Mar 15, 2025 ยท 2025 #10. 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
Is Apple Intelligent?
Human Transcript
Timed transcript
Blocks are grouped by speaker for readability. Expand a block to inspect word-level timing.
Speaker 2
Hello, everybody. It is Saturday, March the 15th, 2025. Another That Was The Week show about the week in technology. We usually talk about the successes in big tech, especially when it comes to AI. But today we're talking about a potential failure, a failure of a $3.2 trillion company, at least as we speak. A company, small company, some of you may not have heard of it, called Apple. According to Keith, who asks in this week's editorial, if Apple is intelligent, the audience says no. And that lack of intelligence suggests that Apple might indeed be failing on the AI front. Lots of criticism from some of its biggest supporters. including John Gruber, O'Malek, and MG Siegler, perhaps the three biggest Apple fanboys, certainly in Silicon Valley. Keith, does all this furore over Apple this week over AI suggest that Apple is failing on the AI front?
Certainly, that's the meme. Apple... not only is failing, but is failing spectacularly, loudly, and dishonestly. That's the accusation. The dishonesty piece being that they announced a whole bunch of features last year at their developer conference in June. most of which were fake, as in... Oh,
Cook their demos. And then they announced this week that the features would not be coming this year, that they may come next year. And that has led everyone to throw the rattle out of the chair, as babies I want to do. And... Yeah, it seems to be a consensus. I take the opposite view, to be honest. Keith, you take the opposite view?
That's the first time that's ever happened. It's the first time. The opposite view meaning that they know exactly what they're doing and that they're a winner in AI or that people are wrong to criticize them for being dishonest?
No, I think people are right with the criticism and I think they're right to note the failure. So I don't disagree with that. I disagree that it matters. I think that, you know, Apple's iPhone is, for most people, the device that sits between them and all software. And on my iPhone, I've got... Certainly does with me.
Right. So Apple's in a great place. There's a lot of AI on my iPhone that Apple doesn't deliver to me. that is significantly better than the stuff that Apple does deliver to me, like Image Playground, which is widely criticised as a bit of a joke. I've never used it. There's way better image generators on my iPhone than Apple's. And so, ultimately, I don't look to Apple to be the deliverer of AI to me. I don't really care whether they have it or not, and I think that's probably true of most people. I do care to have the AI, but I don't need Apple to be the vendor.
Keith... To play devil's advocate, to your devil's advocacy, I mean, Apple theoretically were perfectly positioned to jump on the AI train. Maybe they didn't develop chat GPT, but they had Siri. Yeah. So why the big failure? Why didn't they just do a deal two or three years ago with, open AI and just incorporate their technology into Siri. Wouldn't that be a no brainer?
It would be a no brainer, but they didn't do that. Well, first of all, let's just understand what Siri is. Siri is really a natural language processing input engine. You can ask for things. And it can only give you answers if somebody has curated the answer to that specific question. So it's a curated question and answer engine. Not very smart, in other words. Not at all smart. They did announce that they're building OpenAI into Siri, and they have, by the way. That's one of the things they've delivered. So if you speak to Siri today on the latest operating systems, Siri will defer to OpenAI if it can't answer, and OpenAI will answer, as long as you've enabled it on your phone. So they have done that, and that is the wise thing to do.
So they do have... Apple intelligence is a real thing. At the time they announced it, they were lauded by these very same people as being really super clever not to go... Well,
they're always... I mean, by people like M.G. Siegler and John Gruber. I mean, if they're criticizing Apple, then there's a huge problem. But there will be, Keith, I was thinking about this in the context of your editorial... which questions the future or at least questions Apple's future, there are going to be losers in this game. And they're not always going to be the obvious losers, old media or old industry. I mean, some big tech company is going to lose here. And Apple is probably in the best or perhaps the worst position at the moment, isn't it?
Although there were, I mean, even with Microsoft, there were casualties. I mean, in a sense, Bill Gates, and then certainly Steve Ballmer. So I wonder how Tim Cook is coming out of this. What do guys like Siegler and Gruber say about Cook's responsibility? Who are they personally blaming?
they're not specific, but it, what the implication is, and there was an internal Apple meeting where they, uh, the leader of Siri had on all hands and disclosed the embarrassment, if you will, at this series of events.
Yeah, it wasn't really a show trial. It was that the team who've been building Apple intelligence were blindsided by the announcement that it wasn't going to come out. but they also knew it wasn't very good. And this internal meeting was a psychological therapy session for that team.
Confession, confessional time. So, I mean, if Steve Jobs, who certainly must be turning, spinning in his grave now, if Jobs was around, I know it's easy to ask this, but how would he have dealt with it differently from Cook, do you think?
Well, Gruber actually quotes Jobs at a very similar moment in Apple's history when MobileMe had been launched. If you remember MobileMe, today it's called iCloud.
That's right. And it was kind of stillborn. And Jobs did this meeting. uh where he he came into the room and said what is mobile me and somebody gave a perfectly good answer as to what it what it's meant to be and jobs said well why the doesn't it do that then yeah and he he canned uh the the head of the team was fired on the spot in the meeting and and so apple has this very long history of of not pre-launching things that don't work.
Yeah, and also Apple has, of course, a history of launching things, second generation, which then changed the world. I mean, the iPhone certainly wasn't the first or even the second generation of smartphones. So it is conceivable that they can catch up.
Oh, of course. And by the way, if you ask the question, are they right to not ship Apple intelligence this year, I don't think anyone would say that they're wrong because it apparently isn't very good. And it isn't Apple's DNA to be an AI company.
I take your point, but your point on the iPhone, that it's the intermediary between us and the network. I just waved my phone again at the camera. You're right, but Keith, you know this better than anyone. Nothing ever stands still. We're not going to be forever living in this iPhone age where we have this device that we stick into our pocket that exists between us humans and and the network that's going to change and and ai is going to change it so in 10 years there's not we're not going to have the same relationship with our iphone
So augmenting human capability through devices involves inputs. Elon Musk is trying to do brain implants. Those inputs are going to evolve. And the iPhone surely is not the end of history. I agree with that. Although it is hard. to understand what the next leap will be because the assumed wisdom says that will be some kind of glasses. And to be honest, there's nothing good enough, close enough yet for that to be true.
Yeah, but we know that in Silicon Valley, that kind of statement is very dangerous because next week someone will come out with something that changes everything in the same way as you might have made that remark about smartphones before. Jobs presented the iPhone in, when was it, 2007. So it's always, it's never obvious until it's obvious, in other words.
But I think voice and listening is easily the biggest unfulfilled opportunity. Eyes, not so much, but voice and listening is huge. And Apple's very well placed, better than Google. Google, by the way, probably as a repost to Apple, did announce that they're dropping the Google Assistant on Android, which is their Siri equivalent, and replacing it with Gemini, their AI. So Google's saying, well, we are going to ship.
Yeah, and our post of the week talks about how AirPods can... and we'll get to it later in the show, can work on the fly intelligently to change everything. So just develop that a little bit, put some more meat on the bones. In five or 10 years, are you suggesting that this device, which maybe we won't even have to stick into our pocket, might be smaller,
Only guys like you do that. Actually, it's really good, but I think there's a social problem, which is you can only talk when you're in private, because otherwise everyone around you will be irritated.
Well, everyone talks to their phones anyway. I mean, there's nothing more irritating than being on a plane or in an airport and having to listen to other people's conversations, but they don't seem to be too bothered with that. So there's a difference between talking to a colleague and talking to an AI.
But I think the key point is the technology is already there for that. It really is. Somebody has to exploit it. It's what I sometimes call the unborn child. It's really fully formed, just not born yet. What is interesting is that what that gives every individual on the planet that can afford a smartphone and a pair of earphones, it gives you a way to plug in to the world's intelligence. And not just like a library, but a library with a personality that understands the content of all the books.
And that, you've always been very bearish. You're always bullish on Apple and always very bearish on Google. But if you're right on that, And that should be very worrying for Google because essentially the search engine gets replaced by a pair of earbuds.
Yeah, and they have every chance of, being a player in the next phase, they will be for sure. Like Microsoft is post the internet boom. So I don't think any of these companies are under mortal threat. It's simply how is the deck of cars going to land?
So you talk about being players in the next phase. A lot of people are talking about how Apple can be a player. One link you sent from somebody called Chris Prucha suggests that Apple should buy Anthropic and build Siri from the ground up. You argue in the editorial that apple should buy perplexity but before we even get to the anthropic versus perplexity question would they be wise to do a show an ai shopping spree and go out and acquire someone i mean in in in trump 2.0 america it's probably doable you know
it isn't their dna apple hasn't made a big acquisition for many years so the the likelihood is they will not do it It's also not clear they need to do it because of the reasons we've already discussed. They're in a very good position anyway. But if they bought perplexity, and just to translate that, that's a little bit like buying the next Google. with built-in AI. Perplexity has its own AI, but it also plugs into all of the others and gives you- Yeah,
and you're very bullish on perplexity. I have to admit, I'm less impressed. You told me to sign up, I did, and I don't really understand why perplexity helps me any more than anthropic. It's just more confusing. There's more layers, but you tend to be right on these things in the long run.
Yeah, but I always think, you know, I want AI, so I go to Anthropic. I want web search, I go to Google. If perplexity provides both, but neither are quite as good as the other Google search or Anthropic AI, or maybe not quite as good, but just different.
Well, if you're right, Keith, then... as I said, in the long run, you tend to be more right than wrong on these sorts of things, then it's a no-brainer for Apple to acquire perplexity. I mean, they're only, what, valued it about? I don't know what their valuation is. I mean, their value is, certainly compared to OpenAI, it's peanuts. Yeah.
Of perplexity. And its deep research engine is doing research as we speak. It'll take a while because I use the research engine, which is very good, by the way.
I mean, that's almost embarrassing. Fifteen billion. So if Apple offered, I don't know, a hundred billion or something, which is peanuts to them. So this Anthropic versus... Anthropic versus... perplexity issue. Would it be more complicated to buy Anthropic? You also have another piece from The Times in this week's newsletter about Google owning 14% of Anthropic. And I know Apple have a chunk of piece as well. So would that make it more or less interesting? I know you also have an interesting piece in this week's newsletter from the information about AI circular money flows, which means that I don't know, Apple buying Anthropic would just be another circular money play.
Yeah, I think Anthropic's biggest owners are Amazon and Google. And so you get the incestuous nature of AI there because Anthropic and OpenAI are both startups that attracted capital from...
I didn't notice, but you're right. Anyway, the article about round-tripping of money is that everyone's buying from NVIDIA, and then NVIDIA's buying stuff back and investing in companies, which it is.
Sounds like Web 1.0, that boom. You remember when everyone was buying from everyone else? I remember it well. Ridiculous stock market valuations, and then the whole thing crashed.
But the truth is that for Apple to buy Anthropic... they'd have to want to own an LLM. And I don't think Apple needs or wants to own an LLM. I think it wants to integrate AI into consumer-facing stuff.
Yeah, well, it does have its own as well. If you don't pay perplexity, you use their LLM, which is decent. It's not terrible. But I don't think owning an LLM, I think that's off the table now. And really big companies with enough money could probably build an LLM given enough time and skills. as well. So I don't think buying an LLM makes a bunch of sense in any long-term view. But I think owning the consumer experience and making people delighted with it, that's where I would want to be playing.
So we've got to move on. But finally, on Apple, you're not too worried. In other words, it's not a crisis for Apple. And you're not particularly convinced that they should buy or could buy perplexity. In other words, You've come up with this big story, Keith, which is not much of a story.
But really, when it comes to crises, this isn't one. Steve Cook, not Steve Cook, Tim Cook can sleep well at night and not worry too much. This is not going to cost him his job. Correct. And Steve Jobs might be spinning in his grave for lots of reasons, but maybe not this one.
The embarrassing thing is that they did an ad, a TV ad for the iPhone 16, promoting Apple intelligence as the reason to buy it. And You know, they did deliver some of those features, but not most of them.
Yeah, I'm already salivating, I have to admit, in the way you're presenting it, of having my iPhone as an intelligent companion. There you go. So I don't have to fiddle around with the annoying inputs and put my fingers and mistype and all the rest of it. All I want to do is talk to it and say to the phone, should Apple buy... perplexity or anthropic, what would that mean? What are the pluses and the minuses and what are the likelihood? And it could respond automatically much faster and even more articulate than you, Keith.
Oh, easily. I'll give you a challenge, Andrew. If you get a moment, put your ear pods on, go for a walk in your nearest park, I have a park down the road, the Golden Gate Park. There you go. Use the OpenAI conversational mode. You'll see there's an icon on the OpenAI app on your iPhone. And just start a conversation. I think it'll blow your mind.
Well, you've always been very, very bullish on OpenAI. The OpenAI news this week is that a certain Sam Altman has... called Deep Seek State Controlled. He's been flirting with a certain Donald Trump. Does that make OpenAI now part of the new Cold War against China?
Well, it's because they're so... Sam, for a smart guy, is turning into kind of Dickensian, McCorber-like character where he's trying to manipulate policy in his favor. And the latest, which is a real gut punch, is to accuse DeepSeek of being the Chinese government's tool, when it clearly isn't. It's a hedge fund. And the furthest away from the Chinese government you could possibly imagine. Now, of course, the Chinese government does control some of the content inside the borders of China. But it's far from being an agent of the Chinese government. So it's just a low blow in an attempt to compete.
I wonder, though, whether actually that might suggest that another influences on Sam Altman. Chris Lehane, who's a very well-known operative now, is working very closely with...
with Altman. All sorts of news about that this week. You didn't put it in the newsletter. He had an Axios interview. Lehane is now the strategic arm of OpenAI. Is Altman perhaps losing control, perhaps, rather than being a Dickensian character?
I don't think there's any chance he's losing control, but I do think he's collaborating on the messaging. And the messaging is not driven by Altman alone. is driven... Lehane is a world-class messenger.
Yeah, and Lehane is... Lehane's wife is very close to my wife, so I get all the Lehane news. I would not want that guy working against me. He's a scary character, right out of Machiavelli, but very smart as well. Meanwhile... AI is changing the world. So we're told we've had these articles before. Kevin Roos, who's a particularly good writer on tech, New York Times tech guy, has a piece out called Powerful AI is Coming. We're not ready. How many times do we have to see this article, Keith?
You know, when I read it, you sent it to me initially. And when I read it, I thought, this sounds like it should have been written a year ago. Ten years ago. So it's really, you know, he, Kevin Roos, has had an aha moment that this stuff is for real and it is going to change everything. And he's written about that. And that's all the article says. So it's a bit of a non-event article. It's only worth noting because it's, It's like Bill Gates writing how great the internet is in 1998.
Yeah, except that Bill Gates was running Microsoft. Kevin Ruse is just a tech writer for the New York Times. Yeah, exactly. Although these Times pieces, I remember the Times piece, it was in, I think, 1996, that came out and said, the internet's going to change everything. For those of us living in San Francisco at the time, it wasn't news, but it was news to the rest of the world. And it's these kinds of pieces in mainstream publications, trusted publications like the New York Times, that actually convinces normal people.
I agree. It's in the Valley, they call it the S-curve. This is like halfway up the first S-curve. And yeah, it means that the middle adopters are going to be adopting and the late adopters is probably still two years away.
One final piece on AI and then we'll move on. Interesting piece by Alex Wilhelm. You like his cautious optimism blog. He talks about closed source AI models are doing pretty darn well. What exactly is a closed source AI model? Is that like OpenAI?
Meanwhile, the company behind the best-known and perhaps the most powerful open-source AI, Facebook, has been in the news. You don't have this in the newsletter. You said it's a non-story. A book called Careless People, a cautionary tale of power, greed, and lost idealism by Sarah Wynne Williams, who was pretty senior at Facebook for a while.
I don't know. One would have to know her to know whether their lies are just exaggerated. So why do you think they might be lies? You know, I live in Palo Alto and this book is discussed in circles and the impression I get is that there's one story where she accuses Sheryl Sandberg, who as far as I know is not gay, of hitting on her on a private plane and the The story I've heard is that Sandberg had the only bed on the plane and this lady was pregnant at the time and not very comfortable. And Sandberg offered to allow her to share the bed.
Yeah. And the lost idealism, that bit's true, I think. Which isn't news either, though. No. No. But I've been through it. It's that moment at which your company is past the tipping point where it might fail. And the team have all been hired to grow revenue. And money becomes the primary goal. And principles of money are in a constant battle with each other.
Yeah, although I find it hard to believe that anyone ever worked for Facebook, certainly after the first year or two, without thinking about money. Speaking of failure, Keith, you had a moving piece by your friend O'Malik 10 years later. It's 10 years since... Giga Ohm collapse. Maybe we should have a failure of the week. Giga Ohm seems to be the failure of the week, although Ohm suggests that after 10 years, it gives him some perspective. We've all gone through failures, so we all kind of know what he's gone through and what he feels like. What does Ohm tell us 10 years after Giga Ohm collapse?
It's a very personal story of the reconciliation to failure and the recognition of of what the experience gave you to allow you to be successful in what you do next. I say this a lot, but it's so true, it's worth repeating many times. Failure is the default assumption. And that's true in all fields of science and art. And in life. And in life. So if you didn't fail, you wouldn't learn. Now, of course, you don't want to fail, always fail, but it is inevitable that you will always somewhat fail. It's not a binary thing, like either you succeed or you fail, but within any day, there's going to be goals you can't achieve and lessons to be learned. And Ohm's very good at describing over a period of years, by the way, I think it took him three years or more to fully recover from closing gig at Ohm. GigaOM was significant, along with TechCrunch and VentureBeat. They put on the Crunchy Awards every year together.
Yeah, and of course, by calling it GigaOM, he personalized it. And I think this is a good piece, much better than some of the other... sort of celebrations of failure by obscenely successful people like Reid Hoffman, who don't really know what failure is. So it's well, for entrepreneurs, it's well worth reading because you're right, most of us fail. We've all been through it. I've never succeeded in anything I've done except for this show, Keith. So I think that's healthy. And he is very honest, a very down to earth and a guy I think we all respect. Going on the other way when it comes to success, your startup of the week, Keith, is Cursor, $10 billion valuation for the AI code editor. Actually, $10 billion doesn't actually sound that much.
Yeah, and I was telling you earlier that there was a piece that I missed. I meant to forward it to you about, because you're always talking about how easy it is to use AI to code. And there was one piece suggesting that even for coders, AI isn't that easy to code. Does cursor actually, I mean, what's the value of cursor? Do they make it easy? If I went to cursor, could I code?
Yes, you absolutely could. But you'd go through a very painful learning curve where you failed a lot. It took me about a week to get to the point where I believe... It takes you a week. It's going to take me like 10 years. No, I don't think so. I mean, it's putting up with it not working because you're not asking it... you know, properly. So it's learning how to ask and what to ask for. Interestingly, yesterday, OpenAI released a new interface for programmers, API, and it's called the Response API, and it replaces what was previously the chat API. And the Response API is just a better way to engage with their LLM. And I opened Cursor and asked it to upgrade my OpenAI implementation in the Reflect app that I've written for teens, from the old one to the new one. And 10 minutes later, I had the upgrade. It was working. I submitted it to Apple, and my beta testers could download the app with the new API, all done through Cursor. So it's like anything, you know, like skiing. Once you learn to do it, you know how to do it. Yeah, but skiing isn't easy either. Exactly. So nothing...
But I think it is inevitable that in the future... Most of the time, an individual startup will not need to write code. There's a few articles this week about how little money a startup needs these days coming out of Y Combinator. And partly it's because you don't need to hire engineers if you learn how to use these tools. And that's certainly true. And so the world, that's probably the fastest part of change is engineering.
finally um we would began with apple let's end with apple it occurs to me that maybe sam altman waltz and all is the steve jobs of ai um he certainly associated more with ai and changing the world like jobs was with the smartphone but the real question is who's going to be the johnny ive of the ai world um And we need a Johnny Ive. Your post of the week is from Josh Constine. Is he still with TechCrunch or will he do something else?
So these TechCrunch guys always go over to VCs. He has an interesting post about Apple planning an AirPod feature that can live translate conversations, which speaks of what you talked about earlier about the... uh the ai audio features of uh the new world um but does this need to be i mean is this the right question to ask does apple or somebody else need a johnny ive a real designer to make this viable and exciting for consumers
I think it's less a Johnny Ive. It's really a product vision. This use of AirPods for real-world purposes, I think translation is an obvious one. But there's many more.
Well, it's an obvious one, except most of us don't need it. It's not essential. I mean, it could be valuable. I don't want to make any jokes about... in California about going into your local store and using it. But I mean, most of us, if you're not traveling, you don't really need it. Especially if you're an English speaker. Maybe if you're a non-English speaker, it becomes different.
Right, so the impact it would have in the world is probably quite minor. By the way, I also think it's underestimated the extent to which English has now become a universal language, even in China.
Yeah, but still, what it points to is that the AirPods is a source of information in real-world context. So, for example, let's imagine you're traversing a new city that you've never been to before. I don't know if you've ever done it, but holding your phone in front of you with the map, seeing if you're going in the right direction and such, is nuts.
It's not only nuts, but it's likely to get your iPhone stolen. We were in Santiago, Chile a couple of months ago, my wife. almost had her phone stolen by a motorcycle thief.
So yeah, there you go. So the AirPods, you know, linked to an intelligence would be able to tell you keep going, it's the next street on the left, you know, whatever, whatever, or your trains running.
Well, it's a way of delivering information to the human that they can act on. Yes, the AirPods themselves are not the intelligence, they're just the conveyor. But that relationship between a powerful computer and your ears, imagine the products you could come up with that would be helpful to people.
one of the unheralded products of our digital age. They seem humble. They're under, what, $100, $200, but they do change everything. So it makes a lot of sense for them to become more intelligent. Yep. So, Keith, we have a promising future, do we? Got smart iPods. The AI revolution is for real. Apple's going to acquire perplexity, perhaps ananthropic.
Or not. We will find out next week. Won't happen next week, but it may happen this year. Lots of things changing, lots of potential, lots of successes and failures. Keith, we will talk again next week. Thanks so much. Bye, everyone.