More or Less Podcast

Meta’s $17B Reckoning & Why Parents Can’t Compete With Big Tech's Algorithm

2272 summary words 10 min summary Watch video

Start with the signal

10 min read

Summary

At-a-Glance

  • Verdict: Watch fully
  • Core thesis: The panel argues that Meta's reported $17B child-safety settlement is financially and operationally insufficient unless youth-safety rules apply uniformly across platforms, while proposing agent-enabled, OS-level visibility as a more practical parental-control architecture.
  • Why it matters: The strongest material is not the Meta news itself but the product-policy design discussion: platform-level defaults are easy to evade, whereas standardized data access and agentic monitoring could create a new control plane for families.
  • Best use: Watch for the debate over system-level versus company-specific regulation, the proposed parental-data API concept, and the secondary discussion of Apple's strategic positioning for on-device AI.

Executive Summary

The episode opens with practical examples of agentic workflows, including an editor who says she emailed an agent to build and deploy a headline-scoring application trained on The Information's historical corpus. The app generated headline alternatives, scored proposed headlines, and was revised live in an editorial meeting. The panel treats this as evidence that lightweight internal tools can now be produced quickly, but notes that organizational AI tooling is often fragmented and governance remains necessary because editors retain final authority.

The central discussion concerns a reported Meta settlement with 47 states and other entities over alleged harms to children. The hosts characterize the financial component—described as up to $17B over 10 years—as immaterial for Meta, and disagree about what the settlement's behavioral commitments mean. The measures described include a default two-hour daily limit across Facebook and Instagram for teens, school-hour notification muting, use prompts, hidden likes, optional non-algorithmic feeds and disabled autoplay, and removal of extreme makeup filters. The panel's broad consensus is that isolated restrictions on Meta can displace teen usage toward YouTube, TikTok, Snap, Roblox, or other services rather than reduce total exposure.

Their more constructive proposal is a universal parental-observability layer: require operating systems and major youth-facing applications to provide parents a simple feed or API of a child's digital activity, which an agent could summarize against the family's stated values and flag for review. They argue that current parental controls are too fragmented and technically burdensome, and that a transparent cross-app activity log would be more actionable than a collection of arbitrary screen-time limits. This is explicitly a product and policy idea, not evidence that such an API regime is underway.

The closing technology discussion frames Tim Cook's legacy around Apple's hardware and silicon strategy rather than generative-AI product leadership. The panel argues that Apple Silicon's integrated CPU, GPU, memory, bandwidth, and power profile may give Apple a strong position for local agentic workloads. Their bullish case is that Apple avoided expensive data-center and frontier-model bets while retaining the consumer-device distribution layer; their caveat is that the verdict depends on whether Apple can execute in AI over the next decade.

Key Takeaways

  • Claim: Agentic development has reached the point where a non-engineering business owner can commission a functioning internal workflow tool in a single interaction, but human editorial or operational ownership should remain explicit. | Evidence: One host says she emailed an agent trained on The Information's historical corpus to build a password-protected app that suggests three headlines per story, scores a proposed headline from one to five, and flags scores of three or below; within roughly 20 minutes, it was live and later changed to provide six options during an editorial meeting. | Implication: Ken should view internal AI apps as rapidly deployable operational surfaces, but design them as advisory systems with accountable human owners, feedback loops, and clear override rights. | Caveat: This is an anecdotal success case with a context-rich agent and a narrowly scoped product; the speaker explicitly says editors can override the tool at any point.
  • Claim: The reported Meta settlement is unlikely to materially deter Meta through financial penalties, and its product restrictions are inadequate if competitors are not subject to equivalent rules. | Evidence: The speakers describe the agreement as up to $17B over 10 years and compare that with Meta's reported legal spending of roughly $2B per quarter; they argue that teens capped on Instagram can shift usage to YouTube, TikTok, Snap, or Roblox. | Implication: For platform regulation and safety strategy, evaluate system-wide substitution effects rather than measuring a policy by whether a single large company accepts restrictions or pays a headline-sized fine. | Caveat: The transcript does not independently establish the final settlement amount, legal terms, enforcement mechanism, or the causal effect of each safeguard.
  • Claim: The settlement's described teen protections are mostly a collection of controls whose effectiveness depends heavily on defaults, enforceability, and cross-platform coverage. | Evidence: Measures read aloud include a default two-hour cumulative Facebook-and-Instagram daily limit that requires parental permission to disable, hidden likes for teen profiles and interactions, school mode from 8 a.m. to 3 p.m., prompts after 15 minutes of continuous use, cumulative notifications at 60 and 90 minutes, optional non-algorithmic feeds, optional autoplay disablement, and removal of extreme makeup filters. | Implication: When assessing safety commitments, distinguish between default-on controls, optional settings, and controls that alter the underlying recommendation or engagement system. | Caveat: The panel notes that non-algorithmic feeds and autoplay controls are not defaults, making their practical adoption uncertain.
  • Claim: Parents are structurally outmatched by platform-scale recommendation systems, so merely assigning responsibility to families ignores the asymmetry in data, compute, behavioral optimization, and available time. | Evidence: One host describes preparing a highly deliberate set of alternative shows simply to avoid autoplay-driven YouTube consumption and argues that many families lack the time, resources, or technical capacity to do this; another estimates the platforms operate on millions or tens of millions of computers globally, versus a parent's single human brain. | Implication: Products aimed at family safety should minimize configuration burden and offer decision support, rather than assuming motivated parents can continuously inspect and curate every digital interaction. | Caveat: The speakers disagree on the degree to which parental responsibility versus public policy should carry the burden, and no empirical comparison of intervention effectiveness is presented.
  • Claim: A more useful regulatory and product primitive would be a standardized, parent-authorized cross-app activity feed that agents can analyze, rather than app-specific time caps and complicated parental-control menus. | Evidence: The panel proposes requiring operating systems and apps above a suggested threshold of five million users to offer parents access to a continuous activity stream of what a minor consumes and does. A parent could run an agent over that log to identify school-hour use, risky conversations, channels, or content patterns; one speaker currently screenshots a child's iMessage activity because no usable API exists. | Implication: There may be a meaningful product category around permissioned family data control planes: normalized event streams, natural-language policy configuration, agentic summarization, alerts, and granular intervention controls. | Caveat: This proposal creates major unresolved privacy, child-autonomy, security, consent, data-retention, abuse, and usability questions; the transcript does not address how access would be limited or audited.
  • Claim: Apple's strategic AI advantage may come less from a headline generative-AI product than from Apple Silicon and control of the consumer-device endpoint for on-device workloads. | Evidence: The panel credits Apple's system-on-chip strategy—integrated GPU, CPU, RAM, high bandwidth, and power-efficient hardware—as unusually well suited to local agentic workloads. The bullish framing is that Apple can avoid frontier-model and data-center capital intensity while remaining the last-mile interface to consumers. | Implication: For AI infrastructure and distribution analysis, separate the value of frontier-model ownership from the value of owning performant, trusted endpoint hardware and the consumer operating system. | Caveat: The speakers acknowledge that Apple Intelligence has not landed as hoped and that Apple could still be judged as having optimized the previous era if internal AI execution fails; they frame the outcome as unresolved for another decade.

Detailed Brief

Policy framing: uniform rules versus selective enforcement

  • Claims: The panel argues that youth-safety policy should be established at a societal or national level across comparable products, rather than through settlements targeting whichever platform is politically or legally easiest to pursue.; Meta's effort to condition a tighter one-hour limit on similar commitments by TikTok, Snap, and YouTube is interpreted by some panelists as a strategically sensible attempt to force a broader policy discussion rather than accept an isolated competitive disadvantage.; One speaker argues that social media for minors should be treated more like alcohol or cannabis: age-gated, labeled, and governed through common rules because developmental harms may justify limits even if enforcement is imperfect.
  • Evidence: The panel repeatedly uses an alcohol analogy: restricting one brand's alcohol content while allowing competitors to sell stronger products merely redirects consumers rather than resolving the underlying concern.; Australia's under-16 social-media restriction is cited, via Business Insider, as having failed to reduce teen usage and possibly coinciding with increased use.; YouTube is singled out by panelists as a more serious recommendation-system concern than Meta, while Roblox is described by one speaker as a product that should be banned.
  • Caveats: These are normative policy judgments, not a reviewed body of developmental-health evidence.; The Australia claim is secondhand within the conversation and lacks methodological detail.; Cross-platform equivalence is difficult: messaging, public social feeds, video platforms, games, and creator ecosystems create different risk profiles and legal issues.
  • Implications: A viable regulatory framework would need definitions that cover functional harms and recommendation mechanics rather than relying on company labels such as 'social network' or 'messaging app.'; Any safety proposal should be tested for behavioral substitution, circumvention, and whether it drives engagement to platforms with fewer controls.

Design direction: configurable software and natural-language parental controls

  • Claims: The panel sees an emerging future in which software itself is more editable: users can prompt an agent to modify their personal software stack rather than accept fixed applications and settings.; This could allow families to build a child-specific operating environment that prevents access to disallowed services before a child reaches an app such as YouTube or Roblox.; A startup called Clear Space is mentioned as pursuing a related approach: using a VPN-like layer to understand a child's activity and letting a parent state controls in ordinary language.
  • Evidence: The hosts describe current parental-control environments as a fragmented, irrecoverable mess even for technically capable and highly motivated parents.; They cite personal Spotify listening-history data as an example of the behavioral insights that become possible when a continuous activity log is accessible.
  • Caveats: One speaker warns that broadly customizable software will not automatically produce healthier behavior for adults; users may customize toward more stimulation rather than more restraint.; Natural-language policy layers still depend on reliable device-level and app-level data access, which major platform owners may not provide voluntarily.
  • Implications: The key UX challenge is not merely collecting data but turning cross-application events into simple, trustworthy family-level policies that nontechnical users can understand, change, and audit.; For agent platforms, this is an example of a high-value delegated-control pattern: user-defined values translated into monitoring, classification, recommendation, and reversible enforcement.

Notable Concepts & Terms

  • Agentic display / agentic workload: The hosts describe working with five to six concurrent agents and, in some cases, passing work between models such as Claude and Grok; it illustrates a shift from single-chat usage to parallel delegated workflows.
  • Email-native agent interface: One host advocates email as an effective agent interface because it supports asynchronous multithreading and avoids logging into separate product surfaces.
  • Default versus optional safeguard: A central distinction in the Meta discussion: controls that users must actively enable are treated as materially weaker than default-on protections.
  • Cross-platform substitution effect: Restricting one service can simply move a minor's attention to an unrestricted competitor, making app-by-app safety policy ineffective or counterproductive.
  • Parental activity API / sidecar: The proposed architecture is a parent-authorized stream of a minor's device and app activity that can be reviewed manually or processed by an AI agent.
  • OS-level control plane: Rather than relying on each application to expose usable controls, the panel favors operating-system-level access and enforcement spanning messages, video, music, apps, and other behavior.
  • Apple Silicon / system on a chip: The discussion treats Apple's integrated hardware architecture as the strategic foundation for running agentic AI workloads locally, efficiently, and with strong control of the endpoint.
  • Bitter lesson: The AI-community idea that scaling compute often outperforms hand-designed approaches; the panel uses it to argue that Apple's hardware capability may matter more than its current AI-product narrative.

Operator Notes / Why Ken Should Care

  • Prototype the proposed family-control-plane pattern in a bounded environment: normalized event logs, explicit authorization, a natural-language policy layer, agent summaries, and reversible actions before considering automated enforcement.
  • Treat child-data access as a high-risk security domain: define parent identity verification, consent and age transitions, least-privilege scopes, encrypted retention, audit logs, child safety escalation, and abuse prevention before exposing any monitoring API.
  • For internal agent-built tools, require a lightweight production checklist: source-data permissions, ownership, human override, evaluation examples, access controls, change logging, and a rollback path.
  • Track whether Apple exposes stronger on-device agent APIs, local model capabilities, and OS-level permission primitives; these would materially affect the feasibility of privacy-preserving personal and family control planes.
  • Do not rely on screen-time limits as a primary safety KPI; measure cross-service displacement, default adoption, bypass rates, and whether parents can actually configure and understand the controls.

Source/Metadata

  • Title: Meta’s $17B Reckoning & Why Parents Can’t Compete With Big Tech's Algorithm
  • Transcript words: 12190
  • Duration seconds: 3280
  • Timestamp note: No timestamps or chapter markers were provided. The supplied transcript contains repeated passages near the Meta, Apple, and closing segments.
Search this video Resume transcript at 0%
Full transcript 10264 words · 54 min read
0:00

At this point, I think that we're probably in a zone where social media should just be illegal, similar to alcohol. It's that bad for mental health. YouTube is the most nefarious of all. And Roblox, I actually think Roblox should just be banned. But also, Australia just banned this for kids under 16. And it does nothing. According to Business Insider, more teens are now using it in Australia. I don't like the fact that we are in a territory where screen time is even close in these dimensions. That's the kind of problem we're at. We're going to look back in 10 years, and Tim Cook is either going to look like a fucking genius or an idiot. And Apple will either be a really important company or have completely missed an era.

0:06

Why hello, friends. Welcome to More or Less. Sam, where are you? I'm at the pool because it's a beautiful day. So for those of you who have to watch on YouTube, Sam will only record podcasts in bodies of water. AKA, I'm told my facial expressions make the show. So please watch on YouTube. But also, you will find that Sam, who I can attest has a real office, is 10 feet from his perch on the lounge chair. It's a beautiful day. I wanted to get some vitamin D. Okay. You sound fine, actually. Let's be fair. I'm a little jealous. I think you look very happy right there, Sam.

1:05

Yeah. I figured out how to record Riverside on my phone reasonably, so why would I ever use a desktop again? Well, shouldn't your agents just run it all? They do. The agents are hard at work. You don't even need to use your own computer anymore. The agents run them all. I'm slightly worried because Sam is training for a marathon. And I think to get his long runs in, he's going to start doing them during this podcast. I could do that next week. I did a pretty intense workout a little earlier, but the good news for you is I don't generally do the long runs on Wednesdays. Okay. Oh, thank God.

1:38

Well, Morins, what's new? Britt, once again, your hair really delights. It really needs a callout. Another reason to watch the video, all of you audio listeners. It's really magnificent. Well, I'm saving time in my day because I'm so much more efficient with my agents. And I do have a second display now. So I have my agentic display and then my productivity display. And now I have more time to do my hair. I've gone the other way. I have so many displays in my office, and yet here I am on my phone. I have one phone I'm on to buy things and the second phone to record More or Less. And that's what I need. Wait, you're really double-phoning it right now? Yeah, of course.

1:57

Instagram. Okay, great. I'm really glad we're not interfering. By the way, weren't agents supposed to render the UX obsolete? Why do you need more screen real estate? Turns out it's going the opposite direction. We're not there yet. I'm waiting for agents to become invisible. My agentic workloads have gotten very high. I have typically five to six agents running at a time. And many times, I'm also sending one agent to the other agent, meaning sending Claude to GrokBot or something like that, not just Claude to Claude.

2:07

How do you even know how many agents you're running? Aren't they just a cloud of things being done? I don't know how many agents I have running. They're just a cloud of things. I still don't trust all of them. We have abstracted our workflow, Sam. We have abstracted. We are the less. They are the more. Yeah. Well, to be clear, I have lots of things happening. I just don't think about them as agents. Yes. But Sam, you're using agents in a very strange way. And you talk to your agent over email every day. I don't do that.

2:32

Guys, I got to tell you, listen, the email agent thing, I was on with an entrepreneur, actually two in the last week. And he's like, well, I call the agent and I text the agent. I was like, yeah, but email is by far the best. And he's like, yeah, it is. Email's sick. The multithreading of it is so good, right? I don't know. Anyone who's texting with their agent doesn't get it. What? I have multithreads with my agent and every agentic harness I am using. But how do you manage, how are they talking to you? How about everybody just manages it how they like it? Who cares? Look at Dave. Well, you're welcome to do that and be less productive, but I'm by the pool.

2:47

We have a packed agenda. We have a major Meta settlement. We have a changing garden. Yeah, let's get on with it. No, I'm going to tell one agent story because it's my prerogative. I'm clearly the least agentic among us, but I have had a few nits with Information headlines of late because the news cycle is moving so fast that our amazing team is just getting the stories out there. We're not always giving our headlines love and care. Clearly a problem for AI. I woke up early, didn't really want to get out of bed, but I was like, I should feel productive. So I emailed my agent. Ah, email. What is with you guys emailing your agents? No one uses email.

3:16

Because I wasn't going to log in. You don't have any stuff. I have six passwords, right? I said, can you build me an app? And my bot's trained on the corpus of Information history, right? It has a lot of context about The Information. And I said, I want it to suggest three headlines for every story. And then I also want a section where you can put your headline in and you can get a score, one out of five. And I want it to be red if it's a three or under. So you feel very scared about publishing it, all this stuff. And I don't know, I didn't check my phone again for 20 minutes. I have a full-on app that is live, password-protected. I demo it to my editors at a 9:15 meeting that day. Because they're very smart, opinionated people, they instantly had three points of feedback. In the meeting, I say, we figured out that based on the best-performing headline, it has a lot of biases. And I said, go from three to six options. And it's updating while I'm in the meeting.

3:18

And it's really cool. And the cool thing is, A, just to be clear about our policy, our policy is editors can override this at any point, right? This is a tool for editors. Editors are in charge of what the headlines are. But I'm getting amazing feedback, and I built it in an email. This is one email. Dave built a social app during our date night last night. To be clear, this has been being worked on for a week. Okay, fine. On Sunday, he started it. I added some stuff to it.

3:32

Okay. Well, everyone knows. I'm sorry to cut the mic because you guys all have more robust agent stories. But for those of you listening to this pod, I'm really proud of you, Jess, that you're in your agentic era. Who think you're not a damn Brit or Sam, help is on the way. And it's really cool. Well, especially if you have my version of the deployment going. Yes. Then, like most of you out there in your organization, perhaps you. This caused me to check in on the state of our AI tooling across the company. And let's say it's wholly inconsistent. So that is my next getting ready to negotiate with Anthropic. Wish me luck. Good luck with that.

3:46

Anyway, big win. Okay, guys, we're going to start with some big news this week, which is Meta's settlement with 47 states and some other entities over the collection of child harm safety cases that have been embroiling it for quite some time now. $17 billion settlement that also came with more limits on screen time, more notifications, more of what we've seen in terms around the parental controls. And interestingly, it has an entire campaign attached to it where Meta is asserting it is now the gold standard on all of this and is trying to goad YouTube and others to join it. I like that, by the way. That was smart.

3:51

Of course you like it. It's completely up your alley. Dave, I'm going to go to you first. You have been very outspoken, rightfully so, about social networks and mental health. What do you make of this move? There are still other cases out there, but this is one of the big kahuna, potentially big tobacco-like cases that was hanging over the company. It seems like a nothing burger to me. The amount of money is de minimis. Doesn't matter. They were spending like $2 billion a quarter on

3:58

with more limits on screen time, more notifications, more of what we've seen in terms around the parental controls. And interestingly, has an entire campaign attached to it where Meta is asserting it is now the gold standard on all of this and is trying to goad YouTube and others to join it. I like that, by the way. That was smart. Of course you like it. It's completely up your alley. Dave, I'm going to go to you first. You have been very outspoken, rightfully so, about social networks and mental health. What do you make of this move?

4:08

There are still other cases out there, but this is one of the big kahuna, potentially big tobacco-like cases that was hanging over the company. It seems like a nothing burger to me. The amount of money is de minimis. Doesn't matter. They were spending like $2 billion a quarter on legal expenses or something. How much? I remember $2 billion. I don't remember if that was a quarter or... But still, $20 billion to Meta is nothing. Right. Agreed. Agreed. Over 10 years. This literally is a nothing. Looks to me like you can just abuse children, get away with it, and then get slapped on the wrist.

4:22

Defend that they're abusing children. I'm not going to let that fly. Unpack that. Spicy. It's like I'm not alone in asserting this. Meta's systems are optimized for engagement. They addict you to the feedback loop in the algorithm. It's so good. Is YouTube abusing children? Absolutely. Snap abusing children? Absolutely. And they do this with impunity and with virtually no consequences. And I think these AGs could have taken this much further. I'm actually quite surprised that they settled for so little. What would you want them to get? Meaningful pain. Also, it's over 10 years. So it's like 1.7 billion a year, which is nothing.

4:44

Yeah. I also should say the amount could be a maximum. And also, I believe, which again, we can turn this piece over to Sam, but it will pay more if Snap, TikTok, and YouTube also agree to things. I liked that. That was clever. At this point, I think that we're probably in a zone where social media should just be illegal, similar to alcohol. Whoa. Wait, alcohol is not illegal. If you're under 18 or 21. Probably under 21, you shouldn't be allowed to be on social media. It's that bad for mental health. What would have become of poor you and me without AOL Instant Messenger back in the day? Messaging apps are a different story. I think messaging is probably fine.

5:11

Is Snap a messaging app or a social network? Other than group messaging. How do you define a group? How do you define a message? No, Sam, that's an excellent point. That's an excellent point. I don't know. I honestly think this is a totally logical conclusion for everyone. My sense is Meta, probably like everyone, there's a bunch of emails you just don't want out because that's how these things work. You got to settle. You got to give them a pound of flesh. Well, you're making a practical, like it was wise for Meta to... But it's not a pound of flesh. It's not a pound of flesh, Sam. The limits are a pound of flesh. I mean, it's fine.

6:08

What's the limit, I guess? Give me a little bit more... Like two hours a day? My response so far is only to the money number. I don't know what any of the rest of it is. Here's what the deal includes. Ready? A default two-hour daily time limit cumulative across Instagram and Facebook that teens can only turn off with a parent's permission. Likes will be hidden on teens' profiles and those that they interact with. School mode, or a mute on notifications, between 8 a.m. and 3 p.m. during school days. Prompts noting when a teen has hit 15 minutes of continual use. Then notifications when cumulative use has hit 60 and 90 minutes.

6:43

The ability to choose a feed that is not driven by an algorithm. The ability to turn off autoplay for videos and content. But neither of those are defaults, so let's point that out. Yeah, the devil's in the defaults, so it doesn't matter. Extreme makeup filters will no longer be accessible. That's a key for girls. And... So you do it in grok and bring it over. For that two-hour time limit, it will be reduced to a one-hour daily limit if TikTok, Snap, and YouTube also agree. Sorry, what do you mean? I don't know what you mean by that.

7:19

Sorry, I'm all for Meta taking out an ad in the New York Times tomorrow, which I think they're doing, saying we agreed to these terms. What about you? That's fun corporate business warfare. I love it. But saying that, you're imposing a safety tool and you're going to make it more safe if someone else does it. But that's a business problem for them. I agree with that, Jess. I totally disagree. Okay, I get my point. You get yours. I'm sharing my point. Let me draw an analogy and then you can have it if you're still awake over there. Do you need a pina colada? No, I'm... You want to bring me one?

8:29

Absolutely not. But I could DoorDash you one. To take Dave's analogy, this is what that component feels like to me. And by the way, I'm in favor of any limits on this, even if I don't think they'll be effective. If you say, we studied alcohol and we feel like it should be banned if you're under 21 because developmentally you're not ready. But if Budweiser and Tito's agree, we'll make it 19. We'll throw our principles out the window. That's exactly it, Jess. You guys are so wrong. I agree with Sam. Or, no, sorry, we'll make it 23. My analogy's off. Jess, let me help you out here. Let's say I make his extremely biased point, Jess. It's not a biased point. It's very simple.

9:28

If society had the balls, right, at a national level, to say we as a society, as a democracy, don't want social media... People, kids use more than one hour of social media today. If we could actually pass that and believe that as a society, then fine. That's the same thing for all the players. This is like the equivalent of saying, hey, if you're Budweiser, because Budweiser, we don't like Budweiser, or Budweiser is the biggest target, has the biggest checkbook, you can't let your beer be more than 3% alcohol, but everyone else can have 6% or 8% alcohol. What do you think the teens are going to buy? They're going to buy the 6% or 8% alcohol.

9:48

It's just like selective bias. I understand why they do it. I don't like the fact that we are in a territory where screen time is even close in these dimensions. That's the kind of problem we're at. It's not. But let's just acknowledge we are, and that's the conversation nationally. And you're not going after everything apples to apples. And so I think Meta is making a really good point, which is, look, if we as a society want to make a decision at a national level that we believe that kids should only do an hour of screen time a day in similar services, fine, then we can all agree to that.

10:19

But selectively going after one provider and not the others, I think it's actually a very brilliant thing to be like, look, we're going to do part of it because we want to settle with you guys. We want to move on from this. Can't believe we want to be an AI company. Let's be done with this. But we're happy to give more money, go even more extreme. You guys actually have the legal horsepower, and society believes that we should only have kids do an hour a day. Fine.

10:21

I agree because, to use the alcohol analogy as a continuation of what you said, Jess, we need to label social media companies that are alcohol brands in this case with a label. And we need to make a rule for all of them as a country. And I agree with Sam that Meta is taking a stab here at trying to do that. But we really need the higher powers to come in and help. Just to throw it out there, YouTube is a way worse, way more egregious offender. Totally. I completely agree. Just so I don't sound biased against Meta, YouTube is the most nefarious of all. And Roblox, I actually think Roblox should just be banned.

10:36

Dave, so here's the thing I don't understand about your argument. I totally agree with that, by the way. Then Instagram is like, okay, we have a two-hour screen limit, but there's no limit on YouTube. Kids hit their two-hour Instagram limit, which is more moderated, at least they're trying, etc., and they just bump over to something that's way worse, And we need to make a rule for all of them as a country. And I agree with Sam that Meta is taking a stab here at trying to do that. But we really need the higher powers to come in and help. Just to throw it out there, YouTube is a way worse, way more egregious offender. Totally. I completely agree.

11:02

Just so I don't sound biased against Meta, YouTube is the most nefarious of all. And Roblox, I actually think Roblox should just be banned. Dave, so here's the thing I don't understand about your argument. I totally agree with that, by the way. Then Instagram is like, okay, we have a two-hour screen limit, but there's no limit on YouTube. Kids hit their two-hour Instagram limit, which is more moderated, at least they're trying, etc. And they just bump over to something that's way worse, where there are no controls. That is good for no one. That only hurts Meta with no upside. So it's not... I'm not arguing against you, Sam. Sam, your point is reasonable.

11:56

And I think we can all agree that this should not be dealt with on a company-by-company basis. And I see, to the degree that Meta is trying to make that point and change the broader conversation, I understand that. I am just deeply uncomfortable with the fact that we are now so openly implicit in all of this. Is this exposure bad? And who should decide how bad? Or is it one-hour bad? Or two-hour bad? Or four-hour bad? And that just makes me pretty uncomfortable. I'm not one, you can see that it's as bad as some people think it is, even close.

12:33

But what I will say is, the story to read here is people being like, look, if you want to have a real national conversation about this, let's have a real national conversation. Don't go after the easy targets. Let's have a real chat about it. And if you have the votes slash the staying power, the will to get something bigger done, great. We're on board for it. But don't selectively enforce, which is, I think, a totally reasonable place to be. I don't think that it's necessarily bad. I'm not saying it should be something. I'm just saying, whatever it is, it's a social, it's a societal conversation, not a conversation about a single company.

13:02

Dave, why do you think YouTube, given the nature of the product, hasn't faced the same level of vitriol and scrutiny? And it has faced some, to be fair, but why? What's your theory on that? Yeah. I'm not sure. Maybe they have a better lobby. I think people don't perceive it to be the same type of product. The social dynamics are different. There's still an algorithm and there's a feed of a bunch of crap. But you aren't looking at what other girls that you know at school are liking or saying about you. They're not commenting on your stuff all the time. You're not seeing that on Instagram anymore anyway. Instagram and YouTube are basically the same product now. Not true.

13:43

No. Most teenagers have Finstas. So they have their public Instagram, and they have their private, just-their-friends Instagram that they post other super pretty photos of themselves to, mostly so they get their friends to comment on it, like it. So the popular kids like them. The boys like them. Wait, where do the pretty ones go? Do they go to the public or the private? I'm so confused. I don't know, and I don't know that it matters. Explain teenagers to me. Okay. There's two feeds you have. So you have your outward one, which is connected to whoever, and your mom and dad know about that one.

14:25

And then you have your secret one that only your select people who you approve can follow. You need to keep up your cover story one. Otherwise, people will go after your fake one. Where's your Instagram? Here's an interesting question. Do you think we ever get to the point where an AG sues Mr. Beast directly? Like, skip the platforms? Maybe. Right? They're like, that would be interesting. It's very hard. It's called freedom of speech, baby. It's a tricky one. But it's all tricky. It's an interesting thing about where... Sam, you have this playing out with much more egregious offenders. Like what?

15:17

I don't even want to say their names publicly, but there are other YouTubers. They transmit content which is far more damaging than, say, Mr. Beast. And you've got nation states going after these people, trying to arrest people and put them in jail. You're talking about Tate? I don't know. Is there the Tate brothers? Whatever Dave's trying to do, don't ruin it. I have figured out, by the way, you can't talk about certain things on our podcast or YouTube won't let you promote it. YouTube is not going to be happy about that. So now we can't promote this podcast because I said Tate.

15:55

But do not talk about whatever happens in Washington, D.C. that starts with a P, because then they will not allow us to promote it. I know. I'm so annoyed. We had a great podcast episode, and I can't promote it because we mentioned the DT name. I don't even know what that is. So it's political speech. We can't. We're trying to save this podcast. I'm not sure how to think about that from a policy perspective. I'm not sure that's smart or right. I just enjoyed Sam arguing with a dear, hardworking YouTube account manager over the state of their political advertising. I was like, it's going to get nowhere fast. You got to go to the top. It's not going to get anywhere.

16:26

It just doesn't really make sense. It's a good strategy for avoiding public scrutiny, but it doesn't intellectually make sense. In fact, it's kind of weird to be like, well, we're not going to allow more conversation that's political speech or promote it. And we're only going to have things about crazy celebrity topics. Well, you remember the Russia Today stuff, right? This was all hatched in all of that. I just don't know how to resolve this. No one does. That's why it's interesting. No, I think we label it and set an age limit and a time limit and the same things. When has that ever done it? What are you? What are you, Jill Biden? Alcohol, smoking.

17:02

No, it was Al Gore's wife who put video game warnings. This video game is like, does nothing. Look, Sam, that's pretty unfair. When you're trying to structure a society in a healthy way, you have to have real conversations about law and how to structure things. And I think this is a very real and reasonable thing to be discussing, which is where should the box be? Where is this detrimental to developmental psychology and brain development? For example, there's very well-established science around cannabis, right? We've now recently, in modern times, the last 10 years, broadly legalized cannabis across many, many, many states of America.

17:24

There's very good science that kids under the age of 18 really should not have this in their system. I know that it happens anyway. But we have to set the bar somewhere, right? We have to say there is good research on psychosis and the outcomes around that for kids under 18 that shouldn't impact. Now we know that it's going to happen anyway, but we have to set the bar somewhere. Right. And I guess that's what this conversation still is not turning into. It's like a gobbledygook, like here's all these different... It's all gobbledygook. It is like Brit literally had to take, what, 30 seconds to read off all of the random stuff.

17:46

Nobody remembers, even listening this far into the pod, what she said earlier. And instead of us having a conversation, kids under 18 shouldn't touch this stuff. And yes, kids are going to touch it. Of course they are. Because we can't ever agree at a societal level or do anything hard. So instead, we argue and come up with these ridiculous sideshow things to placate all the political parties and leave every... I just don't think we can make hard decisions anymore as a society about how we want to live and manage. Well, then people are going to have to make hard decisions for themselves and their families. Amen. And communities have to do this.

18:09

This is why I don't think putting the blame or regulate this level makes sense. Dave, I'm with you. I would label this. I would label it. That is why this became what it became, right? The reason that this is flooding through to the... how many AGs were involved in this? 47. It's insane. It's like more than there are states. It's like 83 AGs. But also, Australia just banned this for kids under 16. And it does nothing. Parents feel helpless, man. According to Business Insider, more teens are now using it in Australia. all the political parties and leave every. I just don't think we can make hard decisions anymore as a society about how we want to live and manage.

18:45

Well, then people are going to have to make hard decisions for themselves and their families. Amen. And communities have to do this. This is why I don't think putting the blame or regulate this level makes sense. Dave, I'm with you. I would label this. I would label it. That is why this became what it became. Right. The reason that this is flooding through to the, how many AGs were involved in this? 47. It's insane. It's more than there are states. It's 83 AGs. But also Australia just banned this for kids under 16. And it does nothing. Parents feel helpless, man. According to Business Insider, more teens are now using it in Australia. Oh, according to Business Insider.

19:20

Come on. I love Business Insider. It's my third favorite publication. Shit talking and trolling this conversation is fine. But we're trying to actually have a real conversation about where should this go? I think that just being nothing matters, doesn't care, never going to happen. The reason this is at 47 attorneys general is that parents feel helpless. They feel helpless at the behest of gigantic fucking data centers that have the control over their kids' minds. Parents need to get a fucking grip on it. But it's true. I'm going to start peeing right now. Here's the thing. Guys, I hate YouTube for our kids. I think it's awful, right?

19:54

I keep getting to watch 80s and 90s comedies with me, which is just so fun. And every time they're on YouTube, I'm sad for it. But you know what? I don't even personally have the balls to actually fully ban it for them. It's a home issue. It's a parent issue. I'm not saying it's not a thing. Every family has different setups. If we could have a national conversation about it, I'm actually down for that. I think that's the right of a healthy civilization. But we're not exactly healthy right now as a civilization in terms of our ability to do anything hard or have big conversations at a national level. Because why? Because social media has broken us apart.

20:26

Oh, social media destroyed everything. No, it's not that it destroyed everything. It's that the current structure of it makes it very hard to agree, to find common ground. That's the truth. Well, I think the thing is, it's always been hard to agree nationally on anything, which is why America moves slowly and the federal government isn't designed to do these types of things. States are, communities are, they've all been blown up in favor of a national integrated economy. And the problem is the federal government is just not set up. And you can't do a democracy well with 300 million people. Okay, we've really extrapolated to many layers.

20:48

First of all, I don't think that I am pro giving parents more tools, labels, all of that, or does parents having balls. I'm sorry, you guys, but parents are outgunned. I was going to say today, so I have all the resources in the world. I have all the awareness of the pros and cons of these platforms. And on a regular basis, I have to gear myself up to present my children with alternatives to this stuff on the same level as if I'm giving a keynote presentation to 300 people. Last night, I was like, I know the kids are going to want to watch some YouTube thing. There are very limited YouTube things that I allow. Go Mark Rober. And I was like, I need to present options.

21:06

So I'm like, I know there's this Disney Pixar series called Win or Lose that's based on the Inside Out emotional competency thing that's really good. Okay, is there an episode they haven't seen? Okay, good. I'm going to take the time when they're watching this episode to get six other shows like this episode because even when I'm scrolling the kids' areas of these apps, it's crap. Total shit. Honestly, it's the end of a long work day, a lot of shit. It took centering myself, speaking my truth, a lot of mental work to not just put on whatever plays after Mark Rober. Okay, and I have a ton of resources and awareness, competencies, all this kind of stuff.

21:30

It is simply too much to say all of this is on parents. Especially because half of the kids in America don't have their parents' home or childcare because the parents are working more than one job. And the kids are left with an iPad at home or left at a daycare center or something that lets them watch screens. Nobody has any chance. These clouds, I mean, I don't think Meta publishes it or Google publishes it, but my guess is that it's around 10 million computers in these clouds globally, if not tens of millions. And no one brain has any chance. Parents have absolutely no chance.

22:00

And I really think that's what this is about, is that your brain, your singular human brain, is outgunned. And with AI happening, it's just getting even more intense. The power of the compute is even stronger. By the way, it's not just kids. I want this for my parents, too. Guys, I want this for me. I want this for me. And for me. But I'd rank it kids, boomers, then us. We're hanging on. As Sam scrolls Instagram. No, I actually am scrolling Instagram. Sam, why'd you move inside? Little lawnmower noise? What's going on? It's getting too hot. It's too sunny. You need air conditioning. I needed vitamin D, but not too much.

23:07

But you're still putting your hands behind your head like you're, you know, in Ibiza. So, very nice. Neither of us have been to Ibiza, so I don't think we know what that's like. That's good, actually. I hope neither of you ever go. Maybe Jess has, and you just don't know about it, Sam. I mean, God bless if she wants to go to Ibiza. I'll go with you, Jess. Girls' trip. The societal biggest-level picture thing, where we can talk about individual liberty and the harm principle and John Stuart Mill. He invoked John Stuart Mill. We can talk about liberty and this extensive it and things like that. I think it's all really interesting conversations.

23:28

Well, we can't really have an intellectual or big-picture conversation about this or AI. So instead, we're going to look for reasonably easy targets and then selectively ask them for some money and some changes. It's just that good because I don't think it does anything. I think it's all theater. You started out by saying it's a nothing burger, Dave, in terms of a slap on the wrist. I'd say it's a nothing burger for actually helping anything. Totally. Right? And that's kind of the point, is it's all theater. Right? And it's a huge amount. The only ones who do well in the theater are the lawyers who make a bunch of money on it.

24:00

And you get some AGs who get to say, I signed on to a thing. And we won. And so everyone needs to be able to say they won. But nobody won. I guess that's why I called it nothing burger. Nothing got changed. Nothing's going to improve. No one wins, but everyone gets to go to their constituents. It's kind of like how the hell you get out of Ukraine or Hormuz. Everyone needs to be able to claim a win and then realize we probably shouldn't have gotten into this fight in the first place, or the way we set it up was wrong. And I just think, I wish that we could have real conversations.

24:30

And I actually appreciate Meta at least being like, look, we'll do what we need to do to give you a win. Let's move on. We're happy to do some things. Whatever. But also open the door and be like, hey, if you want to have a more real conversation, we're down for it. But let's do it in an actually sensible, logical way as opposed to this being a selective enforcement game of politics. It's like you're fighting last generation's war. We live in the real world where this stuff is just happening. And so I think we need a lot of different bites of the apple. Okay. Orthogonal idea. AI is going to be so much worse.

25:00

So then what's the, why are we dealing with last generation's war? Okay. Britt is going to kind of culminate this topic. Britt, bring it to a coherent, forward-looking, intraciteful club. I was going to expand it. Okay. We just need to move. So you take us anywhere. I'll expand it and then we can move to the next topic. Okay. But let's do it in an actually sensible, logical way, as opposed to this being a selective enforcement game of politics. It's like you're fighting last generation's war. We live in the real world where this stuff is just happening. And so I think we need a lot of different bites of the apple. Okay. Orthogonal idea. AI is going to be so much worse.

26:10

So then what's the why are we dealing with last generation's war? Okay. Britt is going to culminate this topic. Britt, bring it to a coherent, forward-looking, intraciteful club. I was going to expand it. Okay. We just need to move. So you take us anywhere. I'll expand it, and then we can move to the next topic. Okay. But I think a lot of the hubbub in Silicon Valley right now is, well, making agentic apps is so passe. The new thing is, what if you can create a whole agentic operating system for yourself and just make your own OS? You don't need iOS and the App Store or Android and Google Play.

27:17

And then parents can make their own children's software stacks, and then they won't skirt around the opportunity to even get onto YouTube or Roblox or any of these things because you're controlling it at the OS level. This is not an unreasonable long-term thought, that part of what has made it hard for parents to actually enforce their ideas about what is good parenting is the structure of the technology, the software itself. It's very hard to change the software right now, but that's changing rapidly, right? I think one of the most interesting emergent ideas that Britt is talking about is this notion that all software should just be changeable.

27:34

We've lived in a world for 30 years where software is put in a box and delivered to you, and you don't get to change it very much. You can maybe ask the creator of it to do this feature or that feature. But we're moving into a new world where you can actually just prompt and change your software. Sam, you've talked about this extensively. You're running a lot of your own stuff on your phone now. I am too. And maybe Britt's right. Maybe in a few years, you're just going to be able to be like, please get rid of all of this. The phone will just finally do it. Versus right now, it's really difficult. I think this is good. Can I embrace and extend?

28:03

So, one, interestingly, I think if you gave everyone, I said, all your software is now customizable, the humans will be like, more boobs. Customizable software is not going to lead to restraint. People are going to be like, I want more powerful marijuana strains. They don't want less powerful shit, right? And so... But that's fine for adults, Sam. If you want your OS to be all boobs, fine. I do. But the thing I would say, in terms of structural solutions, let's get into structural solutions. Here would be a law that I actually think I would really get behind, that could make sense.

28:21

Which is, every OS and app produced has to publish a feed of every action of children that their parents can watch with an agent and know what their kids are watching and can have a conversation. They can't stop them necessarily for version one. And we say, as a rule, if you're a parent, there has to be a simple way to get a feed of everything your kids are consuming under any app they're using. I would love that. Then you're like, okay, this is a rule for everyone. This isn't just for Instagram. It's like, look, yeah, kid, do what you're going to do on Instagram. But I get to see everything you're looking at and clicking on.

28:40

What you're going to watch on YouTube, for now. But I want the log of it really easy to access. And then the obvious next step is I want to be able to have a conversation with you about it. In theory, I want to be able to block you from channels in the future. But if you said to me, hey, until you're 18, I'm not going to limit your screen time. That's stupid. I'm not going to try to come up with these netting rules. It's a very, very simple concept. Simple, low level. If you're looking at shit on your phone until you're 18, I have an easy way to sidecar. Well, I'm going to block some shit. That would be amazing. I mean, this doesn't exist.

29:13

I mean, it's important to point out, Sam, that you can't get this right now. The meta API, the Google API, it's not that hard. Come on. Well, I'll give you an example. These are database logs. There could be an API for this, no problem. Even iMessage, take iMessage. They have all these fucking controls. Who can talk to your team? It's too much. But if you said to me, and by the way, I have a desktop sitting in my room where I just logged into my son's iMessage account, and I sit there and fucking screenshot it all day long algorithmically because there's no APIs to it, so I can process what he's talking about and who he's talking to as a parent, right? That's insane.

29:43

It should be really easy to just be like, I get CC'd easily. And by the way, this is beautiful in the agentic world because in an agentic, before we, it's really hard to review all that stuff. I don't want to click on everything. Now you're like, look, I can just run Claude on it and care. These are my values. What should I care about? Things like that. But you could just be so much smarter about the product decisions and, policy-wise, what you're actually shooting for in a way that enables parents and helps people. Right. And people aren't. But again, you need that at the OS level, right? You could also just go after the big ads. Yeah. It's like, I just need a meta API.

30:24

I need a YouTube API. To Brit's point, at the OS level. This is why I always think it's so funny that Apple avoids all scrutiny on this. Apple could easily, with their App Store review process— I don't think they avoid scrutiny on this, man. On the kids stuff. They get to blame— I mean, the stuff like age verification stuff will be so easy for them to do. But somehow it's Facebook's problem. They're pretty good at pushing off blame to others. But to your point, it's everybody's problem. It's not just Apple's problem either. But I'm saying they have the, because of the App Store, they have a very nice—

31:14

Google Play and app stores have a very nice low-level way to enforce and hook into all this stuff. So it'd be good to do the OS level. You can't do it. They're fine. You just say after you have more than 5 million users, one of the requirements is that parents have the option to get the full stream continuously of data, right? But it is at the OS level because it is also group chats on iMessage, Spotify. Yeah, all these things. Dude, what a mess. I have an API that grabs my continuous Spotify and listening history. It's super interesting. I actually know when I go to bed and when I wake up.

32:11

There's all sorts of interesting facts about what I listen to when I work late at night. It's fascinating to look at Spotify at scale. But I actually would love to have the API of what my kids are listening to. Because if they're listening to WAP on repeat, I want to know. There's no reason that you can't just publish that, make that easy. So I think there'd be, this is what I'm talking about in terms of lazy policy. Smart people, you could come up with a few things that would really make a difference and wouldn't even be that objectionable in most cases to anyone. Could really move the needle.

32:36

But instead, we're stuck in AG pissing matches about stuff that doesn't matter. I like this API publishing thing, but we also need solutions that are dead simple for non-technical. That's pretty simple. Just being able to see a log of what your kid looks at. And when are they looking at their device in the middle of school? I don't even know. I mean, I know because their kids don't have phones, but if you're a parent, it's just simple. It's like, hey, and by the way, especially with an agent on top that you can plug any agent you want into, be like, hey, is my kid paying attention or are they chatting with their friends at 2 p.m.?

33:09

I'll figure out what to do with that, right? But there's no way to even know right now. And it wouldn't be that hard to make it very transparent. But instead, we're stuck in AG pissing matches about stuff that doesn't matter. It just, I like this API publishing thing, but we also need solutions that are dead simple for non-technical people. That's pretty simple. Just being able to see a log of what your kid looks at. And when are they looking at their device in the middle of school? I don't even know. I know because their kids don't have phones, but if you're a parent, it's just simple.

33:44

It's like, hey, and by the way, especially with an agent on top that you can plug any agent you want into, be like, hey, is my kid paying attention, or are they chatting with their friends at 2 p.m.? I'll figure out what to do with that. Right. But there's nobody to even know right now. And it wouldn't be that hard to make it very transparent. By the way, the one person Arsa does text with on his iPad has very technical parents who have figured this out, because every time he texts or calls, they reference it to me. Or their agent references it to you. Parental controls that people, this is the thing, is they are so complicated, even

34:10

as a pretty motivated, reasonably technical person. Our universe for parental controls is a mess in our house, total mess. And it's irrecoverable. And I think that's the bigger thing, is you can come with more of these things. You talk about things people can't use. This is what people can't use. You just need to strip it down, allow you to have conversations. Can we make alternatives easier? We have a company that's working on this at some level called Clear Space. So they're taking almost a VPN approach to understanding everything that your kid's doing and letting you speak in English to the AI and say, I want them to do this and not this and this and not this.

34:38

And it starts to implement some of the controls for you. So there are definitely startups taking a stab at this. Isn't it insane that you actually literally need companies that are trying to layer parental controls? It's such a fucking mess. I was saying, Derek, can we make soccer more? No, because we've also ruined soccer. Yeah. Have you seen the latest write-ups about that? Was it Time Magazine? Everybody I know that's got kids in soccer is like, please. It's not just soccer. It's every competitive sport. It's ridiculous. American capitalism sports. Oh, look at us, parents decrying the state of the youth. I bought my kids a model airplane that I built.

35:11

That's the whole idea. It's like a drone technical, like it probably has AI on board. It probably has a GPU. No, it doesn't. It was made in China, so it might have a GPU in it for spying purposes. Okay. Well, we're interested in your opinions on this topic. No, we're not. I am. I said that so Sam could say that, and then we can move on. So. There are a couple other things we want to get to. Worthy of spending the amount of time that we did on that topic. Dolly Parton. We can talk about Dolly. But first, I want to talk briefly about the imminent passing of the torch of Tim Cook to John Ternus. Information reported this week about Tim Cook's farewell. One Republic played.

36:00

Laureen spoke. Tim acknowledged his partner, Mike. And Dave, Britt. Apple of files that you are. Is this a big moment? How are you feeling? Yeah, it's a huge moment. I mean, end of an era. End of the era of supply chain optimization? I mean, Sam, when you take over a company at the scale of Apple and grow it in order of magnitude, you can talk. No interest. I will say, did you see the chart that I posted that Claude made me? No. I'm very proud of this. I said, hey, Claude, I want a chart that basically talks on, obviously, a zero, a percentile basis. You can baseline it. When Tim Cook became in charge of Apple and when he joined the board of Nike and the market

36:50

cap of the two. And it's hilarious because he joined Nike five years as a board director before he became the head of Apple. And if you look at the charts for the first 15 years, or whatever it is, 15 and change, which is the time I said, that is literally Nike's maximum stock price. And it is now back down to basically where he started. And so it's really funny if you look at the chart, which I will post a link to if I remember. So what happened? This is me being really cynical, snarky. But the cynical, snarky take is, in about 15 years, you can supply chain optimize the hell out of a company and make stock go way up.

37:17

But then Apple better figure out how to not be Nike, right, and come up with the innovations and stay ahead of the market because literally it's almost like, to the month, this chart implies he's leaving Apple at the maximum price possible for a Tim Cook-style optimizer. But it's obviously just a silly chart, but it is funny to look at. Do you guys think Tim Cook's legacy, the extent of his legacy, has been overlooked? That his legacy is sort of the successor to Steve, but does that? I don't think so. I feel like he's very respected. I guess it depends on what you mean by that, Jess, but there's been a lot of thoughtful writing

37:52

and content created about the work that he's done. I mean, it's a pretty hard act to follow. And if you were to sort of go point by point on the Tim Cook legacy, it's pretty hard not to look at it and say he followed a really tough act. He made this thing into a multi-trillion dollar company. They did launch a lot of great products. They established a dominant new position in certain form of silicon that emerging is a major, materially different strategic move than everybody else at their scale. Was he the Jobsian innovator that came up with the iPhone? No. But I feel like he's quite respected for what he did do. And I don't think there's anybody out there panning that.

38:31

I'd also argue that the era he was CEO, interesting, but a little boring. There was a mobile revolution and a social revolution in the 2000s that Steve truly took advantage of. The 2010s were a lot of SaaS things and, I don't know, it was fine. A lot of stuff became big, but now we're going into the agentic era, and it's a whole different beast that requires a different set of innovation. And so maybe there's some interesting blessing in disguise here. He did what he does best in the era that's about scale and SaaS and optimization and all of this stuff and commerce and cloud. And now Ternus can take over and hopefully come up with some magic for the next era.

38:51

My first story covering Apple for The Wall Street Journal was Memorial held for Steve at Stanford. I remember when I drove you to drop you off at Stanford. You did drop me off for that. I wasn't invited, in case that wasn't obvious. I did a string of stories, as young beat reporters do when they start beats, which is anytime Tim issued a policy. They started doing more philanthropic matching, which is something Steve hadn't done. And I got a lot of news cycles over the early days of Tim Cook and enjoyed the limited number of times I interviewed him. But I think it is interesting to look back and see, not to forget how much in the shadows he was of Steve.

39:14

And I think there was, to your point, Dave, the results kind of speak for themselves. The financial results are unbelievable. Yeah, no, they are. And I think that's what I meant. But I also thought it was interesting, and one of, presumably, we're sort of at the end of his outward communication of pitching products. The Apple announced the new Mac mini. Did you see, with no AI, with claymation? People were panning that on the internet. Well, the whole commercial for it is a bunch of claymation, which is very fun and cool. But how do you not use AI in a product that is fundamentally about AI?

39:32

What was interesting to me is how much the press release and Tim's own post on X talked about it for AI workloads, which is something that we have been talking about for more than a year, and I know, Dave, you're very passionate about. Honestly, it still feels a little weird to me when Tim talks about AI because I feel like it has not been, they've just taken such a different tack. And then it felt like they had the Apple AI, Apple Intelligence launch that largely did not pan out in the way they wanted it to. But it's certainly looming over his exit and what's to come. But it's also nice, as we've talked about, too, over the last two, three years, the conversation

39:56

about AI has been so co-opted by large language models or a certain handful of apps and certain set. And it's always been much bigger and broader, more interesting than that. What was interesting to me is how much the press release and Tim's own post on X talked about it for AI workloads, which is something that we have been talking about for more than a year. And I know, Dave, you're very passionate about, honestly, it still feels a little weird to me when Tim talks about AI, because I feel like it has not been, they've just taken such a different tack.

40:02

And then it felt like they had the Apple AI, Apple intelligence launch that largely did not pan out in the way they wanted it to. But it's certainly looming over his exit and what's to come. But it's also nice, as we've talked about, too, over the last two, three years, the conversation about AI has been so co-opted by large language models or a certain handful of apps and a certain set. And it's always been much bigger and broader, more interesting than that. I guess Sam would say it. AI is matrix multiplication. It's a type of compute operation that requires real hardware engineering.

40:15

And Apple has always been two companies. It's been a hardware engineering company on the one hand, and a design and software engineering company on the other. And they're completely, in a way, completely different companies. And a supply chain company. Sure. These are completely different modes down there. And everybody actually interprets Apple and its innovation through the software side of things, because that's the buttons you push on and things like this. But Apple is insane at doing the hardware engineering side of the equation and made some extremely smart decisions under Tim.

40:37

The system on a chip, the SoC decision, I think, will go down as one of the most strategically important technology decisions that happened in Silicon Valley in this era. Because the ability to run agentic AI workloads on this chip is extremely, it's just really, really good. And their ability to do it at very high bandwidth with large amounts of RAM and their integrated GPU, CPU, RAM clustering. I mean, it's just amazing stuff that they had a lot of foresight on. And that's because on the hardware side of the house, they understand atoms and heat and energy and how to make boxes filled with the right stuff.

40:53

And so I just think that that's a very misunderstood part of the equation here. And that he's going to go down in history having made a really important strategic choice that's going to play out over the next few years. I do think, Dave, the thing I think you and I can agree on is we're going to look back in 10 years and Tim Cook is either going to look like a fucking genius or an idiot. And Apple will either be a really important company or have completely missed an era. And we don't know yet. I think it definitely will be defined by the AI, what comes next. I don't think it's going to have missed an era. I mean, I don't know that. I don't know that.

41:17

But the scale is tipping towards... This is the question. Look, there's a lot of ways in which Apple actually will come out of this era looking hilariously good. They didn't sink all the money into data centers, right? They're going to do it on top. They're going to get to sit on top of every model as the last mile to consumers. They can maintain that. It's very possible in 10 years that Apple is laughing all the way to the bank and people will be blessing Tim Cook for having fallen into none of the traps, didn't build a big research team, didn't do data centers, didn't get snowed. Then people will be, a hero.

41:44

If that's not true, and it turns out that Apple is internally dysfunctional, can't get AI stuff done, all the negative stories, then the story will be Tim spent the last 15 years optimizing last year's game. It's classic innovators of Memo. He missed the big swerve and he's going to be vilified, but we won't know for 10 years. So Sam, if you had to place a bet today on that, what would you place it if you had to today? We should note that our seven-year-old is putting all of his allowance money into Apple and is very excited about it. He is. It's amazing. He literally was like... And he wants to do it now. He's like, I need all of my cash in Apple. Why?

42:24

I have not really peeled back the onion on why. Is it the Ternus move? Is he a Ternus fan? No, he just looks around and he sees Apple products everywhere. It's not rocket science. He's really into the idea he can finally afford a whole share and he wants to vote. Well, I told him about shareholder meetings, which got him very excited. Try and teach the element of business. But Sam, you're not off the hook. Which do you think it's going to be 10 years from now? Oh yeah, or give a percentage. You can give what percent. I love the humor of it all. So I'm going to bet that it's one of these accidental genius moves, right?

43:25

Which is that he goes down in history and Apple looks fucking great because they avoided an incinerator box of cash and they're going to be on top in 10 years. So I'll take the pro Apple case, largely because I think it will be hilarious if it plays out. And I think it's quite possible, but I don't know. I think I'd take that case too. Not that I think there's going to be necessarily an implosion, but Apple's a very hard company to bet against.

43:42

I mean, my journey covering it at the Journal was like, what is the iPhone number this quarter? What is the iPhone number this quarter? This thing's supposed to disrupt the iPhone. This thing's supposed to, oh wait, it didn't. It went up. Wow. It turns out there are customers in China. Oh wait, there are fewer customers in China. It's still going up. Right. And now at some point that does get disrupted, but I think it's very hard to predict exactly how. And I don't think it's going to be disrupted by OpenAI's puck or whatever they're launching. Right.

44:06

I mean, if you believe the whole idea of the bitter lesson, this idea, this is one of the animating beliefs of the AI community, that the more compute you throw at a problem, better things get. If you were to narrow down to a single decision that mattered to answer this question 10 years from now, it's the Apple Silicon decision, right? They got into the game at a level that nobody in this time period did. And they're going to be in an extremely strong position. Had they spent all of their money optimizing software or building data centers or whatever, those are decisions that could have had dramatic impacts on their future. But they didn't.

44:24

They became good at the one thing that supposedly is the main thing that matters in this era. I would say that the 10-year future is already written. But they're no longer TSMC's biggest client. They don't have a bunch of computers in a warehouse. We'll see. Dave, you're going to be long Apple. Where are you, Britt? I'm long. I'm long Apple, too. I don't think the iPhone's going anywhere. And I think that locks people into all kinds of things. And so we in Silicon Valley see change so much more rapidly than the rest of the world. But it's going to take a long time. So I'm with Dave. That's a way of saying we're one big hype machine. But yes.

45:21

Best hype machine ever created. That's why Hollywood hates us. Yeah, we're better than Hollywood. Better than Hollywood. Hollywood didn't have software. Morens, what are you looking forward to this weekend? Give us a hint. What's going on in the world? Well, Backstreet's back, all right. We're going to the Sphere. It's going to be fun. I have to go, guys. I'm keeping someone waiting. Bye, Sam. See you later. Bye. And I'm making Dave go with me to meet my childhood crush. I can't. Howie. Do it, Brett. Do it. And I will be wearing white. Let it be known that I'm not going willingly. Oh, come on, Dave. It's an experience. We'll have stories next week. So stay tuned.

46:28

I can't stand the Backstreet Boys. They were my first concert. You never know until you hear. Some people just really bring it all out live. I doubt that's the... It's probably good. Yeah. It's going to be a great experience. Well, I wish you guys all the best. I will be figuring out how to get soccer cleats for a four-year-old. And there's a lot of tennis happening right now, too, right? Oh, there's so much, yeah. I mean... There's so much tennis. It is. I'm just a little sad to not be at the U.S. Open because it's also a wonderful time to be on the West Coast, and I do consume and watch a lot of tennis. Let it be known that I'm not going willingly. Oh, come on, Dave.

47:25

It's an experience. We'll have stories next week. So stay tuned. I can't stand the Backstreet Boys. They were my first concert. You never know until you hear. Some people just really bring it all out live. I doubt that's the. It's probably good. Yeah. It's going to be a great experience. Well, I wish you guys all the best. I will be figuring out how to get soccer cleats for a four-year-old. And there's a lot of tennis happening right now, too, right? Oh, there's so much, yeah. There's so much tennis. It is. I'm a little sad not to be at the U.S. Open because it's also a wonderful time to be on the West Coast. And I do consume and watch a lot of tennis.

48:25

But the U.S. Open has figured out a way to add a third week to the tournament. So we. Wow. Are currently in what you call fan week, which features a very cobbled-together but very interesting mixed doubles tournament. So Serena and Carlos played mixed doubles earlier this week. I saw. In really an event staged for Instagram. And then Roger Federer, who's getting inducted into the Tennis Hall of Fame this weekend, played an exhibition set against my favorite podcaster in the world, Andy Roddick. And bested Roddick, who did hit an amazing 130-plus. So. Your favorite podcaster in the world is Andy Roddick. Hands down. No question. Okay. I didn't even know he had a podcast.

48:53

It's called Served. I'm obsessed. I listen to every single thing he says. What does he talk about? Tennis over and over again. Yeah. Brett. There's something called fandom. Usually, I plug Served. No, really, I have met Andy Roddick, but that's about it. But it's just so good. So yeah, we're in a tennis moment. And for all of you out there, The Open Racket, the tennis magazine I'm a large investor in, has 10,000 amazing events that I'm having major FOMO about. So go check out their Instagram and go to their events. With that, we will leave our dear listeners and viewers a very, send us constructive comments. Some hot, spicy, but important topics today.

49:43

And we'll see you back here next week for another episode of More or Less. Bye. Bye. If you enjoyed this show, please leave us a virtual high five by rating it and reviewing it on Apple Podcasts, Spotify, YouTube, or wherever you get your podcasts. Find more information about each episode in the show notes, and follow us on social media by searching for @moreorless, @Dave Morin, @lesson, and @Jay lesson. And as for me, I'm @Brit. See you guys next time. Right. And now at some point that does get disrupted, but I think it's very hard to predict exactly how. And I don't think it's going to be disrupted by open AI's puck or whatever they're launching. Right.

50:06

I mean, if you believe the whole idea of the bitter lesson, right, this idea, this is like one of the animating beliefs of the AI community, that the more compute you throw at a problem, better things get. If you were to sort of narrow down to a single decision that mattered to answer this question 10 years from now, it's the Apple Silicon decision, right? Like they got into the game at a level that nobody in this time period did. And they're going to be in an extremely strong position. Had they spent all of their money optimizing software or building data centers or whatever, those are decisions that could have had dramatic impacts on their future. But they didn't.

50:49

They became good at the one thing that supposedly is the main thing that matters in this era. I would say that the 10-year future is already written. But they're no longer TSMC's biggest client. They don't have a bunch of computers in a warehouse. We'll see. Dave, you're going to be long Apple. Where are you, Britt? I'm long. I'm long Apple, too. I don't think the iPhone's going anywhere. And I think that locks people into all kinds of things. And so we in Silicon Valley see change so much more rapidly than the rest of the world. But it's going to take a long time. So I'm with Dave. That's a way of saying we're one big hype machine. But yes.

51:24

Best hype machine ever created. That's why Hollywood hates us. Yeah, we're better than Hollywood. Better than Hollywood. Hollywood didn't have software. Morens, what are you looking forward to this weekend? Give us a hint. What's going on in the world? Well, back streets, back. All right. We're going to the sphere. It's going to be fun. I have to go, guys. I'm keeping someone waiting. Bye, Sam. See you later. Bye. And I'm making Dave go with me to meet my childhood crush. I can't. Howie. Do it, Brett. Do it. And I will be wearing white. Let it be known that I'm not going willingly. Oh, come on, Dave. It's an experience. We'll have stories next week. So stay tuned.

52:06

I can't stand the Backstreet Boys. They were my first concert. You never know until you hear. Some people just really bring it all out live. I doubt that's the. It's probably good. Yeah. It's going to be a great experience. Well, I wish you guys all the best. I will be figuring out how to get soccer cleats for a four-year-old. And there's a lot of tennis happening right now, too, right? Oh, there's so much that, yeah. I mean. There's so much tennis. It is. I'm just, I'm a little sad to not be at the U.S. Open because I, it's also a wonderful time to be in the West Coast. And I do consume and watch a lot of tennis.

52:37

But the U.S. Open has figured out a way to add a third week to the tournament. So we. Wow. Are currently in what you call fan week, which features a sort of very cobbled together, but very interesting mixed doubles tournament. So Serena and Carlos played mixed doubles earlier this week. I saw. In really an event staged for Instagram. And then Roger Federer, who's getting inducted into the Tennis Hall of Fame this weekend, played an exhibition set against my favorite podcaster in the world, Andy Roddick. And bested Roddick, who served, did hit an amazing 130 plus. So. Your favorite podcaster in the world is Andy Roddick. Hands down. No question. Okay.

53:20

I didn't even know he had a podcast. It's called Served. I'm obsessed. I listen to every single thing he said. What does he talk about? Tennis over and over again. Yeah. Brett. There's something called fandom. I usually, I mean, I plug serves. I've no really, I have met Andy Roddick, but that's about it. But it's just so good. So yeah, we're, we're in a tennis moment. And for all of you got there at the open racket, my, the tennis magazine, I'm a large investor in has 10,000 amazing events that I'm having major FOMO. So go check out their Instagram and go to their events. With that, we will leave our dear listeners and viewers a very, you know, send us constructive

54:01

comments, some hot, spicy, but important topics today. And we'll see you back here next week for another episode of More or Less. Bye. Bye. If you enjoyed this show, please leave us a virtual high five by rating it and reviewing it on Apple podcast, Spotify, YouTube, or wherever you get your podcast. Find more information about each episode in the show notes and follow us on social media by searching for at more or less at Dave Morin at lesson at Jay lesson. And as for me, I'm at Brit. See you guys next time.

Reading tools

Type to find a passage

Appearance
Ask this transcript

Notebook

Export Markdown

Add a note