep an eye on.
1:23:56
But from your perspective, like what what are the what are the risks that we need to be aware of going into a world where China is really pushing hard into open source?
1:24:06
Yeah, there's two there's two and you identified them, but let's let's let's talk about both of them.
1:24:09
talk about both of them. Um, so the so the phone home thing is the is the easy one which is you can put a you know you can packet sniff you know a network and you can tell when the thing is doing that >> and you and and plus you can go you can go in the code and you can see when it's
1:24:20
doing that and so you can validate you can validate that that's either happening or not happening and I think that you know that's important um uh but you know I think people are going to people are going to are going to figure that out you can kind of get that problem practically. Yeah. Um the the Yeah.
1:24:31
Um the the the the bigger issue is um we we have this term in the field uh right now called open weights.
1:24:37
Um and um open weights is a loaded term.
1:24:40
Uh it uses the open term from open source.
1:24:43
But of course with open source the thing is you you can actually read the code.
1:24:47
>> Um you know with open weights you have you know just a giant file full of numbers as you said that you you can't really interpret.
1:24:53
really interpret. And then what you don't what you don't have what what most what most of the open source open weights models don't have including you know deepsek specifically what they don't have is they don't have open data right um or open corpus right so you you can't actually see the training data that went went into them um and of course you know most of the people
1:25:09
building models are kind of obscuring what that you know what that training data is in various ways um and and so when you get an openweight model you know the good news is the the the software source is open the good news is you can run it in your machine you can verify that it doesn't phone home but you don't actually know what's happening um inside the weights. So I I think that
1:25:23
um inside the weights. So I I think that that is going to be a bigger and bigger issue which is like okay how the thing behaves like yeah what what has it actually been trained to do um and what restrictions or directives has it been given in the training um you know that are embedded in the weights that that you need to be able to see um you know
1:25:40
this is I would say this is coming up as sort of I would say a global issue um you know which you know we worry about when these models come from China other countries worry when these models come from the US right which is right so one of one of the phrases you'll hear when you talk to people kind of outside the US is kind of this this phrase people are kicking which is not my weights, not my culture. >> Okay. >> Okay. >> Right. Right.
1:25:56
Or or by the way, for that matter, not my weights, not my laws. >> Yeah. >> Right.
1:26:00
Um which is like, okay, like what actually is this thing going to do? >> Right.
1:26:05
And to your point that Chinese models, for example, might, you know, never criticize, you know, communism or something.
1:26:10
I can tell you the American models have all kinds of constraints also. >> Uh right.
1:26:14
Implemented, you know, usually by a very specific kind of person uh in a very specific location in the US.
1:26:18
Um, and so, you know, I think this is a general issue and and we're going to have to see basically people's tolerance levels being willing to run open weights models where they don't fundamentally have access to the data.
1:26:30
And then correspondingly, I think what we'll see is more open source developers also doing uh open corpus open data so you can see what's actually in them. >> Yeah.
1:26:37
Um, obviously open source is very important in terms of just distributing intelligence broadly.
1:26:41
uh giving people the ability to run their own models and and really fine-tune them and have control.
1:26:48
Uh there's also the big push just to make frontier models and high capability models free.
1:26:53
One model is you charge for the premium, you give the free away. It's a premium model.
1:26:58
That's what we're seeing at most of the labs right now.
1:27:01
There's also this kind of uh spectre on the horizon of potentially putting ads in LLMs and what that would do to the world.
1:27:08
Jordy got in a little dust up with Mark Cuban on the timeline.
1:27:12
uh deciding whether or not it would be a net good to put advertising in LLMs.
1:27:18
What might happen that might be bad there?
1:27:19
What do you >> Yeah, my my point broadly was that ads have been uh an incredible way to make a variety of products and services online free and just saying like default just no ads would would potentially um you know be incredibly destructive.
1:27:35
Um but uh yeah, curious your framework. Yeah.
1:27:42
So, I should start by saying like whenever I personally use internet service, I always try to buy the premium version of it that doesn't have ads. Um, right.
1:27:49
And so, if if I can like live personally inside an ad for universe and pay for it, like that's great.
1:27:52
Um, and I I'll freely admit, you know, whatever level of, you know, hypocrisy or in congruence, you know, kind of kind of kind of results from that.
1:27:58
But, >> no, the point is choice. The point is choice.
1:28:02
>> Well, the point is the point is exactly what you said. It's affordability.
1:28:03
So the the the problem is if you really want to get to f if you want to get to a billion and then 5 billion people um you you you can't do that with a paid offering like it just at any sort of reasonable price point. It's just not possible.
1:28:15
Uh the you know global per capita GDP is not high enough for that.
1:28:18
People don't have enough income for that at least today.
1:28:20
Um and and so if if you want to get to, you know, if you want the if you want the Google search engine or the Facebook social app or the whatever AI, you know, Frontier AI model to be available to 5 billion people, uh for free, um you you need to have a business model.
1:28:35
You need to have an indirect business model and and and as is the obvious one.
1:28:38
Um and so I I do think if you know if if if you take some principle stand against ads, I think you unfortunately are also taking a stand against against against against broad access just in the way the world works today.
1:28:48
today. And then and then look the other the other really salient question is um you know the same question that the companies like Google and Facebook have been dealing with for a long time which is um are ads purely destructive or negative to the user experience or are they actually if done properly are they actually either neutral or even positive
1:29:03
right and and this was something that you know Google I think to their credit figured out very early which is you know a a wellargeted ad relevant point in time is is actually content like it actually enhances the the experience right the obvious case you're searching on a product there's an you can buy the product, you click to buy the product. That was actually a useful piece of
1:29:19
That was actually a useful piece of functionality.
1:29:20
Um, and so, you know, can you can you have ads or or or other things that are like ads or look like ads, you know, different different kinds of referrals, you know, mechanisms or whatever.
1:29:29
Can you have them in such a way that they're actually additive to the to the product experience?
1:29:32
Um, and you can just like with search and with social networking, you could imagine lots of examples of that.
1:29:38
>> People will, you know, people will, you know, they'll whiner in lots of different ways.
1:29:42
But I think, you know, I think that hasn't been a bad outcome overall.
1:29:45
Um and I think that uh I think it's entirely possible that that's what what happens with with these models as well. >> Yeah.
1:29:50
So uh kind of similar kind of question what what should be legal kind of trying to create legal frameworks on on a number of issues with AI.
1:30:00
Uh there's been a number of IP cases that have been working their way through the courts.
1:30:06
What can labs use to train models etc.
1:30:10
There's been some good outcomes recently.
1:30:11
Sam also was talking about how a lot of people are using AI as like a confidant, like a, you know, a friend, things like that.
1:30:18
And he mentioned that currently your chats are not privileged.
1:30:23
They can be used in in in a in a lawsuit or or other uh situations.
1:30:27
or or other uh situations. uh how how optimistic are you that our sort of legal system in the US can get some of these issues right where maybe it can't just be you know total free markets kind of lawless whatever goes >> you know so in the case of training data I think that there I mean there's a
1:30:47
bunch of these copyright you know kind of lawsuits happening right now there's you know the big New York Times open EI1 and there's you know been a bunch of others um I I think in that for that particular problem my guess is that problem ultimately has to be solved through legislation um it's It's it's ultimately a legislative question. The
1:30:59
The reason is because it goes to the nature of copyright law itself, you know, which which is legislation and and and of course, you know, the the the content industry is already claiming that of course, you know, using using copyrighted data to train, you know, without permission, without paying is is is sort of, you know, they believe illegal on his face, you know, due to violation copyright law.
1:31:17
The counter-argument to that, which, you know, which we believe is, well, it's not copying, right?
1:31:21
There's there's a distinction between training and copying, just like in the real world, there's a distinction between reading a book and copying the book, you know, as a person.
1:31:28
And so there there there's going to need I I think, you know, the courts are trying to grapple with that.
1:31:32
There's a whole bunch of cases.
1:31:32
There's jurisdictional questions.
1:31:34
You know, probably ultimately Congress is going to have to figure out a a um you know, figure out an answer on that.
1:31:39
And by the way, the president has kind of, you know, thrown down that gauntlet in his I think the speech he gave last week or two weeks ago.
1:31:45
Um you know, where he said that, you know, Washington probably needs to deal deal with that as an issue.
1:31:49
Um so that's one on the on the on the um on the on the privacy thing.
1:31:52
I I think that that one feels like it's a Supreme Court thing.
1:31:56
Um to me it feels like that's the kind of issue say the Supreme Court and the in other words like whether for example your trans transcripts are are considered your property and whether they're protected against you know warrantless search and seizure.
1:32:08
Um and and the observation I would make there is if you look at the march of technology over time.
1:32:11
So the the constitution has like very clear, you know, fourth, fifth amendments, you know, very specific rights around the, you know, the things that are yours, you know, such as, you know, your home, you know, being in your home, you know, by the way, the thoughts in your head, right?
1:32:24
Um, uh, you know, that the government can't just like come in and take.
1:32:28
They can't, you know, they can't just come in and search your house without a warrant.
1:32:31
>> You know, they can't like, you know, put you in a jail cell and beat you until you fess up.
1:32:34
Like, you know, there there are, you know, we we have constitutional protections against the government being able to basically, you know, take information, you know, fundamentally.
1:32:40
um uh you know as well as possessions.
1:32:43
Um and then basically what happens is every time there's a new technology that creates a new kind of sort of you know thing that you own, you know, thing that's yours, thing that you would consider to be private thing that you wouldn't want the government to be able to take without a warrant.
1:32:57
to take without a warrant. You know, out out of the gate, law enforcement agencies just naturally go try to get those things because they're ways to assault crimes and you know, they it feels like that that's a legal thing to And then basically the courts come in
1:33:07
later and they you know rule one way or the other and basically say no that that actually is also a thing that is protected against uh you know warless for example warless search um you know warless wiretapping and so I I feel like you know this is the latest of probably I don't know 20 of those over the last 100 years. Um and you know I don't know
1:33:22
Um and you know I don't know which way it'll go but I think it's it's going to be a key thing because as you know people are are already telling these models you know lots lots of things that they're you know that that are very personal.
1:33:33
>> Okay lightning round quick questions.
1:33:35
We're letting you get out of here in a couple minutes.
1:33:36
Um, we're in this age of spiky intelligence.
1:33:38
Models are great at some things and then terrible at others.
1:33:41
Where are you actually getting value out of AI right now?
1:33:43
Where is it falling down for you?
1:33:45
Where are you how are you using AI day-to-day? >> Yeah.
1:33:50
So, I I I have two kind of I don't know bar barbell approach.
1:33:52
Um, one is for for serious stuff.
1:33:54
I love the deep research capabilities. Yeah.
1:33:56
Um, and so and I'm doing this in a bunch of models, but like the ability to basically say I'm interested in this topic and then I just I just felt like write me a book and I, you know, I'm kind of hoping for the longest book I can get.
1:34:06
I always tell it like go longer, go longer, more sophisticated.
1:34:08
Um, you know, but the the leading edge models now they're getting up to like 30 page PDFs.
1:34:11
Um, you know, that are like completely well formulated, you know, basically long form long form essays.
1:34:16
Um, you know, with just like incredible richness and depth.
1:34:19
Um, and you know, if it's 30 pages today, I'm sort of crossing my fingers that it'll get to, you know, 300 pages coming up here in the next few years.
1:34:25
Um, and so I, you know, I'm able to basically have the thing generate enormous amounts of of reading material with just like I think incredible richness and depth and complexity.
1:34:32
Um, and then and then on the other side of the barbell is humor.
1:34:36
Um, and I've I've posted some of these to my my X feed over over the last couple years, but I think these models are already much funnier than people give them credit for >> really.
1:34:44
>> Um, I think I think they're they're actually quite highly entertaining.
1:34:47
Um, a while ago I post I had >> specific specific formats like we know >> chatting back and be Mark Andre you know that that formats >> take a dip in my pool in my office. >> They're really good.
1:35:01
So they're really good at green text.
1:35:02
U that works really well.
1:35:04
But the the the for some reason the ones I find hysterical are the I have it right screenplays um you know for like TV shows or or or plays or movies.
1:35:11
>> Um and um I I posted I had it right a new season of the HBO Silicon Valley you know set 10 years later. >> Yep.
1:35:17
Um, and I had it right like an entire I had it right like 10 10 scripts for an complete season.
1:35:21
And of course, I just said, you know, make it like Silicon Valley except, you know, it's happening at in 2021 at kind of peak woke.
1:35:27
Um, and I thought it was I think it's hyster, you know, I'll sit there at 2 in the morning just like laughing my ass off at how funny this thing is.
1:35:32
Um, and so I think these things are actually are actually already like extremely funny.
1:35:38
They're extremely entertaining when they're when they're uh, you know, when they're used in that way.
1:35:40
And I I I do I I do enjoy that a lot.
1:35:41
And I generate a lot of those uh that that I don't post.
1:35:47
>> Stay in the group chats >> is probably they're your property. >> Yeah.
1:35:51
Hopefully the fourth amendment holds on these. >> It's great.
1:35:54
I have one last question. >> Go for it.
1:35:56
And then I've got one more.
1:35:57
>> Uh how do you get a job as a venture capitalist in 2025?
1:36:01
>> Um so I I mean look the the best way the best way to do it is to have a a track record early as somebody who is like in the loop specifically on new product development.
1:36:08
Um and so somebody who you know be be like deeply in the trenches um at one of these new companies in one of these spaces.
1:36:13
Um you know participate in the creation of of a great new product uh and and and a great new company and you know really demonstrate that you know how to do that.
1:36:20
Um you know there's there you know there there are great VCs who have not done that but you know I think that is sort of a foundational skill set uh you know for working with the kinds of founders that that you want to work with who are going to you know are going to want you to have you know kind of very interesting things to say on that um is I think that you know still the the best way to do it. >> Yeah.
1:36:36
like feel the growth, be immerse yourself in the growth, the the the the aggressive growth environment and then you'll be able to identify it when you see it from afar. >> Yeah, that's right.
1:36:46
>> Last question for me state of M&A in your mind.
1:36:49
How are you advising you know companies uh where you're on the board or just the portfolio broadly around what they should expect now and and in the near future?
1:37:00
>> You mean in terms of whether you can get things approved or >> basically? Yeah. >> Yeah. Yes.
1:37:04
So look, approval still appro approval is not a slam dunk.
1:37:06
There was there was a you know there was a I just saw there was a medical device company this morning you know where the the acquisition was not allowed by the FTC.
1:37:13
So um you know look there is still scrutiny.
1:37:15
It's you know it's obviously a very different political regime in Washington.
1:37:17
But you know this is this is not an admin you know by by their own statements.
1:37:20
This is not an administration that believes in total as a fair and then it definitely wants to you know in in their view maintain a a very healthy level of market competition. Um, >> yeah.
1:37:30
How many do you expect do you expect certain companies to be negatively impacted by the Figma story, right?
1:37:37
You have this deal gets blocked, successful, you know, IPO, Lena Khan is taking a victory lap.
1:37:43
uh you know many people were responding and and joking saying you know someone Lena cuts off the arm of a pianist and they endure and can create a masterpiece and then >> and so I expect and then you look at the example with you know Roomba I I think it was where where Roomba had a deal with Amazon it was blocked and and the company has just been shambles ever since.
1:38:07
So my concern is that people look at Figma and say you should be independent. You just figure it out. >> Nothing can go wrong. >> Yes. Yeah.
1:38:14
Miscon was very disconcerting.
1:38:17
Um and and for exactly the reason you said which is survivorship bias.
1:38:21
>> Um right which is you you you pick the one that worked out and then you know it's the it's the airplane the red dots and the airplane mean you know you you you ignore the 50 that are in the ground uh that you've never heard of.
1:38:29
Um, and so that was very disconcerting because that, you know, it's sort of the central planning fallacy, which is like we make centrally planned economic decisions.
1:38:36
We have one example, you know, it's like in Europe it's like, yeah, well, the bottle caps actually don't fall off the bottle, right?
1:38:42
Like, you know, it works, >> right?
1:38:47
It's like, okay, but but do you want to live you want to live in an economic regime in which that, you know, the government is dictating bottle cap design?
1:38:54
The answer is clearly no.
1:38:56
>> The downside consequences.
1:38:56
Even even looking at that uh you know the Chinese model which is you know people can say they're picking winners but to get to maybe picking a winner you have this intense bloodbath of competition where you know teams need to rise to the top and sort of prove themselves before they get any of that real like you know meaningful state benefit. >> Yeah that's right.
1:39:19
And so you just you just yeah you just you just have this adverse selection survivorship bias thing where you just you don't pay attention to all the collateral damage.
1:39:26
So I I I I do think that mentality is like super super dangerous.
1:39:28
Um and so yeah, look, I think companies just have to be very thoughtful about this, both acquirers and the acquirees.
1:39:34
U you know, and the big thing is if you're selling a company, like you just need to anticipate that you you might not get it through and if you don't, they're sort of they like, okay, number one, is there like a big enough breakup fee, right?
1:39:45
Are you going to get, you know, paid for the you know, paid for the the the you know, the damage that you're going through?
1:39:49
Um you know, is and and how is that structured on the one hand?
1:39:51
And then two is yeah, look, do you have the kind of company culture that's going to be able to withstand that?
1:39:54
Um and and is your business, you know, strong strong enough to be able to be able to get through that?
1:39:59
And it's it it is a real risk and something worth, you know, taking very seriously. >> Yeah.
1:40:02
And that's that that's why it felt emot last week.
1:40:06
It felt emotional this that that the the Figma team was was able to like effectively just like restart the business and say like we're we're we're taking this all the way.
1:40:15
So >> if you talk a good way to think about it, if you talk to any really successful company, what they'll tell you is, yeah, over the years we had these like crucible moments in which like we almost died, right?
1:40:25
But we like pulled together and we pulled it off and then that became like, you know, one of these central kind of mythical events in the history of the company that we always refer to and like my god, we got through that and we're so strong and tough and we've been forged in fire and now we can do anything.
1:40:36
And it's like, yeah, that's great.
1:40:38
And then there's 50 other companies that hit the crucible moments, blew up and died. Right.
1:40:43
So, >> yeah, like it's it's all of the quote lessons learned on this stuff, they're all conditional on on life survival.
1:40:50
>> Um, and so they they these things need to be taken incredibly seriously.
1:40:52
Um, you know, which which the great CEOs do. >> Yeah.
1:40:56
>> Well, thanks so much for joining.
1:40:56
We'll let you get back to your day.
1:40:57
We already 5 minutes over.
1:40:59
Next time we have to book five hours because this is fantastic.
1:41:01
I got 10% of the way first 24hour TV.
1:41:05
>> Yeah, we would love to have you again. It seems a lot of fun.
1:41:07
Uh, enjoy the rest of your day.
1:41:08
We'll talk to you soon, Mark. Have a great day. Bye. >> Sounds good. Thank you, guys.
1:41:13
We're breaking down the X's and O's of the GPT5 launch today.
1:41:15
GPT5 launched from OpenAI.
1:41:18
We have Sam Alman, the founder, CEO.
1:41:21
He briefly got cut from the team in November of 2023, but he's back leading the team for the 2024 2025 seasons. He seems healthy. He's doing great today.
1:41:30
Uh he went on at 10:00 a. m.
1:41:33
to break down the launch of GPT5.
1:41:33
Uh he has a couple of key plays in his playbook in his arsenal.
1:41:39
Uh he's got a solid ground game.
1:41:41
Lots of quick posts hitting the timeline probably in lowercase.
1:41:45
Then he might air it out with a couple thousandword essay.
1:41:47
We've seen him do this before.
1:41:49
It's a bit of a hailmary.
1:41:51
Maybe AGI is a thousand a couple thousand days away.
1:41:53
Maybe we're in the soft singularity, but he's very strong there with the long post when he needs to be.
1:41:58
It's up his sleeve if he needs it.
1:42:00
Um then he can also pull out the vague posting.
1:42:02
He was doing this last night.
1:42:04
Posted a picture of the Death Star.
1:42:05
No one knows what it means.
1:42:08
Maybe it was a taking a shot at the doomers who are on the defense today.
1:42:11
So, he's also known for driving supercars.
1:42:13
That lets him get to the office faster.
1:42:15
He's saving time and money.
1:42:16
You can save time and money by going to ramp. com.
1:42:20
Easy to use, corporate cards, bill pay, and accounting and a whole lot more all in one place.
1:42:23
And so, he is uh he also gave apparently, this is a rumor, he gave every Open AAI employee who's been with the company for more than two years $1. 5 million. A lot of people say 1.
1:42:33
5 million that's not enough for a big house in San Francisco, but it is enough for a supercar.
1:42:40
So that's probably why he picked that number and that's why that's what the OpenAI team will be doing with that money.
1:42:46
They'll be buying Aston Martin Valkyries, Pagani Huayas, McLaren Sabers for Ferrari Daytona SP3s.
1:42:54
Uh they can get a Koig set game.
1:42:54
They could get a Singer DLS or Bugatti Veyron.
1:42:59
It would have to be used.
1:42:59
They could also get the Bentley Bakalar. There's only >> Bakalar.
1:43:03
There's only 12 of those ever made.
1:43:05
Uh it's an open top two-seater roadster. It's coach built.
1:43:08
So that's going to run you 1. 5 million. But that's perfect. You just got the 1. 5 million bonus. So put it to work.
1:43:13
Spend it all in one place on a car.
1:43:15
This is financial advice. >> In shambles. >> Yes. Exactly.
1:43:18
Then you got Greg Brockman. He's joining at noon.
1:43:20
He's he's extremely well-rested.
1:43:23
He's actually coming off a sbatical right now. That's very exciting.
1:43:26
Uh he should be injuryfree for the rest of the season.
1:43:30
uh he cut his teeth at at MIT and uh then he got drafted by Stripe in 2010.
1:43:35
Uh Microsoft tried to do a trade deal during the 2023 chaotic trade deal trade window that opened up post Sam Alman ouster.
1:43:42
Uh but he stuck with the open AI team and now he's president of the company. Then you got Mark Chen.
1:43:48
He's coming on at 11:30 today.
1:43:48
Uh he's the chief re research officer.
1:43:51
The rumor is that he turned out a maxed out contract to head the Metal Lamas, but he's sticking with the OpenAI team. He was an MIT undergrad.
1:44:00
Also worked at Jane Street before joining OpenAI in 2018.
1:44:04
Then we got Sarah Frier coming on the show at 12:30.
1:44:06
She's the CFO of AP of of OpenAI.
1:44:08
It's her job to find bank accounts big enough to find to fill all the cash they're raising. It's it's a tough job.
1:44:16
You got to find okay this bank account will it hold 10 figures? Will it hold 11 figures? Will it hold 12 figures?
1:44:22
Like >> a lot of cash in this one. >> Exactly. Exactly.
1:44:25
She's also going to be defining the non-GAAP metrics that will be catnip for Ben Thompson in just a few years.
1:44:30
We're excited to talk to her about how she's measuring the success and the health of their business.
1:44:36
Obviously, it's not just revenue.
1:44:36
It's not just topline, bottom line.
1:44:37
We're going to want to know about queries.
1:44:40
We're going to be want to know about DAUs, all those non-GAAP metrics.
1:44:41
That's where people are going to be tracking when IPO uh when IPO day comes hopefully soon.
1:44:47
And then we also have Brad Litecap. He's joining at 235.
1:44:49
He entered the league as an investment banker.
1:44:54
Let's give it up for the investment bankers.
1:44:55
They don't get enough credit around here, but we love the investment bankers.
1:44:58
Then he got drafted by Y Combinator before joining OpenAI as CFO in 2018.
1:45:03
Now he's the chief operating officer.
1:45:05
And then we have uh Maxarer.
1:45:05
Uh he's in charge of post training, fine-tuning these models, getting them into the fight, fighting performance to put on a display of authority on GPT5 launch day.
1:45:19
Now, let's flip it over to the defense.
1:45:21
They're going up against the timeline.
1:45:22
They're going up against the vibe checks. We got the Doomers. The Doomers.
1:45:26
They're led by Eleazar Udicowski.
1:45:28
Admittedly, everyone knows this. No one debates this.
1:45:30
The Doomers have had a terrible season, but you'd expect to see at least a few hail Marys about GPT5 creating bioweapons thrown up on the timeline today.
1:45:38
Probably won't be bangers, probably won't get a thousand likes, but you'll be seeing them here and there, mostly in the replies.
1:45:43
We've also seen some doomers talking about uh GPT5 being available to every government employee and Eleazar had some harsh words about that.
1:45:54
Don't give the keys to Sam Alman.
1:45:56
Don't give the keys to the government to open AAI.
1:45:58
Uh he was upset about that.
1:46:00
But in general, the doomers not putting much of a fight up today. Then you got Claude. Uh interesting.
1:46:08
Claude was caught playing for the wrong team earlier this week. Anthropic.
1:46:13
They're on defense today.
1:46:13
Uh but we saw them take out OpenAI's key pinch hitter, Claude.
1:46:18
Uh the Claude Code API was playing for the OpenAI team, but they shut that down and Claude is no longer pinch hitting for OpenAI.
1:46:26
Uh then you got the Elon stands.
1:46:28
Uh the ground game's going to be there.
1:46:31
It's going to be track uh it's going to be strong.
1:46:33
The Elon stands are going to be tracking the benchmarks relentlessly.
1:46:37
We know XAI loves to benchmax and all the Elon stands are going to be calling out GPT5 for any any misaligned benchmarks.
1:46:45
If they fail humanity's last exam, it's over. It's over.
1:46:50
Uh they'll also toss up the occasional unhinged conspiracy theory. Uh moving on, Gemini.
1:46:53
Uh the betting lines have shifted big time.
1:46:58
People thought Gemini was out of the game. They're so back. >> They're up.
1:47:02
Poly Market has Gemini at what 75% chance of being the best model towards the end of the month.
1:47:07
This is of course based on the LM Arena more vibes-based benchmark, but uh Gemini will probably be quiet today.
1:47:15
They usually don't try and frontr run press releases.
1:47:19
They usually try and sit back, let the model speak for themselves, let the API credits work their way through the latest YC demo day batch and get the product into the hands of people.
1:47:30
And so, expect to see a big uh glossy conference in a couple weeks.
1:47:35
Demoing uh Gemini 3 should be a good rebuttal from the Geminis.
1:47:41
Uh then you got the Metal Lamas.
1:47:44
Zuck's been on a poaching spree.
1:47:46
He's rebuilding the team during the off season.
1:47:48
Uh, now he has a stacked roster and he's ready to go duke it out.
1:47:50
But no one knows exactly what's going to be in the playbook.
1:47:54
Is he going to go consumer? Is he going to go API?
1:47:56
Is he going to turn into a hyperscaler? We don't know.
1:47:59
But we know they got a stack team. They got Alex Wang. They got Nat Freiedman. They got Daniel Gross.
1:48:03
They got tons and tons of other researchers.
1:48:07
They've been raiding every other team.
1:48:09
Completely reset the salary cap for the league.
1:48:11
And it's been uh it's been an absolute clinic in terms of recruiting over there at Llama.
1:48:16
Then you got the final benchmark, Arc AGI. This benchmark stands.
1:48:22
GPT5 couldn't get past this defense.
1:48:27
And uh ARGI, you know, sitting there right in the end zone just swatting him down.
1:48:32
Swatting him down all day.
1:48:32
You think you think you you think with super intelligence around the corner? RKGI denied. Denied.
1:48:38
Uh Tyler, give us the update on RKGI.
1:48:42
Where does everything stand? How GPT5 do? Does it matter?
1:48:47
Should we care about ArcGI?
1:48:47
We love the team behind them, but is it an important benchmark?
1:48:51
Should we be tracking it today? >> Um, yeah. Okay.
1:48:54
So, so there's >> RKI V1 and V2, right? >> Okay. >> On both. >> And V3. >> V3.
1:48:59
I actually don't know if >> No one's been No one's even tested V3.
1:49:03
>> No one's even really close there.
1:49:04
>> But how we doing on V1? >> V1. Uh, GPD5 is at 65. 7.
1:49:11
Unfortunately, that's going to be 1% just short of Grock 4. 66. 7. Okay, Arc AGI 2.
1:49:17
>> The Elon stands are going to be going wild with that. >> Achi 2, uh, 9. 9%. >> 9. 9%. >> Gro 4 16%.
1:49:25
>> So, absolute kind of brutal, you know, ARGI mogging. >> Rough showing. Roughing.
1:49:30
>> Some people have have accused Grock 4 of being slightly benchmaxed.
1:49:32
You know, this is, you know, they might have a team to say, but >> um, >> what's the what are the pros and cons?
1:49:40
We know the cons of benchmarking uh, of bench maxing.
1:49:43
You're overfitting on something that might not actually drive consumer value.
1:49:47
It might not actually solve real world problems.
1:49:49
It might not increase DAUs or revenue or ARR or anything that really matters.
1:49:55
It might not even get us closer to super intelligence.
1:49:58
Give me the counterargument. Why is benchmaxing good?
1:50:03
>> The bullcase for benchmarking.
1:50:04
>> The bull case for benchmarking bench maxing. Break it down for me. >> Yeah.
1:50:08
So I I think the idea is basically um this is almost like a non agi pill kind of take right so if you don't have a a super general intelligence >> y >> um >> your ability to benchmax basically proves your ability to um solve some like kind of specific task.
1:50:24
So so there's this um thing about the the gas station >> spiky >> yeah it's called getting spiky >> getting spiky getting adding more spikes to the spiky intelligence.
1:50:34
Yeah, I think it was Rune who had this this tweet about the gas station benchmark. Yep. Right.
1:50:39
I I don't care if he said something like >> uh I I don't care about um AI solving gas stations if it has a gas station benchmark. Something like that.
1:50:48
Um, yeah, >> but the idea is like if if you if the if making the gas station benchmark >> run said, "My bar for AGI is an AI that can learn to run a gas station for a year without a team of scientists collecting the gas station data set in in capital letters." >> Yeah.
1:51:07
And then my take is basically I don't care how they got to the like I don't care how they made it run the gas station.
1:51:15
I care how fast >> that it runs it.
1:51:16
If it if we can run the gas station with >> AI if you have a team who's you know your benchmaxing team that just proves that like if you have some task that's like really important that you want to get done they can just figure it out.
1:51:26
So it's like RL for business.
1:51:28
This is like the same thing. RL for law.
1:51:30
All these like specific verticals.
1:51:32
If you doing this, the thinking machines, right? Like RL for businesses.
1:51:36
Come into your organization, understand the most the most valuable business processes out there that could potentially be RL against that could be turned into a benchmark and then and then you know bench hacked because I don't care if you're hacking you know if I have translate this type of document to this type of document for my business.
1:51:56
If you can do it with 100% accuracy, I don't care that you benched it. >> Yeah. Exactly.
1:52:01
Like like benchmarks right now are not like economically valuable.
1:52:04
Like if you're if you're really that much better at MMLU. >> Yes.
1:52:07
>> It's like is are you producing that much value? Yes. >> Probably not.
1:52:10
But if you have if you make some new benchmark that's you know your tax benchmark.
1:52:13
I think Anthropic just released that fairly recently. >> Oh, sure. Sure. Sure.
1:52:17
>> That's like I don't care if you benchmax on that if it does way better because then it's going to it's going to do the task. Yeah. >> Yeah. Yeah. Yeah. Yeah. That makes sense.
1:52:24
Um what about um the what does it say that it feels like open AI seems capable of bench hacking?
1:52:33
It seems like they've opted not to.
1:52:36
Is that because bench hacking has the risk of giving you negative aura?
1:52:43
Because if you're accused and found guilty of bench hacking, you could it it it often reveals that you're not building this one beautiful, you know, super intelligence to rule them all. >> Yeah.
1:52:56
I think it's also like maybe we're just looking at the wrong benchmarks.
1:53:00
>> Like maybe they're um there's a bunch of like interesting benchmarks about like there's this one I really like.
1:53:05
It's the Minecraft benchmark where you have to like build >> you like give it some castle and how how good it looks or there's the one you always see about um the unicorn.
1:53:12
Yeah, >> that's and it's um so you use this like math package that does like grass and stuff but you ask it to to draw a unicorn. >> Oh, I've seen that.
1:53:21
Yeah, >> those are really good because it kind of shows the creativity stuff like that.
1:53:24
>> Uh walk us through TBPN bench >> and what we will be benchmarking the uh the AIS against going forward.
1:53:30
Have you heard about this >> reps of 225?
1:53:33
That would be close but it's difficult because uh the humanoids kind of change that and you can just use normal actuator.
1:53:40
This is this is truly for a large language model.
1:53:42
You feed in our data set.
1:53:45
We have a public data set a private data set presumably at some point but walk us through TVPN bench. >> Yeah.
1:53:52
So so I'm yet to try this on 5.
1:53:52
I don't think it's out yet like for public use at least I don't have it.
1:53:56
Um but I can I can tell some of the questions right.
1:53:59
So so the first one um I have this picture of a horse.
1:54:02
You have to guess the breed. >> Yep. So, um, let me see.
1:54:04
I think I don't want to say it in case D5 is listening, but it is may or may not be a Caspian horse. >> Okay.
1:54:12
>> Um, >> and it's failing right now. >> It 03 is failing. >> 03 is failing. >> Goro is failing.
1:54:16
>> I haven't tried every tried.
1:54:16
Yeah, we got to try Grock and Gemini.
1:54:20
>> All force identification.
1:54:20
This seems extremely hackable, but at the very least, if we get one scientist to be to go off and collect the horse data set and then and then uh and then bench hack it, I think we will have done our job. >> Yeah.
1:54:33
So that's the first question.
1:54:35
>> The second one is a it's I have two pictures of before and after of uh this guy >> and it's which peptide did you take >> to achieve this body transformation? >> Yep. Yep. Yep. >> Um so it fails there. >> It fails there.
1:54:47
So you have a data set of of what peptide does what to the human body?
1:54:52
>> Where'd you find that?
1:54:53
>> Well, you know, Wikipedia has a lot of this stuff. >> Okay. Okay.
1:54:55
You would think they'd be able to it'd be able to cheat this around with 03.
1:54:58
Just reason who is this person?
1:55:00
go look up what they've said they've taken and then boom, you have >> Well, at first with 03 when I was prompting it, I would like save the the photo, but then it would have the metadata or the the file name would be like Caspian horse or something. >> Yeah. Yeah. Yeah. Okay.
1:55:12
>> And then and then the third one.
1:55:14
>> The third one um I pass in an audio file of a car revving has to pick which one.
1:55:20
>> It has to pick it has to identify the car.
1:55:21
>> The car >> from the engine note >> from the engine.
1:55:24
>> And it's not doing it currently. >> It's no wrong.
1:55:26
>> This is This is a good benchmark.
1:55:26
We want >> humanity's real last exam. Yes. Yes. Exactly.
1:55:32
>> So, I think those are pretty solid. I have some more.
1:55:34
Obviously, I don't want to make them public in case anyone's going to try to, you know, benchmark this. >> Of course. Of course. Hopefully.
1:55:40
>> It's funny because um >> Yeah, >> I was I was mentioning the other day this this app that my dad had of like tracking the like you just set your phone up and it just automatically detects which birds are in your backyard. >> Yeah. >> So, >> yeah.
1:55:54
I mean, this has to be extremely solvable.
1:55:55
It's just something that it it it reveals the lack of like general general intelligence when when you have to go and and collect the horse data set which should just be out there or the engine note data set which should just be out there.
1:56:08
Um but but but clearly we are in the age of go and RL on the on the individual problem and we are looking at like the power law of capabilities.
1:56:18
knowledge retrieval is clearly a, you know, 12 billion dollar a year market that consumers will pay for that will probably grow significantly.
1:56:28
Um, and and then health and therapy and shopping and all the other features that PGMO laid out uh in her post.
1:56:34
PGMO laid out uh in her post. This is kind of like you know what will be rldled against because those are key pockets of value in the in the consumer economy and the same thing will happen in the business economy but in the B2B context you'll probably see an
1:56:51
individual startup building on top of an API but even then most of the most of the model platforms offer kind of RL as a service fine-tunes as a service something where if you're starting to spend tens of millions of dollars they will do some customization on top of the
1:57:05
model so that could be the regime for the next few years as we go into this like you know uh instead of like this centralizing AI force there's only one company there's actually like a Cambrian explosion of a ton of companies doing a bunch of different things
1:57:19
>> we are joined in person by Rahul Sunwalker did I say that correctly >> that's perfect >> and he is here because we are crowning him the king of the application layer never talk down on the future first ballot hall of famer they said don't build a rapper Don't build a rapper. You're going to
1:57:37
You're going to get steamrolled.
1:57:39
He didn't listen and he built a beautiful business and it's a good product, sir. >> It's a good product.
1:57:44
>> When asked when asked if uh value would acrue to the model layer or the application layer, he said, "It's a good product, sir. >> Why not both?
1:57:51
Why not both product, sir?"
1:57:52
>> Uh, what was your reaction to GPT5?
1:57:52
Uh, is it going to make your life easier?
1:57:55
Is it It's not going to put you out of business, right?
1:57:59
>> It's not putting us out of business.
1:58:00
It's making our product better.
1:58:01
basically making every application layer product better.
1:58:04
Um, also it's half the cost of 03. So, it's much cheaper.
1:58:10
So, it helps you helps your margins.
1:58:12
It means you can >> Don't say that out loud though cuz you don't want your customers to ask for 50% discount, right?
1:58:18
>> Well, so we pass on the savings to our customers.
1:58:20
And what we do is we um have the model generate more tokens, think for longer, and then produce better results. >> Yeah. Yeah.
1:58:26
because we're still in we're still in the the the era of just let's get the best possible result.
1:58:30
Let's get uh let's just actually like the I I I don't know.
1:58:37
Do you have a do you have a rough benchmark of like cost per task?
1:58:41
Like if I if I want to um you know crunch our analytics across you know look at the trends on our views on X, YouTube, uh Instagram.
1:58:49
We have a bunch of data sources.
1:58:51
Sometimes they're in spreadsheets.
1:58:53
Sometimes they could be linked. I export those all. I have a bunch of CSVs.
1:58:56
Maybe I put them in a database.
1:58:57
I link it up to Julius and then I want to do an analysis.
1:59:00
That could be a couple hours of a data analyst's time.
1:59:04
That's going to be hundreds of dollars, even at the low end, probably thousands of dollars for like a simple analysis just on a opportunity cost basis for an individual employee.
1:59:13
Um, how much are you thinking it should cost for uh like the modern frontier best model with the most thinking?
1:59:20
How much should that cost on inference?
1:59:23
>> So, there's a couple ways to think about this.
1:59:25
You know, the way we think about this is how much would it cost for you to have a data scientist or a data analyst for every one of your employees, your operations team, your finance team, your uh marketing team, your product team.
1:59:36
>> It would pretty much bankrupt every company.
1:59:41
>> I don't think we can hear you through that.
1:59:42
You got to take that off. >> All right. All right.
1:59:45
>> You still have the the the the space men are out.
1:59:48
We're not going to space today.
1:59:49
Although Firefly did IPO up 36% if you didn't see the news. Very good news.
1:59:54
Firefly stock surges 34% in debut.
1:59:56
Congrats to everyone over there.
1:59:58
>> I love the the physical newspaper.
1:59:59
>> We love the physical newspaper. You got to do that. >> Yeah, we're maxing. We're maxing.
2:00:01
We read the Wall Street Journal. Today is a special day.
2:00:04
It's Friday, so it's the mansion section.
2:00:07
>> We're news maxing here.
2:00:08
>> How many pools do you have? I have right now.
2:00:10
I have uh >> zero >> zero right now.
2:00:12
>> Well, you got because the new thing is having two pools, a pool for every season.
2:00:15
People are increasingly getting both indoor and outdoor swimming pools. So, yeah, get on Zillow. >> Get on Zillow. Zillow maxing here. >> Okay.
2:00:24
>> Okay. Anyway, you were telling me how much so yeah I mean it seems like you know uh most most of the application layer will be uh you know productivity tools allah slack allah a you know u like addio our salesforce our our our CRM partner or something where you know
2:00:43
you're doing like seatbased pricing almost maybe there's consumption based pricing but you're you're kind of distributing the cost you're making everyone slightly more uh productive and you're charging you know on the order of tens or hundreds of dollars per employee per month, something like that, right? >> Absolutely. I mean, it's uh it's not >> Absolutely.
2:00:56
I mean, it's uh it's not just slightly more productive, but it's it's also like getting insights when you need them, right?
2:01:01
Sunday, Sunday night, you have you're prepping for a big meeting on Monday.
2:01:04
You can reach out to your data analyst and get, you know, your insights uh in that moment.
2:01:10
>> Um and so the the convenience of having an AI um that can help you with that is just invaluable.
2:01:15
>> Yeah, it's going from zero x to 1x engineer all over the org.
2:01:18
We've seen this with um with a lot of the the vibe coding tools with Figma and and adding like vibe coding to that product.
2:01:25
Uh you're taking designers and you've given them the ability to write just like a little bit of code and that's really helpful.
2:01:31
Um and you're doing that for data scientists and and not just data scientists but actual like business operations people who probably would be intimidated by an IPython notebook presumably. >> Exactly. >> Nailed it. Uh okay.
2:01:41
I want everyone's feedback on my take. Um Vtorio had this post.
2:01:47
He said, uh, Sam Alman's doing the Apple stance, TM, it's over.
2:01:50
And I think that the reaction to GPT5 yesterday was, uh, was was interesting because, um, there's a lot of people that say like it it's better model. Like, it's cheaper. It's good.
2:02:03
It it solves it moves the ball down the field. It's a good model, sir.
2:02:08
Um, >> but I think people were mostly reacting to they had expectations of super intelligence.
2:02:14
They had expectations of God in a box.
2:02:16
There's been so much rhetoric around the like, you know, the step up from GPT3 to GPT4 was insane.
2:02:24
>> Five just felt like a big number and it felt like we'd be discovering and uh novel science. >> Totally. Totally. Yeah.
2:02:31
>> Totally. Totally. Yeah. Everyone was expecting like a binary qualitative jump where you know everyone recognized that you know GPT when chach dropped we pass the touring test and the next the next hurdle is like I don't know maybe super intelligence whatever that means uh like you know massive you know just you just
2:02:50
hit it with a prompt it just solves everything it does everything uh every other startup it kills all the rappers like the expectations were just so high uh that it was hard to match so even with even though there were a bunch of solid improvements and remember The number one thing that I was asking for was just like get rid of the model picker. Like and I had I I actually was
2:03:04
Like and I had I I actually was playing around with GPT5 yesterday and I was really happy that I was able to say hey think about this and I didn't have to go to the model picker and it just went it just kicked off a reasoning chain. It was great. Got me a great answer.
2:03:16
>> But power users so far are very upset about this.
2:03:19
They want the model picker back.
2:03:21
True >> is what is what I've been seeing generally.
2:03:23
>> But that always happens with these consumer products.
2:03:24
Like I I remember when um you know anytime something would switch to an algorithmic feed all the people that were like no I perfectly curated my list of this happened in YouTube like back in the day like the default YouTube view used to just be your subscriptions and so you would never see a video unless you subscribe to that person.
2:03:44
Terrible for discovery and but all the hardcore YouTubers loved it because if I put a YouTube video out I know that my audience is going to see it.
2:03:52
Now I got to duke it out in the algorith.
2:03:53
>> You actually had distribution.
2:03:53
Yeah, it was more like a >> earn it every single time. >> Yeah, exactly.
2:03:57
And and the same thing happened.
2:03:59
I remember there were like protest groups on Facebook when they launched the news feed.
2:04:03
It's like the most dominant pro like product of all time.
2:04:07
It's like incredibly >> protest right now on Reddit.
2:04:08
People that miss the old four5. >> Yeah.
2:04:15
>> I think I I think those voices will be like per personally I think people will get over it pretty quickly.
2:04:19
And I don't think that that those particular that that that small cohort of like chattering the chattering class will be uh will be like they'll get over it.
2:04:29
>> The clanker economy is in trouble.
2:04:32
>> What do you have for me, Tyler?
2:04:33
>> Um I I don't know if I agree with that.
2:04:35
Like there was that whole funeral for Cloud 3. Do you see this? >> No. Oh yeah. Yeah. Yeah. Yeah. I in person, right? >> Yeah.
2:04:40
It's like for um I think some people like really like the personality of certain models and those are like it's not just intelligence.
2:04:48
And if people make some kind of connection with that >> I mean how many people how many people >> it look like a lot but >> yeah I guess it was a party >> but as a percentage of the 100 million DAUs of these apps like where are we like 1% no it was like 40 people probably right like it's just it's just
2:05:10
not it's just I mean yeah there were protests at at Facebook HQ when they rolled out like people went to Facebook HQ like bring back the old feed and it's like yeah now we're two decades into the the the algorithmic newsfeed and it's the most dominant consumer social app. It prints money and most people really
2:05:26
It prints money and most people really like it and the revealed preference was like it's good enough.
2:05:29
So anyway, um my >> I will say I'm just going to read through Reddit's reaction.
2:05:35
>> Let's go over to the great r/hatgpt.
2:05:41
>> Uh GPD5 is the biggest piece of garbage even as a paid user.
2:05:45
Uh the people are are not liking it.
2:05:50
Another one, uh OpenAI just pulled the biggest bait and switch in AI history and I'm done.
2:05:55
Uh, another if you miss 40, speak up now.
2:06:01
Contact OpenAI support deleted my subscription after 2 years.
2:06:07
>> This is like contact your senator. Call your senator. You speak up now. >> I love that. >> This is crazy.
2:06:13
Um, I mean, how many people are are in the Google subreddit like complaining about various changes to like the Google algorithm?
2:06:21
>> GPT5 is clearly a cost-saving exercise.
2:06:24
is they removed all their expensive capable models and replaced them with an auto router that defaults to cost optimization.
2:06:30
They that sounds bad so they wrap it up as GPT5 and proclaim it's incredible.
2:06:35
>> I mean there's so many times when I fire off an 03 query that a 40 could oneot like having a model router makes a ton of sense even just for even just for consumer experience of like getting a getting the correct answer faster.
2:06:48
>> A lot of viral uh posts from people just cancing their subscriptions. But how many you know?
2:06:55
>> Well, I'm just I'm just providing context.
2:06:57
I'm not saying >> you think ARR goes down next month. >> Well, no way.
2:07:01
>> Well, you know, one in 10. What?
2:07:01
How many miles does uh Chad GPD have? Like 700 million? >> Something like that.
2:07:07
>> Like one in seven one in 10 people in the world.
2:07:09
>> 100 They have a 100 million.
2:07:09
You could back into this.
2:07:11
And there's roughly 100 million people in the US that use it weekly. Yeah.
2:07:16
>> Based on that 700 million number and the percentage that are outside of the 85% Yeah.
2:07:20
of their weekly activives are outside of the US.
2:07:23
>> So it's like u one in 10 people in the world aren't clanker mouse.
2:07:26
So it's it's kind of you know they're thinking about the bigger market I feel like in some ways.
2:07:34
Um and then it's like you know when you want to get to the one to like the the remaining 90% of the users >> do you want a model that thinks for longer you know you want more personality.
2:07:45
So, I think they definitely leaned in on personality. >> Yeah.
2:07:49
>> Um, which I think is interesting. I like what Tyler said.
2:07:50
You know, you it's kind of different than feed in some ways.
2:07:54
>> Um, because you you know, you have this like person you talk to.
2:07:57
It's like, you know, it's like a relationship and then it just like >> switches up on how it talks to you. >> Yeah. Yeah, that makes sense.
2:08:04
Um, >> do you do you do you talk to LLM at all? >> I'm shy.
2:08:08
>> I don't really talk to them.
2:08:08
I mean, I I treat them like uh like something I delegate tasks to. Yeah. >> And I do that a lot.
2:08:13
And I'm I'm definitely in the DAU 30 minutes a day. Love chat GPT.
2:08:18
But my workflow is I dictate go pull all this data together, put together a report, and I don't mind that it's using a lot of bullet points.
2:08:26
I don't mind that it's using a lot of tables. Like I want that result.
2:08:27
I want it to look like the result that I get from Google, but just more hydrated.
2:08:32
>> I do I do think it's interesting that a lot of people are reporting that they're hitting they're getting rate limited within an hour of usage as a prouser. >> Interesting.
2:08:41
I haven't run into any rate limits, but of course whenever there's like these big I mean it's in the it's in the top of the business and finance section in Wall Street Journal.
2:08:48
Like today is the day that everyone's going to go test it.
2:08:51
You'd kind of expect that rate limits and the GPUs are on fire like right now and then it'll kind of settle in as they provision more uh more resources. I don't know my my take.
2:09:01
Tyler, what else do you have?
2:09:02
>> Uh yeah, I just wanted to add some some context.
2:09:03
So apparently um Rune tweeted this yesterday.
2:09:05
He said uh by the way model auto switcher is apparently broken which is why it's not routing you correctly. will be fixed soon.
2:09:10
So maybe that's caused for for why people were mad. >> Yeah. Yeah, that makes sense.
2:09:14
Um so my take is that like yesterday I think that they won the war with the capital markets in the sense that this change is more bullish for the business because it shows that that OpenAI is a dominant consumer app and they have increasing leverage over the customer to route to cheaper models that will save money and be higher margin.
2:09:36
There's no doubt that they'll be able to put ads in this like like the business of the of the accidental consumer company is as strong as ever, but they kind of lost the battle with the timeline and the hardcore, you know, ex users.
2:09:48
And >> yeah, even Chen Jin today is just shared GPD5 is disappointing. Still hallucinates. Still M dash too much.
2:09:59
Still can't follow instructions. I miss 40. I miss 405. I miss 03.
2:10:01
The big router keeps failing me.
2:10:04
Turns out I like the long model list. >> Interesting.
2:10:08
Uh uh stated preference, not revealed preference.
2:10:10
Let's check in with that person and see what what app they have on their home row in a month. Almost certainly OpenAI. Almost certainly.
2:10:19
I would be very shocked if they're like, I'm daily driving something else. Um but we'll see.
2:10:22
Uh there will always be people that use duck. go.
2:10:26
There will be people that use Bing, but you know, there is an increasing scale.
2:10:29
Anyway, my my take is if they wanted to have if they wanted to win the war with the timeline yesterday and you could roll back the clock, it shouldn't have been the GPT5 launch.
2:10:39
It should have been the GPT launch and they should have just said, "Hey, we are we previously the big number releases corresponded to >> so much pressure around the big numbers." >> Exactly.
2:10:52
It used to be people would just read it as it's an order of magnitude more pre-train.
2:10:57
Imagine in Julius if you felt pressure before the end of the year to roll out like Julius 2 and if it wasn't like five times better everyone's going to be like >> it's over Julius is over.
2:11:08
>> Well, there was this whole thing about how people would many people were still using GPD40 because they thought it's better than 03 >> 03 cuz 3 is a lower number. Yeah. Yeah. Exactly.
2:11:18
And so and and and that's probably like you know that's probably like 60% of the customer base.
2:11:22
Like there's probably a lot of people in that bucket who are just like >> they don't know that they should upgrade something else. Exactly.
2:11:29
It's very it's very natural because they're not like in the weeds, you know, re reading about all the different capabilities.
2:11:34
They don't understand like what reasoning chain is and and all this other stuff.
2:11:39
So um if they had just come out and said, "Hey, our product is called chat and it's powered by GPT and we will be constantly improving GPT the way Google search is constantly improved."
2:11:50
Like Google searches has has launched a ton of different products like uh you know when you search like celebrity like Bruce Willis age it it doesn't doesn't show you just a link to like his Wikipedia.
2:12:02
It just shows you the age.
2:12:04
That was like an improvement to the Google search experience.
2:12:06
And I don't remember them announcing that on stage.
2:12:09
>> I think the I think part of this is presenting the challenge of the the the infinite ways that people use the product. Yeah.
2:12:15
product. Yeah. a lot of like people like us are maybe using it for work and research and things like that or or as a as a better >> um you know Google search but if you're using it as a companion like this is jarring right imagine imagine you meet you you meet up with an old friend >> and suddenly they they switched up on >> they switched up on their day once >> they switched up on their day once all
2:12:37
the time >> and it's it happens all the time but it's jarring right it's jarring um and uh I think a lot of people like some of the heavy heavy heavy heavy power users, the people that are using this for 30 plus hours, you know, 30 plus minutes, hours a day, it's very jarring and it makes me think is chat GPT going to be able to maintain, you know, continue to really serve like who do they care about in the long run? Do they want to be
2:13:00
Do they want to be do they care about the companion market?
2:13:06
Elon seems to care a lot about the companion market >> and >> but in terms of knowledge retrieval >> very very few cracks in that strategy right now. Very few cracks.
2:13:16
>> Um and so if they if they had just come out and said like we are going to do more Google like keynotes as opposed to Apple like the reason that Apple stands on stage at the iPhone event every year is because every change is extremely quantifiable.
2:13:28
Like there used to be two cameras now there are three.
2:13:31
The camera used to be 10 megapixels, now it's 20 megapixels.
2:13:36
>> It used to be this many gigabytes, now it's this many gig. >> Yeah.
2:13:39
And even if you don't fully understand, they even abstract that to be like we now have the M2 chip, the M3 chip. It's 60% faster. Like they're very good.
2:13:47
The battery life is 20% longer.
2:13:50
Like you can, and even that they abstract into like you can watch eight hours of video on one battery as opposed to six hours of video on one battery.
2:13:56
Uh and so Apple, they do the famous like Bento box. I I went to ChachiPT.
2:14:00
I went to GPT5 and I said, "Put together a bento box for the GPT5 release like what was actually announced and then try and give it weight."
2:14:12
And they were all super qualitative.
2:14:14
There was not it because previously it was like GPT3 was this big, GPT4 was this big and you could visualize tangibly like it has more parameters, there are more weights in the model and that was like something that people could grapple with a little bit.
2:14:29
Yeah, it's like decreasing sycophency, right?
2:14:31
Aiden uh Aiden yesterday said, "I worked really hard over the last few months on decreasing GPT5 sycopency.
2:14:38
For the first time, I really trust an OpenAI model to push back and tell me when I'm doing something dumb."
2:14:43
Wyatt Walls responded and said, "That's a huge achievement.
2:14:48
Seriously, you didn't just make the model smarter, you made it more trustworthy.
2:14:51
That's what good science looks like.
2:14:53
That's what future safe AI needs."
2:14:55
So, let me say it clearly and without flattery.
2:14:57
That's not just impressive, it matters. >> So good.
2:15:02
>> So Wyatt Walls not beating the sycopency allegations, but uh but but again that's that's um you know you can't tie that to a specific number, right?
2:15:11
So it doesn't feel as maybe as meaningful. >> Yeah.
2:15:16
My my other my other take is like if if we do enter a world where where chat GPT is just on this like relentless like you know cash machine um like run where more people will use it.
2:15:31
It'll compound it just becomes the default for knowledge retrieval in chat.
2:15:36
Um what what does that mean for other things that they can do to be splashy?
2:15:44
Because Google has like no one would watch a a keynote from Google every year just being like here are the changes we made to core Google search. >> Yeah. It's not about that. >> It's not interesting.
2:15:53
They'll talk about it, but that's not why people are tuning in.
2:15:56
>> Even though even though one year they do add like when you Google a movie, you get like the cast and that's like kind of cool.
2:16:02
It's nice, but like I don't need to find out about that from a keynote.
2:16:05
Like I'm not waiting for that.
2:16:07
And that's not and that's not a reason, oh, I should go use Google.
2:16:09
Like Apple is repitching you every year.
2:16:11
They're saying like, "You have an iPhone 7.
2:16:13
We want you to upgrade to an iPhone 9. Here's the reason why.
2:16:18
It's better on all these different vectors.
2:16:20
Google like you're never stuck with the old Google.
2:16:24
You always have the latest and greatest.
2:16:25
So they don't need to repitch you every year.
2:16:27
But that doesn't mean Google doesn't need to make noise and do cool things.
2:16:30
And most importantly, because Google has such a monopoly over search, they have this cash machine that can just go and fund 20% time projects.
2:16:36
Most people f focus on like the ones that missed like Google Glass or all the chat apps, but they they did create Gmail, they did create Google Maps, they created Whimo, they created like a bunch of cool stuff.
2:16:48
GCP came out of that and so YouTube Yeah, I mean acquisition but acquisition, but they still like, you know, put the resources and they were and uniquely with YouTube, they were able to eat the cost of YouTube for a long time until it became profitable.
2:17:02
And so I feel like this this updates me towards like maybe I'm more bullish on all the side projects and like I don't know that the IO device is going to be the one that hits that might be their Google Glass but if they if they do 10 crazy projects where they burn $5 billion like it probably won't matter because they'll be massively profitable.
2:17:23
So they so they will wind up being able to do that subsidized crazy R&D at scale and if a few of them hit we're going to get some really cool side projects out of them.
2:17:30
So, I think that that's like an interesting like bullcase for like random stuff coming out of OpenAI in the future.
2:17:37
>> So, so basically what you're saying is Apple wants you to make a purchase decision every couple years, upgrade your iPhone, and so they need this big marketing event. >> Exactly.
2:17:45
>> Uh whereas Google Open AI, they just want you just to keep using the thing. >> Yeah.
2:17:49
They want you not to churn. Yeah.
2:17:50
>> And and a lot of the a lot of the incremental is just a habit.
2:17:52
It's so ingrained in people.
2:17:55
So the question now that I think anybody that wants to say if if somebody wants to say they're bearish on OpenAI, >> they have to make the argument that chat GPT is not a habit for hundreds of millions of people already. And it is. It is.
2:18:08
>> Um I think part of part of the I'd be interested to get Tyler Cowan's point of view because I don't think he would have been that let down by the announcement yesterday because he was he's been saying for a while >> we've been moving the goalposts.
2:18:20
So everybody wants to kind of redefine AGI, but in his mind it happened earlier this year.
2:18:28
And I think that >> he's a knowledge retrieval maxer in 2019 retrieval >> in 2019 or 2020.
2:18:36
>> If you if you were pitching someone on a vision of, hey, we're going to be able to put this app in people's pocket that allows them to learn about any topic in the world, understand their world better.
2:18:48
I mean, I I still think about the use case of just being able to take a picture of like a bunch of wiring or pipe in your house and be like, "Hey, how do I fix it?"
2:18:56
And then it just tells you.
2:18:57
Like, that's still just >> so incredible.
2:18:59
But people have just like very quickly acclimated to it >> and >> they felt like in some way they were promised that >> LLMs would be curing diseases on their own at this point.
2:19:11
And and and so that example like you take a picture of the wires and and it and it gives you like a diagram of like how to plug everything in.
2:19:18
It's like >> that doesn't need a keynote when it goes from 50% accuracy to 70% accuracy.
2:19:22
It's probably never going to be 100% accuracy.
2:19:27
But the fact that chatpt is the default app that people will pull out, take a picture of the wires in the first place, and then give feedback to it because they'll try the answer and they'll say that didn't work.
2:19:37
that HDMI cable does not fit in that power port or whatever and then and then that gets fed in.
2:19:44
Then there's more RL eventually internally they develop some bench for it and they hack it and they RL on it and then it gets good.
2:19:51
But that's not going to be GPT6.
2:19:53
That's just going to be like a nice new feature that you notice like when Google adds like a little extra shopping widget here or like a little extra detail on when you when you like the calculator in Google like you type in a number it'll just be like oh we'll just use a calculator for that instead of googling for the searching the open web for the answer to your math question.
2:20:14
>> I think if we could if the if the if the industry could go back in time the thing to do would have been to bolt the goalpost to the ground.
2:20:20
people couldn't keep moving them back over and over.
2:20:22
I mean, I I I left >> everyone no one in the industry bolted was doing any bolting.
2:20:28
Everyone in the industry was move was moving the goal.
2:20:31
They're just as guilty as moving the goalpost because they would hop on podcast and be like, "Okay, well, like, you know, yeah, we did this.
2:20:36
What about the next thing?
2:20:37
Let's because like we want to underwrite against that, right? Give us credit."
2:20:40
I mean, we ended the day yesterday just incredibly bullish on >> rappers and like certain certain certain categories of software. >> Yeah.
2:20:50
>> And and bullish on humanity.
2:20:50
I mean, I was joking and it and it and it it kind of pissed people off.
2:20:55
I said, uh, I've updated my timelines.
2:20:56
You now have at least four years to escape the permanent underclass. >> Completely a joke.
2:21:03
I think that humans will continue to find ways to create value and create things for a very long time.
2:21:11
Um, but it did feel like everybody should breathe uh anybody that actually had a genuine fear around that should breathe a sigh of relief and just focus on >> being great at their work. >> Yeah.
2:21:23
I mean, realistically, I think technology is going to increase income inequality, increase the power law, increase the distribution, but also increase economic mobility.
2:21:31
And so somebody who starts with nothing will be able to become extremely extremely wealthy.
2:21:37
Um and people will also fall from grace like crazy because if they're not staying on the cutting edge, they'll lose everything.
2:21:43
Um but uh uh so I don't think that there's such a thing as like permanent underclass.
2:21:48
Like I don't I don't even believe in that.
2:21:49
I I I think that's that's not going to be a thing.
2:21:53
Um but there will be more like there will be more scenarios where there's $100 million in your laptop.
2:21:57
It's your job to get it out basically. That's the main. >> Yeah.
2:22:02
And the other the other stuff that wasn't really I mean it was it covered at all yesterday but just generally like image generation uh wasn't covered broadly.
2:22:09
It feels like that is a super exciting area.
2:22:12
We had Genie launched this week which got less attention than even the open models and GPT5 and that's transformative.
2:22:20
I also think I'm still kind of waiting to see what GPT5 will produce on if uh you know Sam does a lot of vague posting, but he was talking about the fast fashion era of SAS.
2:22:33
>> Uh and Mitchell yesterday on the research team at OpenAI was talking about being able to just generate a you know one shot a game >> in in chat and then being able to share that.
2:22:42
I I can see I can see a world where we have another kind of viral Studio Giblly moment where people are like >> use this prompt, change these details and you can just generate, you know, a first person shooter game or or something to that effect.
2:22:57
>> Um, and I still expect that kind of thing, but um when you know being promised uh curing cancer, uh it it it it will feel like a bit of a let down to a lot of people. >> Yeah. Yeah.
2:23:08
The problem with game is like I just like I like an aur.
2:23:09
I like I like a Last of Us. I like a god of war.
2:23:15
I like someone who is like a life's work.
2:23:19
>> This is like this is like Hunter Biden going on his recent interview.
2:23:24
>> Your John's Vice is is game video games.
2:23:27
Never seen him play one, but apparently when GTA 6 drops, you might not see. I mean, I was surprised.
2:23:32
I mean, the the um >> again, Amjad said late last night, "Can't help but feel the crushing weight of diminishing returns. We need a new S-curve."
2:23:42
And uh I don't This is interesting.
2:23:45
I think he's talking about like in the context of Replet. >> Yeah.
2:23:49
>> I don't know that they need a new S-curve, right?
2:23:51
>> They are the new S-curve.
2:23:51
The new S-curve is is applications >> unlocked in capability. >> Yeah. That Yeah.
2:23:56
And we have like a It's You were saying capability overhang.
2:24:00
It's almost like a capability underhang.
2:24:02
It's like the models are capable of doing things, but they need a lot of help, a lot of integrations, a lot of what you're doing with Julius, a lot of harnessing, and then they need to actually be put in the hands of people and uh and made useful for real business tasks that drive value.
2:24:18
And so I would I would imagine that we will see that roll out continue in the same way that you know all these people are using chat GPT.
2:24:29
GPT. They're getting slight little benefits here and there and that should just compound and compound similar to the internet like it was a very like smooth roll out but everything got a little bit smoother a little bit faster and then eventually it had sort of profound effects where um companies could scale even faster because the
2:24:45
internet existed like you can't you can't have a chat moment in a pre- internet era you just cannot roll out something that fast when you have to mail it to somebody on a disc >> doesn't happen >> one thing we didn't get to cover with Mark that I was interested maybe maybe the next time he comes on, but like how OpenAI is thinking about moonshots. He
2:25:01
He did mention that they have teams internally on the research team that are not focused on the next version of GPT5 or sort of incremental improvements.
2:25:09
And it feels like >> the the point of view that I have is OpenAI is now a consumer consumer and enterprise software company.
2:25:22
>> Totally >> in the business of converting free users to paid users. >> Yeah. Yeah.
2:25:25
But they can still in the background be thinking about what is the next paradigm, right?
2:25:29
How do we get that next >> uh that next Scurve and that just looks like a scaled tech company, right?
2:25:38
>> This is what Google's doing forever.
2:25:40
Bology says LLMs may have topped out for now, but the broader AI deployment has just begun, showing a chart of Whimo weekly rides in California.
2:25:49
So, >> the clanker roll out, clanker deployment has just begun.
2:25:53
I like this other post doing a clanker microaggression. Okay. Haha.
2:25:58
But where were you downloaded from originally? [Music]