May 31, 2024 ยท 2024 #19. Read the transcript grouped by speaker, inspect word-level timecodes, and optionally turn subtitles on for direct video playback
Edit labels for this show, save them in this browser, or download a JSON override for the production folder.
Transcript Playback
Stand Back, Think Big, Dig Deep
Human Transcript
Timed transcript
Blocks are grouped by speaker for readability. Expand a block to inspect word-level timing.
Speaker 3
Hello, everybody. New month, new week, new technology with my old friend, Keith Teer. That's the old feature of our show. This week's That Was The Week newsletter says, Stand back, think big, dig deep. It's a defense of utopia and to kick off, Keith, We're going to play a little bit of your music. So let's rock.
It's really an honorable mention to Rohit Krishnan, who wrote the essay of the week this week is called With a Utopia. Rohit's a British, I think he's British born, big thinker who stands back and digs deep. And this essay is about the history of um utopian thinking and utopian experiments mostly in the 1800s and in the early 1900s leading up to the first world war and it's an essay that reflects on the fact that macro thinking like that is now frowned upon um if anything it's feared and criticized for being um i don't know uh overarching and kind of authoritarian, whereas micro-thinking has taken over. That's his thesis. He blames a lot of it on technology. He thinks technology is delivered on some of human needs, therefore people don't think big anymore. But he's kind of lamenting the loss of big thinking.
Yeah, and your editorial this week focuses on a defense of his peace with a utopia. Krishnan talks quite a lot about Robert Owen. Why is Owen relevant today, the late 18th, early 19th century social reformer and unashamed utopian?
Well, I mean, he's not the only one, but he's... He brought together an understanding of the Enlightenment thinking about progress with, if you will, thinking about equality and the human condition. And often the two are separated. Like when we talk about AI today, we often talk about the business of AI, but we often don't talk about the human condition in the context of AI. So Owen brought those two things together and began to experiment with different ways of organizing societies, mostly failures, by the way, almost every utopian project that tried to build cities based on a way of living failed, mostly due to disagreements between the inhabitants. So it wasn't successful, but it was at least thinking about the human condition.
Yeah, and Krishnan in his, it is a very good piece, talks about the, we need the spirit of what he calls pioneering ambition. Of course, we're fearful of that. We've been through so many cycles where we've been promised ambition. to borrow Peter Thiel, flying cars, and all we got were 160 characters. And we've also been promised utopias by people like Mark Zuckerberg that haven't been realized. What needs to change then, according to Krishnan or yourself?
He doesn't really say what needs to change, other than he'd like to see more time spent on bringing together thinking about technical progress and social progress in the same conversation. And the reason I seized on it is it's very clear to me that in the history of knowledge and thought, when you go from the pre-First World War era, and you could even call it the pre-Russian Revolution era, to the period after the Second World War, you move from social engineering on the big scale to postmodernist microengineering. And any conversation about dramatic social change really doesn't exist anymore. I'm not going to blame the woke people for this, Keith. I'm not going to blame the woke people, but the woke people are symptomatic of it.
I didn't mean to bring woke into this. It was interesting. Another piece that sort of perhaps could incorporate Krishna's piece is a piece you add to the newsletter about Chinese scientists developing a cure for diabetes. Technology is accomplishing, always has done, of course, medically. But should we be more open to some of the some of the breakthroughs that are happening, maybe less in Silicon Valley and more in China?
Well, the human race will be open to them because even if the Chinese don't commercialize these things, Western pharmaceutical companies will be all over the scientists and the patent owners and the inventors to commercialize them. So, yeah, we should definitely pay attention. If you look into the detail of what they've done, it's all really to do with biotech. It's understanding at a very deep level how the human body works and how viruses work and how illnesses work and being able to reprogram human beings to remove things like diabetes. That is technology. And you can definitely look forward to a time when many of the diseases that kill us today won't. And that's a great symptom. I put it in mainly because Rohit's essay is really calling for optimism on tech, but optimism on social progress as well.
Yeah, and you and I, Keith, we don't always agree on everything. We disagree on a lot of stuff, but I couldn't agree with you more on this anti-Chinese sentiment, this... trade warfare going on, not just on the right, but also on the left, hostility to TikTok. So I think the more we recognize breakthroughs in China, obviously good for Chinese tech and Chinese medicine, Chinese state, but also for everyone, it's important. What about AI? There was an interesting piece in the New York Times this week that I forwarded you that you included in the letter this week. If AI can do your job, maybe it can also replace your CEO. We've had lots of promises about the flattening democratic potential of tech. Could AI, do you think, replace CEOs and thereby replace these stratospheric incomes and differences between ordinary workers and CEOs in America?
You know, the answer has to be yes, doesn't it? How could that not be true? But I think it's worth talking about what a CEO does, because it isn't everything that a CEO does that can be automated. But a CEO has to do a lot of kind of administrative work. Like I'm sitting here now, I almost shared it with you, but I've got a an Excel file here. Maybe I can. Let me just see if it will let me do that. Yeah. Click on that, Andrew. This is part of my job as a CEO. I've got to fill in for my auditors who controls the IT systems at Singularank. And between you and me, this is not my favorite job. And it's a job AI could easily do. So I think there's many things a CEO does that can be automated by AI. I don't think you'll ever get rid of a leader in a company that's responsible for setting strategy, telling the story, possibly even execution at a certain level and management. But execution and management is going into the domain that AI could... That's not a CEO, that's a COO.
Another piece that you linked to today is about Sequoia now investing massively in what they call Musk world. And Musk, love him or hate him, is enormously influential. AI is never going to replace an Elon Musk, is it?
And more and more pieces about how he's talking to Trump and maybe even supporting him in the next election. So he's clearly a very controversial, not everyone loves him. You love him, I'm more ambivalent.
I wouldn't say the word love. You admire him. I recognize what he adds. There are times when I... really disagree with him strongly. He's just tweeted something about Trump before our show started more or less saying that the trial was a political trial or a trivial trial that shouldn't have been held.
Oh, and by the way, Keith, speaking of Musk, one person we know doesn't like him. And Trump is our mutual friend, Steve Gilmore. I do have an apology to make to Steve if he's watching. If he's still watching, I know he threatened never to watch again. Last week, I noted that he was a fan of Paul McCartney. And the reverse is actually true. He's not a fan of Paul McCartney. He loves John Lennon. So let's be clear on that one. Is that good enough apology, Keith?
No, George Harrison comes second. And I think McCartney.
Words and timings
No,GeorgeHarrisoncomessecond.AndIthinkMcCartney.
Speaker 3
Now we've got at least one loyal viewer. It's Steve. So hopefully we haven't lost him. So back to Musk and Sequoia. Is this a big deal, Keith, this news? Or is it just Sequoia's way of hedging their bets?
it's more of the same because Sequoia has backed Musk through many iterations over a long period of time. So, you know, PayPal, if you remember, going all the way back, Rulof Buta was CFO there. Rulof today runs Sequoia. And Musk and Rulof go back a long way.
But it's relationships built over a long period of time, just like you and me. I mean, we first met ages ago, decades ago. And you don't, those relationships, normally at least.
So back to this AI CEO issue, I take your point that AI can fill in forms, but that's not what a CEO does. That's why Musk can be the CEO of three or four companies or Jack Dorsey or Steve Jobs. You may not like them, but they are clearly geniuses.
Yeah, well, we should acknowledge there's different types of CEO. Some of them are more like COOs in how they operate, and they have a head of strategy doing strategy. Others are more like me, which is I hate micromanaging, and I will micromanage if I'm allowed to, but that would be a bad thing. Yeah, and so I'm more big picture and strategy setting and storytelling and money raising. They're all different.
Well, I agree with you. Yeah. We've got to find something to disagree with on this show. But I think Krishna's piece is an important one. I think he's right. I think that the zeitgeist has shifted too much the other way and that we do need to recognize, as Krishna says, the spirit of pioneering innovation and to begin to think deep, utopian thoughts. Anyway, back to. Reality, back to day-to-day Silicon Valley stuff. Salesforce had a bad week, Keith.
Benioff said we need 50% more pipeline. That's a terrible indictment of their own salespeople. Their stock is down substantially. Is Salesforce one of those companies that are just going to die out on the vine in our new AI age?
I think every company is threatened with that unless they innovate. And I wouldn't put Salesforce high up that list because I do think Mark Benioff is a great leader and thinker. He does stand back. Not that great.
Yeah, but that's because the current product set, he actually explains it in the call and it's in the article. The current product set is producing a pipeline of customers that converts at a certain rate. And the pipeline is only half the size it needs to be to make the numbers that they wanted to make. So they have what's called a top of the funnel problem, which means customers are not showing up to buy the products. Now, the answer to that is product innovation. AI is probably part of that. And I do think Benioff is an innovator. He buys what he can't build. And he's shown that time and time again. He bought Slack, for example.
Yeah. He's got competition from HubSpot now. HubSpot came from nowhere to have $700 million of annual revenue from CRM, which is a Salesforce core product. So he needs product innovation as well as
I mean, if you stumble in this economy, things are moving so fast. Can you stumble and survive? I know Adobe also had a bad week. These sort of mid-level or upper middle-level companies that used to be very powerful in Silicon Valley, pre-AI, the Salesforce's and the Adobe's, they're really in an existential doom loop unless they can reinvent themselves, aren't they? No company can stand still.
Yeah, well, you can see it best and most clearly because it's playing out in public at Google. I mean, theoretically, Google's... You say about Google, Keith, otherwise they'll get the police on you.
No, but just to characterize what you said up, Google's core product is a search engine with search results that generates advertising revenue. they have innovated. That's what I'm saying. They're having to change rapidly as well, and they are changing. And they have to change in a way that doesn't kill the goose that lays the groundwork.
They want to be both the old search company and the new... I mean, it's like what Microsoft had to deal with in the 90s. And this comes to, I think, another very good piece you linked to today from Alex Kantrowitz, one of your favourites there. about chaos coming for startups, ads, and online businesses as generative AI eats the web. The end of paying and optimizing for referred traffic, that's going to change everything, not just for Google, but for every company.
Absolutely will. The entire business model of the internet is based on links being clicked. You know, there is no money on the internet unless links get clicked, resulting in videos being watched, content being read, podcasts being listened to, and so on and so forth. So link clicks are everything. And in a world that moves to voice input and AI, you know, it's funny because the world comes around again. But in 1999, when I was running Real Names, I struck a deal with Phone.com. for real name keywords and the Press release for that deal said URLs will not be able to be spoken But keywords will and we're moving gonna move to natural language Inputs which was about 20 to 30 years too early But now we are with that really is happening and I think I think the web is meaning the narrow definition of the web, which is HTML pages, are not likely to be the primary feedback loop anymore. And that means all those business models for advertising threatened now this isn't going to happen fast it's just like you know the radio star wasn't killed by TV neither will the web be killed by AI but the balance of
attention is going to shift and I have a I sent you another link which I don't think you included but I thought was a good piece coming back to this it was It was in today's journal about... Oh, yeah.
Hold on. Let me just get it up on the screen. Because it speaks of the hype. I know a couple of weeks ago when I wasn't here, you had the hype about, or you had a show about OpenAI's new...
A lesson in AI hype from the journal. about GPT-4.0. I used it. I think it doesn't strike me as being that revolutionary. So there's a lot of this stuff, that promising stuff. I know you think that this new version of OpenAI is going to change everything, but at the moment, it's still just promise, isn't it?
Yeah. Now, why did they do it? They did it because the Google announcements were the following day and they wanted to get ahead of the Google announcements. But the actual product took a couple of weeks to start rolling out. It has begun to roll out now. And it does have some of the features that were demoed, but not all of them. So the time is exactly right. I still believe they will ship all those features. And we will end up getting them.
Yeah, it's like in the long run, we're all dead, Keith. So what does it mean for companies like Perplexity that are trying to reinvent search? What's their business model?
Well, they just announced something this morning, which I didn't put in the newsletter because it was too late. But they announced that search results are now going to be shareable pages. So everything that comes in a search... Well, so almost like a PDF, but... HTML version so that if I do a search for chat GPT hype, it's going to bring back really good results, well designed and everything. And I can say share with Andrew Keen and you'll get the same search results page that I just got. So they're trying to introduce social sharing with AI results. How do they monetize then? Well, presumably they'll put ads on the pages. Well, then it's not that different. So it wouldn't be that different, exactly. Perplexity, remember, is a search engine, so it needs a web page with a result set. So it probably is as threatened by this direction as Google is, and even more so because it's a smaller company with less revenue.
Everyone's having to think really fast on their feet. But on the other hand, when you think of Microsoft, people are still buying desktop publishing. I still pay an annual subscription for Word. The old economy isn't just going to go away. It's going to take 15, 20 years for it to finally disappear, isn't it? I mean, the old economy being, let's call it the Web2 economy of Google and Lynx.
I think that comes down to a kind of interesting question, which is human beings are always going to share. So then it comes down to what is the atomic unit of sharing? Is it a document? Is it a link? Is it a post? What is it that we're going to be sharing? And how will you render? You know, like video is rendered in YouTube for the most part these days or Instagram or TikTok. Text is rendered in... posts on social media, in newsletters. Word probably is no longer the main place words sit. Spreadsheets probably still rendered in Excel, but will that stay true? So I think that's the place the least change is happening. Like what is the atomic unit of sharing a particular type of content? It's more to do with the knowledge in those documents. That's where the main change is, how knowledge is being retrieved and shared.
And are you still, you've been very bullish this year on Altman and OpenAI. Do you still remain as bullish? Is he making all the right steps, the wrong steps, some right, some wrong? What do you make of what they've done? And you didn't put this in the newsletter this week. They've done a series of deals with media companies, including the Atlantic, who are very hostile
Yeah, well, those deals are where they're paying. The more interesting deals to me, which I did put in the newsletter, one with PwC and one with Apple, is where they are being paid. PwC is putting 100,000 of its employees on a platform that has ChatGPT. And Apple, I don't still know if this is true, but the very strong rumor is that OpenAI has won the deal with Apple to be the dominant AI on the next version of the OpenAI.
You put that in the information, and because I don't subscribe, I couldn't put that one up. So when it comes to PwC, what will the PwC team be able to see that ordinary OpenAI subscribers won't be able to?
They'll be able to train models on PwC's own data that's behind their firewall. And therefore, the knowledge that that chat GPT will have will be much more specific than the knowledge that the general one has. And that's going to be a trend.
No, it's still an LLM, but it's trained on specialist data. OpenAI gets paid whenever a third party uses its APIs to train a model on data that ChatGPT itself doesn't have access to. And so that's the revenue. It's already over $2 billion, by the way. It's probably going to $50 billion of revenue in the next two or three years.
Yeah. No, they're trouncing the competition. I mean, Microsoft probably is the smartest kid in the room because it owns a piece of everything. There is an underlying... challenge which is transforming open ai into a for-profit company which i think
they will eventually do and and then what is in terms of this potential imminent deal between apple and open ai who who is that More threatening to Microsoft or Apple? Microsoft or Google?
Google, by far. Microsoft will get some of the money that Apple pays OpenAI. Google will get none of it and will potentially lose some revenue from search that Apple pays them. In fact, it's the other way around. I mean, Google pays Apple, but it only pays it a fraction of the revenue it makes from Apple. That will go down. So Apple will probably keep Google as the default search engine, will make ChatGPT the default Siri, if you will. And over time, user behavior will either shift towards the Siri-like experience or stay on the web search experience. And whichever way it goes, Apple wins.
So these can exist in parallel, the OpenAI-powered Siri-like experience and the traditional Google search? Yes. And why wouldn't Google give Apple an open check, as they did with Search, in terms of incorporating their AI, which is as good, many people argue, as the open AI?
Well, I'm sure that if that hasn't already concluded, that conversation is still going on, or it's being concluded. The rumors suggest it's being concluded, and Apple chose open AI. But there'll probably be a time limit to that agreement. And it's always open to Google to try to get it back. I think, you know, having used them both, I do think OpenEye is vastly better than the Google current offering. But that won't be true forever.
That'll be a very interesting game to play out. Let's move on to... startup of the week, which I was particularly intrigued with, Lux, Keith. Not Paul Graham. We're going to come on to Paul Graham later. New from Lux, makers of Halide, is something called the Kino Pro Video.
Yeah, this is a bit of a geeky startup of the week. I love my iPhone. I love the video features. I use the Blackmagic camera on the iPhone, which is also a very good cinematic software to use the iPhone cameras. This Kino is a simplified version of that. It has all the power. but it's been presented in a much easier to use way. So I think that Kino Pro is 10 bucks, by the way, to buy it for life. And it changes your iPhone camera experience into much more of a cinematic experience with color. Whatever that means.
The bullet points are, this is a bit geeky, but it's using H.265. HEVC is the codec, which means much smaller files. It is using Apple ProRes, which means that it gets very good color rendition. And it's building in color grading, which you can actually turn on as you're filming. So you don't have to go through a step of color grading your footage. It's already baked in.
Yeah, Lux is an interesting company. I'm interested in Orion, which apparently turns an iPad into a portable HDMI monitor. So that's a company we need to look out for. And finally, he's back, Keith. I thought we'd got rid of him forever, but you managed to... infiltrate him back into X of the Week, your old friend and heartthrob Paul Graham. What did he X this week?
He decided to go on the record about whether Sam Altman was fired from Y Combinator or not, which has been an underlying assumption for the last few years, that Sam is basically a difficult guy and got fired, in quotes. And it turns out he kind of did get fired, but Paul Graham's trying to present that he didn't. I mean,
I don't think there's anything wrong with being a difficult... Everyone knows geniuses, whether it's Jobs or Musk or Altman, you've got to be difficult.
Yeah. So the facts are Altman announced he was going to be CEO of OpenAI. At that time, he was president and effectively CEO of Y Combinator. And he was told by the founders, the Graham family... husband and wife family that he'd got to choose one or the other. And he chose OpenAI. And so he left. And, you know, the comment stream there, some people say, well, so he was fired. And other people say, no, he had the option to stay, but chose not to.
A little bit different. Only that there are two choices and they have to make one. That is the common element, but that's true in all life. Do I need a girlfriend? You've got a wife. You don't need a girlfriend. Exactly. It's the same. Easy answer to that question, Andrew.