Big Technology Podcast - ChatGPT As An Insult, NVIDIA's Moat, A New Musk Profile

Episode Date: August 25, 2023

Ranjan Roy from Margins is back for our weekly discussion of the latest tech news. We cover: 1) ChatGPT used as an insult in the Republican debate 2) Whether Generative AI is leading to real business returns 3) Bing's poor performance vs. Google 4) NVIDIA's blowout earnings 5) NVIDIA's moat 6) ARM's forthcoming IPO 7) Ronan Farrow's new profile of Elon Musk in the New Yorker 8) Government's role in privatization. Enjoying Big Technology Podcast? Please rate us five stars ⭐⭐⭐⭐⭐ in your podcast app of choice. For weekly updates on the show, sign up for the pod newsletter on LinkedIn: https://www.linkedin.com/newsletters/6901970121829801984/ Questions? Feedback? Write to: [e-mail]

Chad Cheap-T is an insult at the debates. Nvidia is surging, but how high can it actually go? And a new profile of Elon Musk hits newsstands this week. We'll break it down right after this. Welcome to Big Technology Podcasts Friday edition, where we break down the news in our traditional cool-headed and nuanced format. We have a lot to talk about this week, plenty to talk about regarding generative AI. Some questions are being raised about whether this new technology has a ceiling.

We talk about a lot on the show. But this week, we're going to break down some of the recent news, including a very interesting survey and what actually just happened at the debates where chat GPT has been used as an insult. Ranjan Roy is here with us today. Welcome, Ron John. We did not mention generative AI last week once, which was a record. So I think it's time to devote a good amount of this episode to Gen AI. Yeah, definitely time to pick up on it.

And we have plenty of interviews on the Wednesday show that are coming up and have run. Saragua was just on this week of conviction. We have Colin Murdoch of Deep Mind coming up next week and then after that, Surita Ramoswamy, of Neva, now Snowflake, we're all going to be talking about this stuff. So we're right in the think of it. And it's at this extremely interesting moment because we're at this place where we've had the hype around generative AI. Then everybody got building. And now we're like at this moment where it's, hey, was that hype all misplaced? Or is all this building going to lead to something, and we're in the place where we don't really know yet. And that's kind of where it got interesting because they're starting to become this narrative

around generative AI that's turning a little negative. So, for instance, during the debate this week, Chris Christie said that Vivek Ramoswamy was a guy who sounds like chat GPT. I've had enough already tonight of a guy who sounds like chat GPT standing up here. and using that as an insult basically saying you sound like a chatbot even though we've been talking about how this chatbot is brilliant and the next maybe the next artificial general intelligence so ronjan i'm curious like what you made of that and what you think it signals yeah i think to have that in kind of the zeitgeist as an insult especially on the republican debate stage is a moment it's a sign of where we are right now and i think that it's important to understand that the hype cycle, we're now, let's say, the launch of chat GPT publicly last November 2022, I believe, is when this really got moving and really kicked off. We're, you know, closing in on a year of just massive hype.

So I think we're definitely going to start to see a little bit of fatigue, a little bit of resistance, a little bit of disappointment with the technology. And I think it's kind of amazing that Chris Christie, brought that to the debate stage was a great zinger and here's gary marcus who's a you know a friend of the show notable critic of generative ai uh i think gary goes a little too far on on some of this stuff but i'm curious i'm going to throw out his line and we can start to take it apart he says in less than a year chat ch pt has gone from being mistaken for artificial general intelligence to being the butt of a joke and an insulting shorthand for robotic incoherent unreliable and untrustworthy is that

fair? I don't think that's fair, but I agree with you on that side. But I think Gary Mark has also had a piece, is generative AI a dud? He wrote the rise and fall of chat GPT just this week. I think he's playing an important part in this. And again, you know, someone who has worked with deep learning and artificial intelligence forever, who has, you know, successfully sold a company to Uber, who was a professor at NYU who's I mean I honestly like he helps make sense of this technology in a way on the episodes he's been on here in a you know in a very accessible way so it's good to have someone kind of pouring a little water on this however I think the idea that it's a complete dud is wrong I think I see even in everyday life I mean even people I never would have expected

just actually using chat GPT or using some kind of generative AI text genera in their just day-to-day life. And we can get more into, you know, other utilizations of code interpreter on chat GPT or Dolly or other image generation. It's just becoming so commonplace now that at least to generate mediocre things, everyone is doing it now. I mean, I think we should definitely get into the enterprise adoption side because I think there's something to be said there. But I think in terms of just popularity, I think it's getting more ingrained in how people live and work. than ever. Yeah. And I think since we've moved past the hype moment, this is sort of where we now run into reality. And reality is two things. First of all, it's bringing into focus that this stuff

is difficult to build. And second of all, there's a building phase. And it's not going to happen immediately. And I think, you know, you mentioned the enterprise side. And this is something that Marcus links to in his piece. It's from Axios. Companies struggle to deploy AI due to high costs and confusion. And the piece says that nearly 70% of respondents at the SMP Global Survey said they have at least one AI project in production. But about half of those companies, 31% of respondents are still in pilot or proof of concept stage, outnumbering those who have reached enterprise scale with an AI project, 28% of total respondents. But I think the framing, so, so I think the framing of this story is completely off. The fact that 28% of respondents already have something,

you know, with AI in production, you know, sort of shows that this has been an exceptionally fast rollout. And we had chat GPT came out last November. We're August. We're not even a year past. And we already have a third of companies in production. Then we also have a third who are working on it. And, you know, this stuff isn't, it's the change management is tough. We had Sarah on this past week talking about how integrating this stuff into new workflows is tough. And we're getting to like the what happens when you put humans and technology together and that's going to be inherently messy and it's the human part that's messy. It's not the technology part that's messy. And the nice thing about it is the tech is there and it's just the human side is going

to get worked out. Yeah, I think in some ways I think we're agreeing here. But also I would say that when you hear 70% of respondents to this survey said they are working on an AI project and then 28% are enterprise scale. Given this is a loose survey and everyone right now is being trained to just say AI, AI, generative AI, generative AI, I almost would take those numbers with a grain of salt and almost discount at some degree that how many people actually have projects at enterprise scale. But I say that all because I am so happy right now. Finally there's a bit of like a reality check in all of this as someone who has again worked with natural language generation since 2015 with my startup since 2021 have been working with gpt3 and other lLMs like a lot the last year has been

really interesting for me because in some ways when you are working on something that becomes so incredibly overhyped it's kind of cool you feel like you know you you got you started listening to the band before they came became cool but in other it's it's kind of terrifying because you're like can this and I've written about this I think we've debated about this can technology be hyped out of existence and I still always wonder with crypto did it have a potential future and actually transforming the financial infrastructure that we live with if it didn't turn into this insanely hype technology and with generative AI if we don't have a bit of cold water on this I do worry about that because then what's going to happen

and I think you're starting to see it now, once everyone actually starts building. And anyone who's really worked on building with this technology knows it doesn't work well the first time. It certainly doesn't work perfectly. So whether you're willing to go through the iterations and kind of perseverance to actually get it where you need it to go is a whole other debate. But I think everyone honestly thought you're just going to plug into an API and suddenly just stop having to work. Right. And it sort of goes to sort of one of the purposes of this show, which is to bring some nuance to these big conversations. You know, the loudest voices in the room have held really two positions.

There's been this one group who talked about how generative AI or AI is on the path to wiping out the world, standing on 60 minutes and saying everybody look out. And then the other side saying that this is a dud. And it is somewhere in the middle. And we're like in the middle of an S curve where that building is happening. And the building is messy. And you look at some of the statistics in that Axios piece and you're like, okay, well, this makes sense. And one of the things that I highlighted is that companies are finding their data isn't organized for AI. It's saved in different formats. It's in disparate data sets. It's sometimes still on paper. This is a change management exercise among everything, you know, among all everything else. And so there's a huge part of it is just trying

to figure out how to organize to take advantage of this technology. And I think people are going to mistake the organization part and getting ready part for the fact that the technology doesn't work. And that's what my perspective is here is that the technology does work. And in this messy middle, it's going to open up a big moment. I think we're about to experience a big wave of criticism about generative AI. And then ultimately, I think those critics are probably going to have to, you know, eat their words. You know, even though it might not be, I never thought it was going to be the wipe the world out. technology but it is an impressive technology and I think that you know as we get deeper into

the build phase we're going to see it rolled out this is incredibly frustrating because we're agreeing on way too much today because to me the best part of that axios piece was also I agree the data infrastructure side and it was almost there's a line about data warehousing as a competitive advantage which sounds like the most boring thing in the world but again having built with LLMs and generative AI technologies, I've seen it firsthand and you see it that you have to have clean data. If you want to have like garbage in, garbage out, if you want good output, you need very well-structured clean data.

And everyone has just kind of been sold this story that you're going to just plug in every file you've ever had and just whatever sources it is. And LLMs create magic and can read anything and understand anything and then generate anything, and it's wrong. And I actually, I think that's exactly where we're going to see a lot of problems. And especially when you get into giant corporations where data cleanliness and, you know, like how there's data structured and accessibility, these things are known to be messy. They're going to be the first ones that have definitely issues trying to make things happen. be a reason why a company like a snowflake, which is all about data warehousing, will acquire aniva, which is, you know, Google competitor that's been working on LMs.

We'll talk about that with StreetR in a couple weeks. But, you know, all right, let me do a mea culpa here. On this show, I had talked about how I was a Bingboy and, you know, thought that... I was going to, I was going to bring this one up. I have to apologize. I have to bring up Bingboy. Because I had thought that having chat GPT like functionality or GPT4 like functionality baked into Bing was going to help it cut into Google's lead. No, it hasn't. And there is a story this over the past week that talks about how Bing's market cap or Bing's share, sorry, Bing's share has not even budged in terms of where it was fighting against Google. So this is from the Wall

Street Journal. Even AI hasn't helped Microsoft's Bing chip away at Google search dominance. And Microsoft said they could add $2 billion in revenue if it could pry away even a single point of market share from Google. Here's the story. Six months later, it looks like even one percentage point could be a tough target with some new data showing Bing's place in search has barely budged, partly because of how Microsoft handled its high profile rollout. So the last part of that is important. Microsoft made people use the edge browser in order to use this functionality. I think that was a disastrous strategic decision because if they made Bing chat available on all browsers, you know, people aren't going to go out and download a specific browser in order to be able to

use your technology. I mean, that's just like the high point of arrogance. They're going to want to be able to use it anywhere. That's sort of like what's made software work in the past and what's made websites take off. And people having to go to Edge, I have Edge downloaded on my machine just so I could use BingChat, huge stumbling block. And, you know, I think that obviously contributed to the reason why people didn't want to use this Bing alternative to Google. But the other side of it is, you know, and I think this has sort of been a running theme here, is you got to, I mean, I have to admit, right, as someone's so taken by this functionality to

begin with, that using generative AI for every search query just doesn't make sense. Like it will make sense in certain areas, but more often than not. regular search is much better. And I've really seen this as Google has released its generative search, which I have enabled through labs. And I immediately scroll past it and just go to the blue links. And that's sort of like, I mean, it's a real revelation, I think, and something that I'm definitely changing my position on. Yeah, I think on the rollout, I agree that it was incredibly clumsy, especially knowing who are the type of people that would actually start, who are going to be the early adopters of this technology. Exactly the kind of people like myself and many others

that probably have tricked out their browser and spent years putting the exact Chrome extensions, which I do, so I can like enable every specific workflow and being super intense about it. So I'm not going to switch browsers. And I just same thing. I downloaded the edge browser and then occasionally would open up Bing and try to do something. And it was good. But I think that I agree that was a huge mistake. I do think I am. starting to see on the search side in terms of generative capabilities, I think it's still as promising as the initial kind of view of this, but I don't think it's there yet. But what I think could be interesting is like, I have started to certain things, again, even now recipes and things,

like instead, which are the like worst search experience on the internet where you get, you know, and I'm sure everyone, many of you are familiar, you get like an entire story that, shows a ton of ads. You have to scroll way down to get down to the actual recipe versus now it'll be like, chat GPT, give me four variations on this recipe. And it's clean, it's good, it's nice, and it works. What I think gets interesting is to me Reddit is still kind of the the massive potential here because almost all of my product searching, pop culture searching, is on Google, but I will search, you know, X Reddit, a pair of running shoes, Reddit, new movie, Oppenheimer, Reddit, and then I'll just go to Reddit. I don't start necessarily at

Reddit.com, but I start and I go to Reddit, and that's where I find my information. To me, one of the most important things Reddit did early on was cut off access to their API. And there's a huge backlash, and they risk their entire community. But to me, it was an important strategic decision because what they were doing was stopping training on their data for any generative competitor because they know they have probably the richest data set in terms of search query data, in terms of bringing together information. I believe Stack Overflow has claimed they're trying to do something similar because they've been in a lot of trouble around, you know, developer answers and queries and stuff like that.

So I think it's really interesting this kind of vertical. of search is almost to me a bigger threat than Microsoft's Bing suddenly taking over, or Google's generative search product itself, becoming the only place to be. I need to do a timeout for productivity tips here, because you mentioned that you had your browser tricked out with a bunch of different extensions, and I just became instantly curious what extensions you're using to make your life easier. Oh, like Notions web clipper, amazing for, again, recipes. And I have like my own little database there for music.

What do you do with the web? So you basically will pick different like paragraphs and stuff and it will send it to like essential database. It takes the whole page, images, text, clean, very clean, very nice. Read wise is honestly for anyone, did you ever use Instapaper? Yes. I mean, I use Pocket. No.

Okay, use pocket. So read-wise, I've used for a long time. It was originally like a highlight manager. You could import all your Kindle highlights. They have the best read it later. Like I had used the pocket in the past. I love it. It's, and it's, it's a super clean pocket to me, not to get too nerdy on read later for, for the listeners out there, but pocket kind of like was trying to push too hard into the recommendation side. Like we're going to become, as much of a recommender as a pure utility for you to read things later. Readwise is purely around, and it's readwise.io. It's purely about reading stuff later, being really good at parsing text,

actually pulling it off, highlighting, taking notes, all these kind of things. So, as you can tell, stuff around reading and saving information for later is a big part of it. No save to Kindle. Do you, Kindle? I read books on a Kindle But I I don't I've tried like doing PDFs on a Kindle But again, the read-wise reader actually can handle PDFs pretty well

Okay Actually do you read on a Kindle? I do Okay Wait but do we Are you listening to audio books now? I know this conversation has come up On Kindle

Or just in general Or anywhere yeah I've tried but no not really Yeah I like to read I tried on my vacation. I tried on my vacation. Audio books, just, they don't do it for me. Well, I listened to a very well-written audiobook, and I got annoyed because I wanted to enjoy that through reading.

I was like, this would be an amazing experience reading it and it's cheapened listening to it. But to those who enjoy the spoken word, I want to say, keep going. I appreciate you to all who listen to the show. Podcasts, I think, are a different category. Podcasts are a different category. But to all always day one audiobook listeners, thank you. I spent three days in the studio trying to record that. I'm recording that and completely lost my voice by the end.

But it was a very cool experience. All right. On the other side of this break, we're going to talk about Nvidia earnings. We're also going to discuss this new profile of Elon Musk and the New Yorker and maybe a few other topics if we have time for it. All right. Thanks for listening. We'll be back right after this.

Hey, everyone. Let me tell you about the Hustle Daily Show, a podcast filled with business, tech news, and original stories to keep you in the loop on what's trending. ending. More than 2 million professionals read The Hustle's daily email for its irreverent and informative takes on business and tech news. Now, they have a daily podcast called The Hustle Daily Show, where their team of writers break down the biggest business headlines in 15 minutes or less and explain why you should care about them. So, search for The Hustle Daily Show and your favorite podcast app, like the one you're using right now. And we're back here on Big Technology

podcast. Ron John Roy of Margins is here with us. Ron John, what did you make of Nvidia earnings this week, I feel like it's become a meme that every time in video reports, by the way, for those lists, I mean, I'm sure you know, but like this is the technology that underpins a lot of this AI momentum. And, you know, we can see that real investments are being made. But what did you make of the earnings? It's like every time they report, it seems like they just blow away expectations, even though those expectations are inflated. Yeah, I think the earnings unquestionably were good. You know, they, you know, they crushed it. You know, they crushed it, they had $2.70.70 earnings per share and it was expected $2.9.

Revenue was at $13.51 billion, whereas expected at $11.22. It's an unbelievable beat. Again, over $2 billion in revenue beat. I mean, that honestly, it blew my mind. Like how you, and as you said, the expectations are already severely inflated and they're still crushing them. then you start getting to a point one of the things that came up that was really interesting to me and i don't know if it's a good or bad thing for them they issued the fifth largest share buyback

of the year 25 billion dollars 20 25 billion dollars it was the fifth largest of the year this is when your stock is already up 220 percent on the year it's trading 45 times forward earnings when the average in the NASDAQ is 19 times. Like, you have the most expensive stock ever short of a meme stock. Right. But it's a great company. Is it, like, why are you buying back your shares? Do you really, as a management team, know something or believe something that it's still

severely undervalued? That, to me, made me almost more interested in these earnings. And do they, you know, like, are they really that confident in where this can, could continue to go? Or do they have to do that as a signal to investors who might start to think that it's a bubble as well? I mean, clearly it's the market cap and the revenue. I mean, the revenue is impressive, but the market cap is out of control in terms of what this company is doing, revenue-wise. That's a good point, actually. And I didn't think about it that way.

And it is interesting because I'm trying to think of an example. So again, like, I mean, it's a $1.1 trillion company, $25 billion to keep that valuation where it is is actually relatively small investment. Yep. No, no, no, you're right. You're right. That's actually interesting. And again, it's an authorization of a share buyback. It's not actually the buyback at that moment. So it's all my, typically, again, a share buyback, the signal, when it's done from a signaling perspective is if it's a beaten down stock and management wants to show that we're, we're confident in this. We think this is undervalued. It's almost never you would see it in that case.

But that's a very interesting point that maybe at a certain level, when a stock is so expensive, management has to show it's expensive, but we believe there's a long way to go. I was going to say it feels very Zerpey, but we're not in ZERP anymore. Well, actually, hold on this topic of ZERP, because this came up in another discussion I was having because someone was like, what do you think of all this AI hype? Because interest rates are, you know, at 5% or whatever. Like, it's, I thought you were saying that all this bubblishness goes away once we cross one or two percent, once we move away from that zero bound. The thing that's interesting to me is I actually think there's a genuine, the difference is there's a real story here. Again, I am a strong believer in the long term potential for AI and generative AI and these things are going to be transformative. So to me, this kind of like concentrated explosion of interest and excitement makes sense.

Like this is a rational bubble to me. Right. And I think it's probably going to be a bubble. And as we talked about, I think there's going to be some cold water and some downturn on all these inflate evaluations. But overall, it's a rational bubble versus a GameStop or a NFT, JPEG or whatever else we don't talk about anymore. Yeah. And I think it's worth talking about. why invidia has been able, has been the company that has bubbled in this way.

And it's just a mid-so the New York Times had a recent piece about the moat that Nvidia has around its technology. And there's a couple really interesting things about it. So first of all, the technology is just so far advanced beyond its competitors. This quote really stuck with me. Customers will wait 18 months to buy an Nvidia system rather than buy an available off-the-shelf chip from either a startup or another competitor. And I was just like, wow.

like we rarely see tech leads especially in like we were used to chips and hardware being commoditized but we rarely see such head starts so ronan can you explain exactly what about the technology has given invidia this unbelievable moat that i mean to wait 18 first of all i don't know if that's really true i mean if you wait 18 months for the chip like what are you going to do fold by the time like your your next round is coming but you know even the sentiment there is like fairly fascinating No, to me, what was really interesting about this article in the Times about their moat was it's essentially their community of AI programmers, which is kind of cool because like when you think about hardware is typically something that I never would think is would have a competitive edge purely or like primarily based on community. You know, you hear about again in many other areas or even social media and all these things. It always starts with community. and whoever owns the community. Even Facebook, like, that was the people choosing to develop apps on that platform.

That's what makes it a winner, Apple and iOS and all these kind of things. But it's they own the AI programming community. And then the article dives into, the more people develop on it, the more they can quickly scale their hardware to make it more powerful, which makes people want to develop on it more. And again, that amazing flywheel of software meeting hardware, meeting software, meeting hardware, and also everyone from a pure brand perspective, everyone who knows what they're doing and is in this community wants to build on it. Right. Yeah, this is a very fascinating line from the story that Nvidia has offered customers access

to specialized computers, computing services, and other tools of their emerging trade that's turned Nvidia for all intents and purposes into a one-stop. up shop for AI development. I mean, that really gets at what you're saying. Yeah. And I think like there is there is one another point in this where Jensen Huang, the CEO, had mentioned how in 2012 there was this massive breakthrough. It was ImageNet, which was in research where it was essentially the first time like it was proven computer vision could work, that a computer could reliably and confidently recognize a cat. And this one thing that's crazy for me is I actually remember,

and we've talked about this in the past New York meetup, a New York Tech meetup. I remember, so there's a startup called Clarify, like AI at the end. The founder, this is years ago. This was around 2013, I think, was talking about he would won some competition image net that their Clarify algorithms could recognize objects in a better way. And I remember sitting and being like, oh, that's kind of cool. And having no idea. That was the start of a revolution.

That was the, and this is one of those moments where, because when I was reading this article and they were talking about that 2012 image network and realizing, you know, in technology, when you have these moments where you're sitting there and you hear or see something that you have no idea just how transformative that thing or that moment's going to be, this really brought me back to there 10 or 11 years ago. And also reminded me, it was only. 11 years ago that a computer could start to recognize a cat and then like rely where we are now and look at where we are now. Okay there's something that that also uh poked my antennas up that I wanted to

speak with you about because it seems like we've there's a number of tech companies that do this where they they'll invest in startups and then have those that money used to buy their technology we've talked about it with Palantier for instance actually invidia is doing it as well so inflection AI which announced this is from the New York Times story which announced 1.3 billion in funding from invidia and others in June, and the money was used to finance the purchase of 22,000 H-100 chips, which are the Nvidia chips that are used to build AI. And this is another part of the story. NVIDIA has also directed cash and scarce H-100s to upstart cloud services like CoreWeave

that allows companies to rent time on computers rather than buying their own. And CoreWeave used, sorry, owns more than 45,000 Nvidia chips and raised two, point three billion in debt this month to help buy more is this you know so we've talked about in some cases this is kind of shady financial manipulation where does this sort of fall on the scale of like good business practice to like or i mean obviously their chips are like totally in demand so you know less like trying to prop up numbers but i'm curious like where do you where does this fall for you in terms of like how to view this business practice well i think and the grander scale realization of this is Microsoft and Open AI. Microsoft, quote, unquote, invests in

Open AI. The valuation skyrockets, but then a lot of that investment is in cloud credits, which Open AI then uses on Azure. And we've talked about this. What was interesting for me this week is on the topic of Hugging Face, which is actually a very cool, more open source community-driven platform to access generative AI models. They raised money at $4.5 billion, and I believe it was like Google and Invidia, or it's a bunch of the big tech companies were involved in it and potentially stand to benefit from it for their services. Roger McNamey had tweeted about, during the internet bubble, it was common practice for

industry giants to fund startups and then spend the proceeds on services from the giants. it can be illegal the management of home store went to prison the scale with AI is a hundred times is great extremely fishy um so i was like kind of fascinated by this so i was looking up so there's this company home store and the guy in like 2001 2002 was found guilty it was called a round trip scheme and it was the same thing it was you know uh taking an investment money to spend on other companies that were also being in that they were also investing in for services that they did not necessarily need and a dude went to prison for a quote unquote round trip scheme so it made me because to me it's always felt a little uneasy round trippy round trippy yeah but but on the other

side like it's you know it's standard practice any investor in their portfolio is always going to kind of like, hey, you should maybe check out this other company that I've invested in. They have something I think would be good for you. Like network effects like that are almost like part of every pitch of a BC, I would imagine. So like, so in general in investing, it's just human nature, it's standard practice. But when it becomes part of the investment, even with Microsoft and Open AI, it's just, it definitely starts to feel different, uneasy. I don't want to say illegal.

Like, it's just, it's, it's, it's weird to try to process. I just think that, I mean, so the one thing I'll say here is, like, there is demand for these chips. Like, people are falling over themselves to buy NVIDIA chips. So, like, in video, like, financing companies with, by, you know, giving them some of the chips or giving them the money to buy the chips doesn't seem, like, as, like, alarming. But it is something to watch on, watch, I mean, in one way, in one hand, if you have this sort of scarce resource, which are the, these chips, and you can trade them for equity in promising companies. You do that. So, but it is, it'll be interesting to watch. This might not have like a very clean ending is all I have to say. I mean, there's even one part of the story that talks about how invidia is a company that's not

afraid to bet against its customers. So this is all going to get very interesting and very intertwined in ways that I think, you know, in the next couple of years, we're going to start to see, you know, start to see that messiness come out. Yeah, I think. The simplest thing is, again, all this stuff works when it works. As long as every demand is going up, even on Microsoft and Open AI, it makes total sense. Like, what is Open AI spending its money on? Azure. And processing.

No, no, but they have to spend it somewhere. For sure. The entire, it's like a car company, it's all the parts of a car. There's going to be the majority of the cost. Like, it's, that is their cost. So to receive investment from Microsoft and then get preferential treatment on cloud and compute, it all is so logical and makes sense in that way. And as we talked about in the past, like, how do you recognize cloud revenue when it's your own money kind of in and out of another investment?

Is that different than cloud revenue from another type of customer? This is the kind of stuff that just from boring accounting feels like it's going to need to have some, more answers to it versus what's out there now? Very briefly, since we're speaking about chips, we should talk about how Arm is filed to go public, and it's soft banks, and it's probably going to be the biggest IPO in 2023. They are a designer of chip circuits,

so they're not actually a manufacturer, but they're a very interesting company to watch, and they're going to, like, really, they're a crucial company in SoftBanks portfolio. Did you see anything in the announcement of its decision to go public that sort of made you think, oh, this is interesting or anything we should be paying attention to? The most interesting part of it for me is everything looks super clean and normal and good. And for a soft bank IPO, for a soft bank related IPO, that was the part that stood out the most to me.

And again, it's something that it's a company that's positioned correctly. It's here at the right time. They essentially licensed the design, as you said, to other firms to actually kind of like help power the AI revolution. And it's interesting to me because what was the Masayoshi Sun thesis? It was AI is going to take over the world and we want to be the kind of like, but what's so fascinating to me is he was essentially correct on that. circa 2015-16 when all this we were dog and wag and all this other stuff was happening and clearly this is going to looks like it could be one of the only real bets on powering that entire revolution that that looks like it can make a good amount of money for them in a reasonable way

now they do a lot of work with smartphones and this sort of struck me from the report that there's a real slowdown going on with smartphones and they said there was a set 7.8% decline in smartphone sales in the second quarter. What do you make of that? I mean, is that just people like holding onto their COVID days upgrades and not buying new phones? Or is this a cyclical slowdown? No, what I make of that is I bought every iPhone after waiting in line for the first iPhone until the 12. I still have the 12. I have not upgraded since. I was talking about this the other day, but I got the Apple Watch Ultra. I got like I'm like other parts of the Apple ecosystem are trapping me in and keeping me happy.

So I think I, but I was talking about this the other day. It's interesting. You don't need to upgrade as often. Yeah, like what's the new thing? Camera is a little better, good. Obviously, my battery is kind of dying a little bit now and faded like degrading. But overall, smartphone innovation in itself is really, I don't know, like what do you have now? I have.

So I am counting down the days for the iPhone 15 to come out because I have an iPhone 10. Oh, wow. Okay, the screen is smashed. There's a bright line going down the left side of it. It's slow. The battery dies. I plug it into the car, for instance, and it overheats and I have to put it in the air conditioner. So it really is time to upgrade.

I cannot wait for the 15. And the 15 already has also has USBC charging, which is why I didn't go for the 14. I wanted to wait for the new charging port. Thanks to European regulators. Thank you, European regulators. We always like to shout out the EU's competition committee. Thank you guys and gals. But yeah, I do think that that's going to be nice to have USBC charging. But yeah, I agree. Once I get this 15, I don't want to replace it for four years. Yeah, no, I think smartphones to me, I don't know, maybe there's going to be some amazing thing that comes out or some kind of innovation or maybe again like

two years from now you're going to need such a powerful chip on your phone to do on board compute and processing for generative AI and other types of artificial intelligence that could happen but as of now having a slightly better camera that takes kind of like slightly better pictures and video it's not that exciting but the European style charging port I mean but a USB I would be embarrassed to walk around in public with a regular lightning cable. A lightning cable. No, I'm just, I'm just joking. Let's round it out with Elon Musk, this Elon Musk profile in the New Yorker. A lot of it, it's by Ron Farrow. Okay, worth reading. And a lot of it is build up about how, like, Elon has, like, sort of influenced the war effort in Ukraine. And we talked about this a little bit last week. And the point, and it's like thousands of words. And then it's like, points to this, it sort of builds up to this paragraph. And it says, and Farrow writes,

in the past 20 years against a backdrop of crumbling infrastructure and declining trust in institutions, Musk has sought out business opportunities in crucial areas where, after decades of privatization, the state has receded. The government is now reliant on him, but struggles to respond to his risk-taking brinksmanship and caprice. I mean, that to me says it all, which is that we have a government that has been that has stood still that has not made the right investments in infrastructure and some of that is by design but other that is by incompetence and it just leaves the door open for musk to be sort of doing this type of things and these type of things and so like if you have a whole whole piece complaining about like his

influence over the war operation in ukraine like there's only one reason why that's happening and that's because you know for all the defense spending that we've done And all of this boastfulness about how our infrastructure, it's infrastructure week every week, like we've actually failed in the U.S. and globally in certain ways to build the infrastructure necessary, to sort of be able to live in this current world. And that's where you end up going to the private sector. And by the way, I think that all the people complaining about Musk in this piece would much rather have his Starlink internet infrastructure that's basically enabled Ukraine to even do

battle with Russia than not have it. So that was my takeaway from the piece. My takeaway was a bit different. It was more, Ronan Farrow is publishing. It's going to be thousands, if not more than 10,000 words. I was hoping for something a bit more interesting, a bit more exclusive. And we can definitely get into the kind of public, private, because I think that is the most important part of it, the story. But to me, a lot of it, It felt like, and this is someone who has very closely followed Elon Musk for years, has written about him for years, been on TV talking about him. Like, a lot of the stuff I get is stuff that felt so common, like common knowledge for

me that I was a bit surprised or I was a bit like, you know, disappointed that that was what made it into this story. Because in some ways, it felt like someone who had never probably still only thought, the way it even reads is it's like Elon Musk you think is a hero but really here's what's going on which at this point I felt like that kind of narrative for me is not as interesting I do think the Starlink info it's it's out there but but let's not forget why did Elon Musk at the beginning the entire Ukraine thing it was such a weird moment because it was such marketing on both sides. It was product marketing for Starlink. It was the getting the Ukrainian generals. Do you

remember those tweets and pictures of like guys on like generals unloading satellite dishes out of their? Yeah, but it was more than marketing. Like it was critical to their war effort. Yeah, but but is it the only way to actually make this happen? Like, or were they perfectly positioned? Because let's not forget, Elon Musk, again, in all his brilliance, like, no Zelen is looking for anything to grab a good amount of attention at that moment. So getting a Starlink satellite and a photo of a truck, you know is going to bring more attention to the war effort separate from the actual infrastructure and capacity side. You can make Starlink. Wait, for real. I mean, Russia takes out their internet, okay, to begin with. How else is Ukraine going to get Internet?

You tell me. I don't know. I don't know. I don't think there is another way. Like, this is actual. I mean, that's why it's so interesting is that this is mission critical infrastructure for ukra no no so this is the part i will say from the ron and fair piece and the and i agree the way he paints this picture is actually making elon musk look more capable and important and powerful than anyone ever could have imagined okay i agree is that that's the way it's the pictures that's painted is that is that his ownership of low earth orbit satellites is going to be absolutely critical to all anything involving satellites, which is going to be a good amount of the world

and the economy coming up. So, yeah, I agree. The way the picture is painted in the story, he makes Elon Musk not look like a bad guy in my mind or not look, like, as you said, it's more a critique of 30 years of privatization and deregulation than Elon Musk, which maybe what he was going for.

And look, I think that, like, just to bring up full circle, wouldn't it be nice if we had a serious government and serious politicians? I mean that this story came out in the backdrop of, you know, the Republican primary debate, which I know you didn't watch, but I did. Well, I caught pit bits of bits of it and the highlights. And it was the most unsurious discussion about politics and policy I've seen in a long time. And people can criticize me for both sides in it. But, like, I don't think that our politics broadly is serious about policy issues these days.

And we don't, we do not have, like, I thought for a Republican primary, wouldn't it be nice if we had one candidate who actually focused on restoring infrastructure, an actual economic policy, which has been their bread and butter that's going to put the country in a better position and the world in a better position? And instead, we get, you know, Vivek Ramoswamy screaming that climate is a hoax. Well, half of the United States is in the middle of a smoke alert and don't even get me started on Canada, right? Like, it's from these wildfires. I mean, I just drove through them this past week, you know, starting in Seattle, making my way down the Oregon coast into California. And it's amazing how so much of the West Coast is choked in smoke. So, like, to spend your time talking about how climate is a hoax and really not focusing on economic policy,

not focusing on the fact that we do need infrastructure or else we're going to put ourselves at a strategic disadvantaged. It just seemed like a complete shame to me. Yeah, but that's the problem. To me, it's not a lack of seriousness in policymaking and government in general. Again, we just hit the one year anniversary of the Inflation Reduction Act. And in terms of combating climate change, it was like the most significant piece of legislation and actual policy that's ever passed. And even at a micro level, talking to people who are actually retrofitting their house to get tax benefits, like real policy action is taking place. It's just less interesting than Vivek Ramoswami's tennis swing or I don't know

if you saw him rapping. Oh, I did see that. Lose yourself. Yeah, no. See, I'm, I'm guilty as well. I watched that on repeat and that's what you talk about. But it's, to me, policy is not interesting on cable news. It's just, it is what it is. It's what Donald Trump learned to exploit. Even the front page of the New York Times the other day was all Donald Trump stories and zero policy or Joe Biden stories when he got his mugshot and the debate and everything. Well, that is front page news. I mean, when the former president of the United States. Yeah, yeah. Okay, that's fair. But not participating in the debate is not necessarily interesting. By the way, you know what he's doing. There was also a full story in the times about just the symbolic nature of Trump's mugshot.

And I started reading it and I was just like, you're going to do 800 words on this, on the, like an art critique of the mugshot. Like, come on. Come on. But I think the broader point that I'm trying to make here is, yes, by the way, I think there were good parts of the inflation reduction act, even though the name was kind of hilarious. speaking of marketing, but broadly, I think we've had policy leaders leave us in a place where we do have these huge gaps in areas like infrastructure, which Elon Musk has stepped into fill, and then people complain that Elon Musk has filled them. So that to me is like the real issue.

I will push back again. It's not that this gap existed and he stepped in to fill it. It's he also helps create the gaps in the electric vehicle industry. to like avoid any interaction or engagement with regulators and to you know and to build your entire company that way and look at regulation as a nuisance and kind of position everything around that maybe we would have again waymo and crews actually working directly with cities and regulators and now and actually i'll be in san francisco next week and i am more excited than anything to ride in a waymo but there are companies that show that you actually buy

cooperating with regulators and doing things in a constructive manner, we have self-driving cars going around roads right now versus Tesla, which took the opposite route, and then kind of like both fed into the deregulation and privatization, but also helped drive that climate over the last number of years as well. Yeah, I thought you were going to say that, you know, we can't take away the fact that Ilamos did basically work to invent this in this, you know, Starlink system. But you went a different direction, but I think there's both, both of that. But I, I had, like, I'm going to give the Waymo CEO who was on here last week a little bit of credit.

No, no, a crew CEO. No, I agree. But like, look what's happening to them. They're driving into wet concrete. No, no. I think, no, I think, no, you're, I think you're totally right there. And I do think, so I think that, that a lot of the criticism of Musk is indeed merited, but I think that, like, some of it can be so myopic that it sort of misses the detail.

And again, that's sort of what we're here for is we want to talk about these stories with nuance that it's way too often missing in the rest of the discourse. So we'll get things, I mean, I'll admit, like I'll get things wrong sometimes, get things right sometimes, but that's the aim here. So to the person who sort of complained about me not threading because it was a support for, support for Elon Musk. I'm back on threads. It's not like that. but are you are you back on threads i was threading this week yeah all right all right i'm still using but the thing is the thing is by the way first of all i appreciate any feedback that we get so thanks for that i guess like the point the point of that i'm trying to make is that like we're also

like at a moment where like you're on it's it's become less about the product and more about like the tribe right like are you on Elon must tribe you tweet are you on the not Elon must tribe you thread are you in like the Donald Trump camp you truth social and rumble the rum or rumble you it's just like it's another shame is that well at least social media technology maybe that was it was always destined to go that way but it's become more about like who you support and less about like the product experience but I digress anyway Ron John thank you so much for joining plenty to talk about we cover our two favorite topics genera i and Elon Musk will be back again next week to break down the week's news. And appreciate you coming on, as always.

All right. See you next week. All right. Thanks, everybody, for listening again. Colin Murdoch, I know I promised him this week. We had a scheduling changeup, so he'll be on on Wednesday, followed by Sridor Ramaswami coming on the following Wednesday. Ron John and I will be back on Friday to break down the week's news. Who knows, maybe there'll be another debate or we can finally get to the story I've been really wanting to talk about, which is Yvgeny Progoshin. Fascinating story, by the way. Maybe we talk about it next week.

But in the meantime, we'll be breaking down all the tech news for you. So thanks again for listening. And we hope to see you next time on Big Technology Podcast.