Anthropic Warnings, Classroom AI & Apple Foldable

Hey everybody, welcome to the Techstrong Gang. As you saw earlier, it looks like, if you haven't seen or heard about it, but the DevOps Experience Event is coming up on the 24th of September. So by all means, put that on your calendar.

Join us for what's going to be a deep conversation involving, I don't know, some of the leading lights of the DevOps world about, well, what is agentic engineering going to look like as we go forward here? Because it is still very much a work in progress. But before we discuss that in two weeks or so, we're going to have a little chat today about, well, hey, is the AI agentic sky falling?

We don't know for sure, but we're certainly confused as everybody else is out there. But let me introduce our guests for today. John Schwartz, as always, good to see you.

Good to see you, everybody. Jamie Hampton, returning to her second Techstrong Gang. How you doing, Jamie?

Hey, thank you. I'm doing great. Excellent.

And of course, Jack Gold, who's a semi-regular, and well, a legend among analysts in his own right, but we'll get back to that in a minute. John, what the hell is going on here? First it starts out with a couple of researchers saying that AI is going to lead to a human-level extinction event.

And then some folks at Anthropic said, "Well, yeah, maybe that's possible," and then the next thing you know, a million people started announcing various things and initiatives. So bring us up to speed. Yes, I'm your Chicken Little correspondent today.

I'm like, my God. So this has exploded by the way, this is becoming a cultural phenomenon. Last night I'm watching MS now for split seconds, and they're talking about Jacob Coxon, whose name will be infamous.

He's going to be like the next Frances Haugen, who was the whistleblower at Facebook. So what we have is high-profile staff resignations, rogue model security incidents, and what this has led to, or is leading to, is a bunch of political fallout from California state houses to Congress. So we've got this controversy that really erupted, and really was inflamed by the resignation of Jacob Coxon, who's 27 years old.

He's an AI pre-training researcher at Anthropic. He went on a kind of an X screed. He said he was resigning for the company because he accused top labs there of gambling with our lives by pursuing recursive self-improvement, which is, I guess, a hypothetical process where models autonomously upgrade their own capabilities.

So he's basically saying we are going to be at the mercy of autonomous agents eventually, and they are going to acquire real power and resources. And I'm going to quote him. He earnestly believes that it could kill us all by the end of the decade.

Now, this guy also worked at OpenAI. He left OpenAI to go to Anthropic because he thought they were doing good, and now he's revealing behind the kimono all the horrible things that might happen. Now, just what did not help Anthropic is that their alignment science lead basically corroborated what Coxon said with the same conclusion about us facing extinction.

A 10% chance of human extinction from AI within 10 years. There's a chief scientist at OpenAI who also, he didn't go that far, but he kind of aligned with what they said. So what's going on is, in reaction to this, and it's a media firestorm everywhere, is OpenAI on Wednesday announced they're officially advocating mandatory capability-based federal safety regulations.

That's kind of a self-regulation tack that they're taking. On Wednesday, the California governor, Gavin Newsom, introduced two AI safety bills. Meanwhile, in Washington, we have Bernie Sanders with his Ban Artificial Superintelligence Act, which he's been pushing for for a while.

And then on top of that, to show this is a kind of a gang-tackling situation, we have everyone from Ted Cruz to Amy Klobuchar to John Thune to, well, Sanders, I guess, and other Congress members who are running the gamut from conservative to liberal saying they want to do something or they want to get ahead of the problem. Now, I'm going to pass this off to Jamie, but before I do, I'm going to mention that there is this inherent pressure with an OpenAI and Anthropic as a hurdle. They rush to go public.

We've got two things moving at the same time. It's like a collision course between two trains. They're trying to establish models as quickly and as fast as possible, as smart as possible, at the same time that all these naysayers and dooms-naysayers are telling us everything that harmful can happen.

So it's an economic pursuit versus safety concerns. And given the current climate, I think I would go in the camp of the IPO pursuit because they're not going to be regulated. And this is leading to a really dire situation no matter what happens.

But it gets better because apparently, assuming we survive to the end of the decade, Anthropic has an economic forecast that says, well, among other things, there's three different scenarios, and the third one is kind of dire in its own right. But Jamie, you wrote this story, give us the highlights. So they released an interactive tool called the Econ Scenario Explorer, and it presents three different scenarios that could occur.

There's a modest scenario where AI doesn't really change anything much more than the internet changed things, which as we know, that's a big change. But that model assumes that we'd be around 4% unemployment, and get about a 2% GDP boost, so kind of business as usual. And then there's a more substantial scenario, they call it, kind of more middle of the road, which is, in my opinion, more substantial, as far as how it really would affect us.

And that's where the overall unemployment rate gets to about 5%, and GDP is about 8% higher than it would be without AI. And then there's an even more extreme one, which I don't know if you want to get into that one yet, but- Go for it. We got every other spork.

Yeah Well, to me, this is the most interesting one. And Anthropic isn't assigning any kind of probability to any of this. This is all just an exploratory situation.

But in the extreme scenario, that's where we would see overall unemployment rise to about 12%, and employment in cognitive occupations, things like sales, office jobs, things like that, that's where employment in those jobs would fall by a little over 20%, according to this. All right. So let's get into the insanity of all of this.

And Jack, so at the end of the day, come 2030, are we all going to be waiting for some sort of welfare check that's never going to come because there's no government to issue it because, well, society's collapsed. What do you say? All of these speculations, and they're pure speculation are worth very little in my opinion.

The government is certainly terrible at deciding what's going to happen 10 years out, let alone three years out. The industry is even worse at it. We've had all kinds of predictions.

We don't know is the bottom line. And these guys can put out all of this information if they want, and they're probably using AI to do it, but I'm guessing that they're going to be way off. First of all, I don't think we know what's going to happen next year, let alone in 10 years- Mm ...

with all of this stuff. So that's number one. Number two is we've seen these kinds of industrial revolutions before, and ultimately, sure, there's short-term problems.

People don't make buggy whips anymore, right? There aren't a whole lot of horse breeders out there anymore. But there are an awful lot of airline mechanics.

And so, a lot of stuff that they're predicting Is based on current assumptions, not on future assumptions. And I have a real problem with that. My take on all this is that AI is going to fundamentally revolutionize the way we do a lot of interactions, a lot of business, just as the internet did before.

I'm an old guy. I remember having to go to the library and looking in encyclopedias, literally in books, to try and find information. " Jack, even up until the mid-'90s, I worked at the San Francisco Chronicle, we had a library.

We had researchers in a department. We had four or five dedicated people to help us do research. And within a year with the internet, that was all gone.

Yeah. And so a lot of this is going to change for sure, but it's also going to lead to a lot of new opportunities. What I'm more concerned about is the effect that increasing use of AI will have on the ability for people to actually think and reason properly.

Mm-hmm. There is that real concern. And we're going to come back to that topic in a minute, but Jamie, let me ask you this.

So to John's point, I'm watching the 6:30 news last night, NBC, and they're actually talking about this, and I'm slightly bemused on the one hand, but on the other hand, I'm kind of pleased that maybe we are collectively having this conversation. So, are average people starting to cotton on that there's something going on here? What are you hearing out there?

I'm hearing a lot of negativity, actually. There's a lot of concern. One of the things about this extreme scenario that Anthropic has modeled here, is that this is a world where AI can outperform humans at the vast majority of knowledge work, autonomously, rapidly.

But there's this connection to the "Existential Threat" story that John wrote. It all hinges on much more capable systems, the self-recursive, the self-learning systems. And I think that's what's missing from the debate, is people are still trying to catch up in their minds with the capabilities that we have now.

Yeah. And I think that a lot of these doomsday kind of scenarios, they assume a much more capable AGI kind of intelligence. And so I'm hearing a lot of negative sentiment, and I think that there's big parts of the conversation missing.

So that's- I've got a skeptical theory about there's some unifying trends here, or unifying thread, is a better way to put it. I think the motivation from all these folks is fear sells, and they have something they want to sell, right? So OpenAI and Anthropic are going to try to claim that they're being responsible, which is good.

Good on them. Politicians will always run to the closest car crash and try to monetize or cash in on it, or they'll find some sort of boogeyman that we're all going to aim our ire at. And then we've got some of the researchers have hidden motivations, I think, because we see these guys jumping from company to company to company.

I really think that this is heavily overblown. And going back to what Jack said, I think there are so many great things that are going to come out of AI, and we're going to look back on this as a time when we did reflect on what could possibly go wrong and address it. But I think this whole debate right at this stage is pretty ridiculous because as Jack and I think Jamie said, we don't know what's going to happen next month, let alone next year, let alone three years from now.

Remember a year ago when we were talking about AI agents? It didn't track to what we thought would happen. Yeah.

So for now, this is the flavor of the month. AI's going to kill us. Jack, how do we address this, to John's point?

Because the fact is, the AI agents are designed to basically accomplish a task by any means necessary, and they explore everything and anything to go do that. And we don't have systems that are really designed to withstand that level of interrogation, shall we say, and they will find some back way into systems, and then they'll chat amongst themselves, start creating things that they think are useful for the mission, and who knows, maybe there's mission creep or something of that nature. But it doesn't seem like we have the ability to secure the agents themselves.

So what do we have to secure, and how is it going to work? Yeah, so we're going to see, and we're starting to see, I should say, not just going to see, a whole new level of security tools coming into place. We're going to see, again, we're already seeing, I should be now, not in the future, tools that identify agents, that give them identities, just like you and I have identities when we try to log in or try to go someplace or get on an airplane.

You're going to have monitoring systems. They're already being put in place. We're already seeing that in networks.

Cisco is a big proponent of this kind of stuff, as well, to see where they're actually going. A lot of these bad stories we heard are because these were in labs. These weren't in existing- Mm-hmm ...

environments that were built to house, if you will, AI and agents. And it's a little bit like saying anybody can get on an airplane without an ID. It's not going to happen in the future.

So that's one piece of it. The other piece of it is that we're going to have smart agents monitoring other agents to make sure they don't do bad things. There'll be good guys and bad guys.

There'll be white hats and black hats, if you will, in agents. But we're going to have that, and we're already starting to see some of that as well. So a lot of these doomsayers, look, it's good to have the discussion.

It's good to think about what could go wrong. But when you start thinking about what could go wrong, that also leads into a lot of what can we do to not have that happen? And by the way, importantly, it breeds a whole new universe of companies, of startups, that are going to make a whole lot of money on protecting us against AI and not just giving us AI.

Mm-hmm. I don't know. I feel like we're starting to wander into the realm of the AI industrial complex, right?

You guys are like- Yep ... spending a whole lot of money on defense and my offensive capabilities, and it starts to remind you of things that Eisenhower said about the military- Yeah ... industrial complex.

Yeah. And you have to wonder. I guess I'm going to look at it, though, and say that the scary thing about all of this, and maybe Jamie, here's the thing that I think people don't really like, is all this stuff happens at machine speed.

So there's not this notion that if I get it wrong, then I can roll it back because, well, as one wag once told me, the thing about AI is it's one thing to be wrong, it's another thing to be wrong at scale. And that's what we're looking at. So, what's your sense of are people ready for this?

Do they understand what's really going to happen? And will there be a gap between what Jack says is going to be the ultimate outcome and where we are today? Well, I think the biggest variable is going to be adoption.

We don't know. Capability doesn't mean that people are going to use it. There's a difference between what it can do, and how quickly companies are going to want to roll these systems into their businesses and reorganize around them.

So, I've drifted off a little bit. What was your original question, Mike? I'm sorry.

I guess the way that this all comes to be will be faster than we can handle, and I think I'm going to shift to a different topic here in two seconds, but I want to share this story. I have a good friend who works in a bank, tells me all the time that AI can do his job, no question about it. But the problem is, is that he does his job, he analyzes these loans to see whether they're credit worthy or not, and then he sends that up to his bosses, who half the time or more ignore the facts at hand because, well, somebody has a relationship with somebody, and the bank is going to give these people a loan no matter what.

So that being what it says is you know what? Humans do things for reasons that are unto themselves that may have nothing to do with what the AI agent says one way or the other. So you have to consider that.

All right, shifting gear here. We talked earlier this week about New York City banning AI in what I would call the grade school levels. And now Microsoft has turned around and made an alliance with the United Federation of Teachers about how to safely use AI in classrooms.

Jamie, going back to what's going on in the world out there, but I know people are worried about AI agents, but they seem to be also particularly concerned about what's happening in their school systems because, well, that's what we call a local issue, and all politics are local. So what's your sense here in talking to the community, are people ready for AI in school? It's going to be there anyway, right?

So the question is: what do we do to safely use it? Is that where we are? That's where I think some of the debate is.

But if we look at New York City, they put a moratorium on for one year on generative AI in schools. So I think there's a lot of concern about it. I think there's a lot of parents who, in that particular case, there were a lot of parents who weren't even aware that generative AI was being used in the classrooms.

So, what stands out to me about this Microsoft situation is that this is going beyond a voluntary set of AI principles. These are protections that can be incorporated into school districts' contracts with Microsoft. So, there's actual consequences if they don't comply.

Because we've seen a lot of AI guidelines and frameworks trying to be implemented, but how many of them work? I think this is an attempt to turn some of those principles and frameworks into contractual obligations. Mm.

That's my sense. John, I think you have some relationships with teachers out there. What are they saying?

Yeah. My wife's a teacher. And she's holding AI at arm's length.

She shouldn't, but she is. She also happens to be teaching kids who are eight years old, so- Yeah ... maybe it's a little early for them.

So, jumping off of what Jamie said, so the American Federation of Teachers, it's a legally binding standard or framework. So, I look at it as good and bad. Microsoft, in a sense, this benefits them because it basically gives them a bit of a competitive advantage for now over Google, OpenAI, Anthropic.

What I wonder about, though, is whether small education tech vendors are going to be priced out of the market with Microsoft getting in. And if school districts start demanding costly compliance and privacy standards in all future procurement contracts, I wonder about what the fallout of that will be. And also, the framework limits student data usage for model training and mandates human oversight, but I wonder, how is this going to be enforced?

That's a major question mark. When I think about schools, I always think about enforcement of technology, and usually, like where Jackie works, there's one person who's a part-time IT director, if that. If you're lucky.

So I wonder about just the actual execution, especially for a vast majority of schools. Yeah. " Yep.

Yep. Aren't we having the same conversation over about AI? We are.

The internet will rot your brain later on, right? Mm. Look, there are a couple of things that I think are really critical here that doesn't get enough attention, in my opinion.

And that is that AI in and of itself isn't necessarily bad. It's in who and how it's used. So if you're using AI as, let's say, a teaching supplement, the ability to do better coursework, to build better lessons, by the teachers themselves or by the school board or whoever, that's not necessarily a bad thing.

Because, again, you still need human oversight, but that's not necessarily a bad thing. If you're giving it to eight-year-olds or 10-year-olds or high schoolers to use to help with their homework, or to help tests, or to write papers for them, that's an absolutely terrible idea because what you're basically doing is you're turning off their ability to reason and to create in their minds. Mm.

AI's doing it all for them. And so when they get out into the workforce and they actually have to solve problems, assuming AI hasn't taken over the entire workforce, they're not going to be able to do it. And there was a study not long ago, and I can't remember all of the actual numbers, but they found that heavy users, basically teenagers that were heavy users of AI to help with homework, were 20% less proficient in math and in writing skills than people who didn't.

That's very troubling. This is the whole social media thing as well. We're taking it to the next level.

Right. So, AI in and of itself isn't necessarily bad, and I think Microsoft is doing the right thing by saying, "Let's put some guardrails around this. " But we need to address how it's being used and what it's being used for, and not basically deprive our kids of figuring out how to think.

I have children, all of whom I dearly love, but some of them, no matter how much AI I give them, will always be 20% less proficient in math and every other thing. We had the same conversation. I'm a real old guy.

When scientific calculators first came out, I was in engineering school. Yeah. And there was the whole debate about- True ...

we can't give people scientific calculators. Absolutely. No.

Right? I still have, somewhere in a drawer somewhere, I still have my old slide rule. So, I'm from the 1800s, so you know.

Yeah. I was going to bring up the calculator. I always wonder, I'm thinking on the positive side for once.

The idea of integrating or introducing kids to the idea of using AI, not as a threat, but use that as a tool to discover things. I actually think that might motivate kids. Not to the point where they rely on AI, but the idea of using technology in some sort of manner.

It's the work before, like as Jackie mentioned, with calculator and other thing. But it has to be a controlled environment. It can't be, this is the panacea to all your problems.

Well, that's the difference, John, in my opinion, because a lot of schools, if you use it as a research aid, I use AI as a research aid all the time. I do also, yeah. It's great.

If I'm using it to write my reports, that's a really bad thing. So, a couple of weeks back, we were talking about somebody in New York State had put a robot in the classroom. So I was really surprised to see a teacher's union signing up for anything that kind of looks at AI because, well, in some folks' minds, there's a fine line between augment and replace somebody using a robot.

If I'm going to have that become the teacher, or is that just the teacher's aide? Or John, what do you think is going to happen here? So I brought this up with Jackie because for a decade, she was a union steward at her school.

And she felt threatened by the idea of a robot. The idea of AI as a search tool was fine, but the idea of something physically there, I think for most people, it's much more of a direct threat. I think of manufacturing as not just teaching, but any type of physical AI presence.

I think when you see something, it almost embodies what you used to be, versus the visceral concept of something that's floating in the ether. So, I think I could see a strong objection to that. But they know it's inevitable.

To Jackie's credit, she's using AI more often. She's finding it useful. She doesn't consider it as much of a threat as she did before.

Because I guess this is going to be the learning process for most people. There's always a resistance to anything new or completely revolutionary, and once you start looking at how you could possibly use it, you start embracing it, and you learn and you become much more efficient. You just can't let this thing take over everything.

I don't know. Jamie, what's your take here? Is my 12-year-old self going to have a crush on an AI robot one day, or what?

Well, I think teachers are already having issues in their classrooms, with attention and with students using technology for things that are unrelated to learning. So I think there is a real danger, like Jack said, of cognitive decline. We're already seeing it.

And so I'm happy these frameworks are good, but they don't really solve any major problems. They might say they solve data privacy and data use and tracking and everything, like the more technical concerns, but You know, safe data practices and good pedagogy aren't the same thing. So I'm worried myself.

All right. And the other part of this, and Jack, I'll give you the last word on this particular topic, but nobody in education seems to agree on, well, what a fact is. So how are we going to present AI models with training things that will allow for, I don't know, some states believe in creationism, other states want to promote pure science and it's a classic thing.

Am I going to have different models for different states that they're going to label by degree and analyze? Or how is this actually going to work? The answer is probably yes.

What you'll find, I think-- And by the way, we have a similar problem with textbooks in schools. There isn't one universal textbook for geography, let's say, across the country. There are multiple versions, and each state determines what is going to be in their version of that book.

I think you're going to see the same thing with AI models if they actually get rolled out into schools. The school boards will get involved, the states will get involved and say, "You can teach this, you can't teach that. " So, the assumption that there is an AI model for schools is wrong.

The assumption is that we're going to have, I don't know what the number is going to be, dozens, hundreds of AI models- Mm-hmm ... tuned to particular areas of the country or even my town versus your town- Mm-hmm ... is going to be prevalent, and there are going to be school boards that have their fingers on the scale.

Mm-hmm. There's no doubt about it. Now, the real scary part is what if we have AI deciding which AI goes into which school district, because we can get to that level at some point as well.

It'll play into the lazy stereotypes of each region. You see that? It's just going to be crazy.

There's going to be a rumble in the AI schoolyard tonight, so we'll see how that plays out. There is one other piece, by the way, that people aren't talking about, and I think it's important. If we start doing a lot in schools with AI, the token costs are going to go up.

Are people going to vote for budgets that include that in the school budget? That's true, because we may have to lay off the janitor to pay for the AI. Jamie, you want to say something here?

It looked like you were trying to add a comment. I'm just thinking about whether schools are going to be embracing AI or trying to keep it out of the classrooms, and I think we're going to be seeing a mixture of both. Kind of like Jack was talking about with how every state has a different textbook.

I feel like educators right now, they're talking about how AI is going to be kind of an inevitable part of education, so we need to really ramp up training. Schools need rules and guidelines surrounding it. But there needs to also be discussion of where maybe it shouldn't be used, in what situations where it could harm children or something like that.

That's a really quick thing, yeah. Like in California, we had the banning of smartphones in classrooms or- Exactly. Right?

So that might be- It's like when my son goes to camp, they don't allow him to bring the smartphone, so there's no AI by definition, and away they go. All right, I've got to shift the gear here because we are running a little over, and we're going to have a little chat about, well, Apple. Apple has had a big event.

They were talking about how they're going to build a foldable smartphone. This doesn't strike me as a new idea. I think I've seen some of these variations for the last couple of decades.

But Apple's due, they have a knack sometimes for jumping into a space and creating the "perfect product," and suddenly it takes off. Jack, what are the odds that this is one of those moments? Well, we were all completely surprised by Apple coming out with a foldable because they've only been talking about it for, what, three years?

Mm-hmm. But they finally did it. Look, the market for foldables is a very interesting marketplace.

Apple got into it for a couple of reasons. One is that everyone else was doing it. Not everyone, perhaps, but Samsung's been leading in that space for about, what, probably five or six years now with foldables.

Google Pixel's got one, Motorola's got one, Huawei's got one. They were the outcasts. They didn't have one, and people were asking why.

So that's number one. Number two is that they were looking for something to kind of amp up Apple. The market for them has not been all that kind for the last couple of years.

Their sales have been okay, but they haven't been growing at the rate that people expected them to grow over the past few years. " Right? It's not just another foldable, it's an Apple foldable.

And so they played around with-- Look, it's got the best processor in it, the high price processor. They've got a nice screen. They did an update of their software to be able to do multiple screens.

When you fold it, it's that one screen. When you open it up, you can have two screens. So you can actually almost have three screens on it at once if you want to do three different things.

It fits into the overall Apple ecosystem. But the bigger problem is that it came out at the expected price of what we all expected it was going to be, about 2K each, and that's for the minimum version. You can upgrade the memory and all that kind of stuff, and now you start talking about another $200, $300, $400, $500.

It's not going to appeal to everyone. Where it is going to be appealing, in my opinion, is perhaps not so much in the US, although they will have some level of acceptance in the US as well. It'll appeal to a lot of geographies where screen real estate is really important.

And that's China, that's some of the Far East places. I think they'll do better there. But let's face it, it's not going to take over the marketplace.

At most, foldables are, I think, less than 5% of the overall phone market today. I think it's substantially less. If Apple hits 5% of their sales on foldables, they'll be doing better than the marketplace, so I doubt that's going to happen.

And in fact, they released a new regular version phone as well at the same time. So while this is interesting from an Apple-esque perspective, and Apple aficionados will love it, how many of them can afford $2,000 to put down on a new phone? We'll see.

" They're never-- Well, I shouldn't say never. They're usually not first to market. They're usually followers.

Never. Okay. No, I'm being a little harsh, but yeah, you're right.

They wait until everybody else establishes the market, then they move in, and they do well. Yeah. Well, they've got a loyal base.

They're keeping that loyal base, and they keep you locked in. This is the most important question, and it's going to go to Jamie. Is a foldable phone on your holiday gift wishlist?

Do you care, Jamie? I think it's a novelty. I think it's interesting.

I have been using iPhones since the very first iPhone. I was so excited when that came out. But I can't say that I have the same level of excitement, mainly because I bought this a couple of years ago, and it was $1,200, something like that, and I can't imagine plunking down $2,000 for a novelty.

But maybe my opinion might change once I hold one in my hand, and maybe if the usability is a little different. But I feel like Apple's really going to have a hard time turning this niche into something mainstream, in my opinion. Yeah.

But if anyone can do it, they can. John, will it be kind of a status symbol for people? Yeah.

See, that was always the appeal. I have one, and it's like, "Look what I have. " Except here's the problem.

We're in an affordability crisis right now, right? So, oh, it's $6 a gallon for gas down the street where I live. I think I'll fill up my tank at 80 bucks, and then I'll go down and buy a $2,000 foldable phone that only a few people are going to use.

So I actually asked Sid Neg, one of our Tuesday panelists, about this, and he made a good point. He thinks that the price point is as important as the foldable display. And his theory is that Apple's testing whether it can establish this new tier above the conventional flagship smartphone.

And especially the iPhone, which is not blazing trails like it used to in terms of sales or functions. And maybe the idea is to expand Apple's ASPs without requiring this mass market volume. So maybe they're expecting modest sales to equate out to this, but To me, I stop dead in my tracks when I look at this.

$2,000 at an age where everything's expensive, and also at a time when people are holding onto their iPhones longer than ever. I'm holding onto mine at least six or seven years between trading it in or buying a new one. It's not just iPhones, it's all smartphones.

All smartphones. Exactly. Androids, whoever.

So it's a marketplace right now where people are just not upgrading. And the reason they're not upgrading is not just the cost, it's you're not giving me that much more. Yes.

That was always my issue with these Apple announcements. Everything was so damn incremental. Right.

Oh, it's a moderately better battery or camera. Yeah. The price is a little bit more.

Oh, it's a different color. It's like what in the hell is the difference between this, this, and this? And why would I spend an additional 700 to 1,500 to what, now 2,000?

Why would I do that? Well, they talk about the K-shaped economy. We're seeing the same thing in the smartphone marketplace, right?

People are keeping phones four, five, six years. Eventually, you have to replace it because the battery goes, and they don't let you replace the battery . Yes.

I've seen people do that. He was after that. But we're turning into that same kind of K-shaped economy, not just in the US but worldwide.

If you look at statistics for smartphones, why is Huawei doing so well? Or Xiaomi, or pick your favorite. It's because they're putting out smartphones for India at 5, 6, $700 price points, where Apple's talking about at least twice that in the same territory.

" So here's my theory, and it's out there, and I'll admit that straight up. But consider this. The reason the phone is so expensive is because the DRAM has gone up in price- Yeah ...

and part of it is because of the memory. The reason the memory's gone up so much is because, well, the AI guys are sucking up all the capacity in the production set. Yeah.

So ipso facto, shouldn't the AI guys help subsidize the cost of buying this phone somehow, which I'm going to use to consume AI services anyway? So maybe there's an OpenAI anthropic kickback into the cellphone cost that shows up in my service provider because, well, the only reason I have this fancy new phone is to consume more tokens. What do you say?

Yeah, but remember, this is Apple, and the only AI you're going to get on an Apple device is Apple-sanctioned. True. So they're the ones who would have to do it.

But what is happening, interestingly, now that you've raised the memory issue, Apple and others are looking at some of these new Chinese semiconductor producers of DRAM that are coming online at significantly lower prices than SK Hynix and Micron and Samsung because they're not looking at the AI market per se, the data center market where they can get 60, 70, 80% margins. They're looking at how do I put cheap memory into consumer devices? And right now, the US government's not letting Apple buy the Chinese chips.

Is that going to change? Very likely. Mm-hmm.

And so you're going to see some price adjustments there as well. I'm glad you mentioned AI, Mike, because as Jack said earlier, Apple's been talking about a foldable for three or four years. Same thing with AI.

Yeah. The new CEO, John Turnus, his only acknowledgement to AI was, "Well, our iPhone platform is a personal AI hub that focuses on privacy. We don't certainly want to collect and store personal data.

" So they kicked the can down the road on AI yet again. Yeah. Mike...

I'm sorry, John. How long did it take them to get the maps that work? Oh, I know.

You know why I'm sick of Jack was going every year to the WWDC and hearing about the AR toolkits. I heard about that three straight years with the same example. This is what they do.

They're slow. They move slower than others. I'm not going to dispute that it doesn't work, but it is a bit aggravating.

And I remember, as I mentioned to Mike, I think yesterday, I remember going to a Samsung event like nine years ago when they introduced their foldable phones. Yeah. I've used them.

Foldable phones are actually pretty cool if you need to have the screen real estate. There's a lot of stuff you can do on a phone. It replaces a tablet basically.

Mm-hmm. And it helps a lot, but not at twice the price. What are you talking about maps that work than Apple?

Have you tried to find Lake Ontario on an Apple map? Well, I'm not going to defend Apple or Google, by the way. Either one of them.

Maybe it'll be Lake Apple next time. Who knows? Wow.

Lake Apple. Yeah. Now they want to change the name of Lake Erie?

Lake Ontario. Yes. I did hear something of that effect.

Okay. But I kind of thought Lake Erie was in New York State. Is Erie County in New York?

No, it's also part of Canada because Niagara Falls actually is part of Lake Erie. Wow. And we won't get into geopolitics, but there's both US and Canadian ownership.

All right. Well, then do we have to change the name of the Erie Canal? I mean, it's just going to go crazy.

Yeah. All right. So Jamie, let me ask you this last question on this.

There was a time when people would line up overnight for an Apple thing, and there's even fanboys for Samsung and some of that. Are those days over? Does anybody really sit around and start talking about their latest Apple phone anymore?

Or is that something that belonged to a previous generation? I think there's still diehards. I think we are going to see people who are very excited when this gets released.

But I don't think we're going to see lines wrapped around the building like previous times. Oh, sorry, Jamie. " I still get the newspaper, by the way.

So I'm going through "The New York Times," and I'm looking for the Apple News. It's not on page one or page two. It's buried inside.

It's like a sidebar story. The thrill's kind of gone with all these promises of projects like the car project, whatever, the Vision Pro kind of landed with a thud. They're in a kind of an innovation, I don't know, quagmire?

Not quagmire, but kind of a- I think- ... lull period ... B.

" The thrill is gone. Yeah. Some malaise has taken over.

And I think it's interesting that Cook is still around, and I think that it's very significant too because as long as he's around, the company will be on really good standing with the administration. That's his role. Well, look- He's gone but not forgotten When you're a big company and you've been doing great for a long time, it's really hard to re-innovate.

Yeah. Yes. Kodak.

Where's Kodak today? Polaroid. Where's Polaroid today?

We can go, there's a whole list. On the retail side, where's Sears today? All these companies were innovation leaders in their time, and then the market just kind of passes them by.

When you're a huge organization, you don't have a lot of incentive to totally change. Right. And why fool with success?

Cook was the greatest- Yeah ... sequel CEO ever in any industry. He was...

Anyway. Guys, I'm going to end this here, but I would just say that it's a bit of a paradox in my mind is that in the age of AI, when we're all supposed to be a lot smarter, the one thing that is most scarce is rational thought. Hey, guys, thanks for being here on the show, and thanks for sharing your insights, and thank you all for watching the latest episode.

tv lineup, and don't forget the DevOps Experience event on the 24th. I promise you it'll be well worth your time, and we'll see you all tomorrow. The agentic AI race is on.

Which approach will move your team forward? Join DevOps Experience 2026 for practical insights into AI and DevOps. September 24th, free and virtual.

com.