Hello and welcome to the broadcast flagship podcast of Lake America, a federally mandated database change that our nation's mapmakers will now have to contend with. I'm your friend Neil Patel. Jake Kastrenakes is here.
Exciting week, Neil. Glad to be here.
So much going on. Uh, David is here. Welcome back, David.
We didn't scare the audience away with our camera crash out last week.
No, there was a lot more I would have loved to crash out with you about, but there's still time, guys. We don't have to talk about Max. We could just go straight into...
I didn't notice that we were on time last week due to your interference.
I know. Everyone is mad at me for not letting the show go for another two hours. It would have been fun.
I would have talked about 3D models of fake tracks of phones on video and it would have been great.
Oh, but these just are Slack chats put into...
Yeah. Podcast. The problem is, and this has come up, you know, years depending on who all the hosts change, but the demand that Vergecast ever longer exists. The problem is that it just slowly fades away into like a video of me crying about computers.
Which I think people would also be okay with.
That's just the last 30 minutes of your four-hour live stream, just me looking at like an original Macintosh and being like, 'What we lost.'
It's not good. How we got banned from Twitch. Let's see your Macintoshes. There's a lot going on. Jake, what are we doing today?
Yeah, there's a lot to talk about. There was some big beta news. Nvidia might have saved the US economy. But first, we've got to talk about Apple. Big week for Apple news. So, Apple did two things this week. One was announce the next iPhone event, and we'll talk about that in a little bit. But first, they announced new Macs. So, there's a new M6 chip and the M5 Ultra. There's new Mac Studio and Mac minis. I'm going to read Apple's descriptions of all the new devices, and I want you guys to tell me what stands out here. All right.
So, it said, 'The Mac Studio is,' quote, 'the ultimate desktop for on-device AI and the most extreme pro workflows.' The Mac Mini, quote, 'delivers a massive leap in AI performance, supercharging a leading desktop for always-on agentic computing.' And the M6 and M5 Ultra deliver, quote, 'a big leap in performance and AI compute.' So what market are we looking at here?
Well, at least we've gotten away from every computer for creators and creators alone.
This is true. That's a nice improvement.
The thing that I am really interested in here is the way that most people use AI today is through cloud compute.
And Apple has gotten really into bragging about their local AI compute possibilities. And I am interested in how many people are buying these devices for that purpose. Right? If they're advertising a MacBook Air as being, you know, the leading machine for AI compute, I would be very confused. But the Mac Mini is sort of the sweet spot, right? Like those were blowing up at you went to like Claw Con, right? The Open Claw Convention. Is that still...
Are the Mac Minis still the hot device for this local AI stuff? So, I saw some people tweeting a couple weeks ago like, 'Remember when everyone bought Mac minis for OpenClaw?' But all of the replies were, 'But it's still my favorite computer I own, and I don't regret buying it.' So, Apple is in this kind of unique position where it's like people are starting to sort of toy with this idea of local models. You know, maybe the M4 Mac Mini was not the fastest local model runner on the planet, but people see a day in the future where they will relinquish their chains from Anthropic and OpenAI and they're going to be able to run Mini Max or whatever on their 512 GB Mac Studio. So, I feel like Apple's kind of just dipping their toe in that market and being like, 'We know people want this and it's an aspirational idea to be able to unchain yourself from OpenAI and Anthropic.' Maybe we can kind of sell to those people, but if we're not there yet...
We can still sell to all the same people we've been selling to for the last 40 years.
And it sounds very cool that they have all this new additional AI compute abilities. I mean, this is the direction that Apple's headed in some ways, right? Like macOS 27. My understanding is there's already a split happening in where the compute is. There's some Siri stuff that happens locally and some that's in the cloud. Neil, is your impression that Apple wants this stuff to be happening locally as part of their sort of big privacy promises, or is this just sort of marketing still?
This is a lot of marketing. I would actually pull apart the Mac Mini and the Mac Studio claims here. That's fair. So the Mac Mini delivers a massive leap in AI performance, supercharging the leading desktop for always-on agent computing. That's they're just talking about OpenClaw. A leap in AI performance is like meaningless on this scale. Like some AI is happening here. Maybe you're running a local model somehow, but oh, the leading desktop for always-on agent computing is like a bunch of people bought this to run OpenClaw. There's a new one called Instinct which just raised a bunch of money. Uh, and it's a privacy nightmare from what everyone can tell. Like they will literally train on your user data and they got to fix it. Um, but like you're going to buy a small computer to sit in a corner of your house and do computer stuff that you command over WhatsApp or iMessage or whatever. Like that's the Mac Mini. And I think that the AI claim there about performance has almost nothing to do with always-on agentic computing.
Like I think that that's one sentence with two parts and Apple is just excited for you to conflate the two parts. The OpenClaw fans were enough to what, put the Mac Mini on back order for months. Yeah, but this is also a pretty small market.
It's a small market. But what I'm getting at is, you know, if you bought a cheap Windows PC with enough power to sort of like be a computer, you like you would spend a lot of tokens with Claude just trying to close, right? Which is like not a problem on the Mac. And that's like honestly a reason that I think a lot of people bought these products.
In the footnotes of that press release, uh, they're comparing it to the M1. So these leaps are not all they're made out to be.
Yeah. Yeah. Well, I'm just saying I think the Mac Mini is one kind of product and it is in many ways the best sort of price-performance out-of-box computer to let a thing like OpenClaw run on, which is the always-on agent computing, and most of those tokens are being burned in the cloud to control your computer.
Right. And you're just giving Claw a computer and it happens to be a Mac Mini. So that's one thing.
The Mac Studio is the ultimate desktop for on-device AI is a wild claim. Just a wild claim because another kind of desktop you can buy is a PC with an Nvidia GPU in it.
Right. And I don't think there's a world in which Apple is truly claiming that the M5 Ultra competes with high-end Nvidia GPUs at inference. If that's what they're claiming, then we should run some benchmarks and find out. But I don't think that's the thing that they're saying. And I'm very curious to see in this context what that means. It's the ultimate desktop for the most extreme pro workflows, which is the other thing they're saying. Yeah, I buy that. That's the thing it's been good at for ages. But the ultimate desktop for on-device AI, like I feel like a lot of people running actual GPUs on their desktops have a lot to say about that.
Probably true. They do offer a 512 GB RAM model now. They do keep upping the amount of RAM that they can have on these things. So...
I think they're trying to court those people. I mean, I ran OpenClaw. I tried to run OpenClaw. I couldn't figure out how to really get it to work when you insert enough tokens.
Wait, you didn't ask Claude to configure OpenClaw? This is the loop you have to give yourself up to.
Yeah, I'm not Silicon Valley plaid enough for this, but um...
You have to set up one agent to set up your other agent.
Yeah. Exactly. But you have to download, you know, this large local model, you know, and I was running this on like a Surface Book 2 from 2015 or something. Could not really run a large local model on there with 8 gigs of RAM. So, I feel like they're trying to court those people that are like, I want to run the two trillion parameter models. As far as inference speed, I'm not sure, you know, in comparison to like Nvidia's GTX computers or whatever they're called.
Well, this is what's so interesting. Like, I combed the announcements to see all the different specs they're highlighting and the benchmarks and what percentage better all these things are. And you know how Apple is with fake charts and numbers. And when it came to the M5 Ultra chip, which is what's in the Mac Studio, what you can spec it out with if you really want power, they had almost nothing when it came to AI, right? Which I think speaks to Neil's point about they're just trying to add some AI hype there, right? That the Mac Studio is by all means a very capable computer. I think for creatives who are doing video editing. Yeah. I don't know that it's really the thing for running LLMs and doing inference. Yeah, probably true. Probably true.
Yeah, I mean, we're going to find out. They're going to come out, people are going to put them to the test, and certainly we'll see if the upgraded neural engine can compete with the blah blah blah when you're running Quinn, right? Like, yeah, there's a whole bunch of stuff you can do with these computers that are interesting. I think in the context of them also announcing the iPhone event this week, it's a pretty big signal that that event is packed. Yeah.
And here are the new Macs.
And these Macs just kind of sell themselves.
Right. Like you don't need to go market a Mac Mini or a Mac Studio. The people who want to buy them can buy them. And especially the people who are super serious about doing local AI workloads know exactly what they want to buy.
Like those are not consumer products yet. That's not a thing regular people do. It's a thing people who are really invested in wanting to do it are doing right now. And they're having fun. I'm not saying they're not having fun, but they know exactly what computer they want to buy. I'm also extremely curious exactly how much power you can get out of open local models that you can run yourself these days, right? Uh, you can do something on the Mac Mini, you can do something a little bit better on a Mac Studio if you spec it out, but at some level these open models are, they just don't hold a candle to the tip-top flagship models from OpenAI and Anthropic. And if you want to do the beefiest work, you're probably just paying those and you're running your whatever agentic harnesses so you can have a hundred different agents in a bot swarm, you know, run all of your coding for you.
I think it kind of depends on what you're benchmarking against, right? Because if it was like, I think it was MiniMax or the Chinese model that basically was open source and free and definitely distilled on Anthropic because when you asked it what its name was, it says, 'Hi, I'm Claude.' Um, that, you know, whatever benchmarks they released were like 95% of what Anthropic had at the time. And I think what people are buying these $18,000 computers for is the dream that they can be freed of this leash. Whether or not that's actually true, I feel that is a whole another story. I mean, I would dramatically prefer to talk to these things locally and it just does not feel like an option. And to the extent that I even could run something on my computer that is, you know, somewhat capable, I'm like, listen, this computer is going to fall apart if I ask it too many questions for this LLM. Like I don't want to destroy my MacBook Air just to get some bad AI answers.
Okay, so a couple other things about these. Number one, we should point out there are some price hikes here. Everything's getting more expensive. They actually very quietly hid a pretty big price hike for the Mac Mini. Um, all the news is that it went up by $100, but that's because they dropped the 256 GB config this spring and then they added it back. So, it's actually a $300 price hike, which is, I think, one of the bigger ones for Apple. I think we're expecting probably some more next week.
Um, there was one price cut though, and have you guys seen this? I've not seen it in person.
This is big news. This is big news. It is the Apple polishing cloth.
It is now $9 instead of $19. I have heard the legend of this cloth. It is allegedly a very good cloth.
Yeah. Because it comes with like a few different expensive Apple products you can buy. It comes with it. And it's like it's beyond microfiber. It's nanopiber.
No. They would have tried to find it.
But it is very good at polishing things that are already clean.
How is it? How do you know it's polishing then?
Well, wouldn't that be what the polishing is?
Getting the dirt and grime. No, you got to use something else. But if you want to polish your screen, which is why it has a compatibility chart, that's what it's for. But it is a good non-patented nano fiber cloth. Yeah, it's very good.
Wait, hold on. I have a number of questions about, are you really sitting on the word polish as a definitionally important word here because actually to polish means to slightly remove an amount of material, like do you know what I mean?
That's fair. Yes, that's true.
I don't think that's happening on the iPhone screens.
Right. It's just wiping grease away.
Grease. Grease and fingerprints. Yeah. Yeah.
All right. I'm just saying if we're calling it a polishing cloth, what we're technically doing is very finely sanding our phones.
Maybe that's what we're doing. Maybe that's why there's a compatibility list. Like these are the products that you can finally araid in order to achieve a mirror-like shine.
For the record, it only goes back to the iPhone 5. So, if you're going earlier than that, your screen will be damaged.
Here's what I want you to do, David. I want you to make one of those time-lapse videos that's like polishing this phone until it's a mirror, but instead of like one of those cast iron pans, it's just you and the polishing cloth and it's like every day you give it a wipe, but it's like 10,000 years long.
Yeah. My big question is why did they reduce the price by like 60%?
Because they had to say something got a price cut. Do you think that they actually, I mean it did get them the goodwill that they hopefully were probably going for, but is that the reason do you think why they cut the price on this thing from $19 to $9?
Listen, I'm just going to say we reached out to them and said you got to tell us what happened here. Did the materials change? Did the product change? Is the size different? They would not comment on the record. So, I don't know the truth. We're going to have to put this thing to the test. It's the only way to find out the truth about this. I mean, ironically, it did kind of make people just talk about it, which takes away from as much noise about how expensive everything else got. You know, there's some goodwill.
No, here's my, I mean, I'll just do my like Tim Cook supply chain spreadsheet conspiracy theory.
So, you got to imagine Tim Cook. It's his waning days as Apple CEO, right? And he's like, 'What is my legacy going to be? I got to set this kid up for success. He doesn't know what's...' He opens Excel in his Vision Pro because he uses every Apple product every day. So he puts on the Vision Pro and he's got like 40 Excel windows open in mirroring obviously on his Mac Studio and he's looking at where all of the margins for Apple products are and he's like, 'Man, I can't charge a 400% margin on a RAM upgrade on a MacBook anymore. The prices are through the roof. That's gone. Where am I going to reclaim this margin?' So then they re-rack the storage and they bring back 256 gigs, but it's at a huge price increase. And he's like, 'All right, I need to catch some goodwill. Like I'm not going to sell accessories, dongles at 200% margin anymore because I'm going to build in enough overhead on the storage and the RAM to make sure that further price shocks don't knock Jon off his game.' So he's just moving the cells around.
And he's like, 'All right, we're bringing down polishing cloth. I don't need to make the 10 extra dollars here because I'm going to make $30 extra dollars on storage to hold us over until this market stabilizes.' And he's just saying all this to himself in his very charming southern drawl alone in his office in the Vision Pro in the dark. And this is why Apple's share price is through the roof is that kind of superb supply chain management.
I'm so excited about our new RAM prices.
Yes, we are so excited.
So that was on Tuesday. All the Macs, the polishing cloth excitement. Then Wednesday comes, Apple sends out invites for its next iPhone event. Tagline is 'Surprise and Shine.' They are going to run out of these someday. They're going to have to figure out a new approach to the taglines. It's September 9th. It's the Wednesday after Labor Day. Uh, notably a Wednesday because usually they're Tuesdays, but I don't think that they like press having to fly on holidays. They probably also don't like having their employees have to fly on holidays. That's probably the number one thing.
So, this is going to be Turnis' first event as CEO.
And normally this is the iPhone and Apple Watch event. We're expecting the full lineups.
Some things we're expecting different this year. There's supposed to be no standard iPhone 18, just the Pro models, and we're supposed to get an iPhone Fold. I'm curious from both of your perspectives. It's been a big year for foldable phones. Do we think they are compelling enough yet? Right. This is Apple's putting their flag down. They're saying this is enough for mainstream audience theoretically. Do we think that foldables have gotten there?
The price is going to be important. I think it's notable that they just recently announced the leasing program because they're going to be leasing a lot of these foldables. Everyone from my hairdresser to my oil change guy has been like, 'I love my iPhone. I really love my iPhone, but some of those foldable phones on Android are kind of cool. I might switch over to that if Apple made that.'
I am serious. I am serious.
Really? Your oil change guy?
You're talking to Trump voters and diners. I mean, it's after I tell them what I do for a living. So, maybe they're just trying to report.
I just think people are intrigued. I don't... The price is going to be important. Very important.
Yeah. Every time I see anybody with a like a razor, other people always ask about the razor because they're fun to look at.
I think the question of whether you can go into an Apple store and look at something that will almost certainly cost north of $2,000.
Yeah. And be compelled to say, you know what, I want an iPad.
I mean, maybe that is a question, right? Are they trying to make this a mainstream product?
No. No. I think the answer is no. And I think one of the reasons you don't have an iPhone 18 is that you would have no choice but to raise prices on an iPhone 18, whereas keeping the existing base model iPhone let you keep riding the cost curve down. Like all that tooling is long since paid for, right? Like you don't have to pay new cost into that phone. So, you can hold the price, which I think is very important for Apple right now. And at the very high end, you're just like, is this going to work? Not to be totally rude about it, but it feels more like a Vision Pro shot than a future phone shot, right? So, they can introduce this at whatever price they want. They can see what the market will bear and what the worst case scenario is. They attract a bunch of people to come to their stores and check this thing out.
Yeah, I think that absolutely. Again, we haven't seen it. We don't know how it's going to work. Mark Gurman had some reporting from Apple employees who've used it and they're like, 'We love it.' And it's like, what else are they going to say?
Yeah.
But I do think the promise of these devices always rests in how good is it when you unfold it.
And we have just seen over and over and over again that people buy these phones and then they do not unfold them.
They end up using that cover screen a lot. And so maybe the form factor shift here into that more square form factor will lead to a new usage pattern. I don't think we've seen enough people with the Fold 8 to know if that actually happened. But that's the new form factor. And I think Apple has to answer a great question about do you want to use an iPad mini sometimes? And maybe that, if you're reading this as cynicism and it's not cynicism, just Apple basically has to sell people an iPad mini. Like the software experience here is going to be, and you get an iPad mini, and whether or not Apple wants to say it or cop to it or think about it, it's going to be the thing people say, 'Oh, it's an iPad,' when you open it up. And then that comes with all of that. And I'm just very curious to see how that unfolds both on Apple's side and how they talk about it and on the consumer side and how people react to it.
I do think that's going to make for a very interesting event beyond the fact that we're getting new Apple hardware, which is not that common. It's going to be a software event in a way that the September event usually is not because we're going to get to see what they are doing to try to make the foldable experience feel like something that is worth opening your phone for.
David, do you want to have a camera crash out?
Yeah, do it. I can do two of them.
Felt me. Neil felt me trying to move along because...
No, absolutely not. We're going to sit in this for one minute.
Because there's no telephoto. Is that what you're talking about?
No, dude. Telephoto is one thing. I do think having the most expensive phone not have the best camera system is an interesting trade-off. Although I think at the high end, not everyone buys the Pro phone because it has the best cameras. That's just like an interlin causality. Like that has been true for the longest time so that people think it is necessarily true. Yeah. I think people spend money on all kinds of things that are not the best thing all the time.
So maybe you're buying the form factor. Maybe you're just buying the ability to unfold your phone. You know, like 2000s rappers would always talk about how they would drop their Bentley keys in the club.
Maybe this is just a thing that I grew up with, but this is like a thing. Like you're flashing your keys. Like maybe you just want to flash your phone. Great. The thing that I'm really curious about is we have spent all of this time verticalizing video. Like we have switched the aspect ratio of culture in like very serious ways. And Apple's like, it's landscape again. And I'm very curious if you have a point of view on this.
Yeah, I've heard a lot of takes about this from a lot of people. There have been people that I've talked to that have said, 'I really think the future is we're going to go back to horizontal.' I don't think we are. I think that when you unfold this phone, you're then going to turn it vertically and use it like an iPad mini. And that's going to be the way that people are still used to using their device. But I don't think that there's, I mean there's a world in which, you know, it's the content consumption device so you watch videos in a vertical format but people still turn their phones sideways to watch long form video. So if you're watching long or short form video with the screen open you're still turning your phone sideways. I don't think it's going to have a tangible effect on how...
Wait, I just want to be clear. Your bet is that iPhone Fold owners are going to open TikTok, unfold their device, and then turn it and then scroll in a giant iPad format.
Maybe the first two times they ever do it. Yeah. Yeah.
I'm very curious about that. When I say like we verticalize the culture, what I mean is the hottest money in Hollywood is like AI generated vertical micro dramas.
Yeah. Because that's the format.
Who knows? I'm just like I'm dying to find out how Apple contends with the TikTokification of media in this event when they're like, and then you unfold the phone and you have a giant landscape display.
Yeah. Where you can fit two vertical video feeds.
That's the move. And that's 100%.
They're not even super vertical, right? It's like because it's not the same candy bar style where you have two full screen displays on the side. It is very squared off. So, I don't know. I don't think people are going to start watching more long form content again. I don't think that's going to happen, but...
Okay. And then here's my other camera crash out question. Okay. I'm ready.
Again, I just I can't wait to figure it out.
You're going to open it and then you're going to have, it's going to be a big viewfinder. This is the other piece of the Gurman reporting.
You got the big viewfinder, which means you were taking landscape photos with a giant viewfinder.
That feels like we're going to get some really weird photos out of this thing, right? Like just framing wise, everyone's going to go from vertical default framing to landscape framing. That's one thing. But then they will have a giant viewfinder, which necessarily just changes how you would compose a photo.
My bet is that nobody is going to do that. I think that everyone's going to keep the phone closed while they take their photos. Personally, that's...
You get a weird like paperback book. I mean, these are the same people who are holding up their iPad at a concert, right? Like that's what this boils down to.
Yeah. The first moment that you get the person with the unfolded iPhone Fold at a concert, it's going to be like broadcasting behind them size screen. That's a meme. Like, we're just waiting for that to happen. The countdown is on as of today.
If you see it, you need to send us the photos immediately.
I already get annoyed when people are recording the concert and I can't see because their phone's in front of me. If it's the size of an iPad mini, I'm going to be really...
Well, your view will be better because the screen will be bigger. No, the only rational thing to do is to bring a full-size iPad and then stand in front of that person.
Like just a full arms race. Because doesn't the full-size iPad have the full camera array?
Uh, I think it has two. I don't think it has a telephoto camera. It's got LiDAR. Yeah, for some reason.
So you can map the concert and relive it in 3D in your Vision Pro.
There. That's your 3D scanned on a rail. You're like, I've got a full 3D avatar of you saved on my iPad for further use.
I don't need your Alicia Keys concert. I got my own. We got Alicia Keys at home.
Okay, those are my camera crash questions. Obviously, more will be revealed, but I'm dying to be revealed.
I mean, people are going to use this thing. It's going to get away from Apple in like the best way, right? Yeah, that's the nature of these things is people start using them in ways no one can predict. And everyone's camera the size of an iPad mini is just one of the most unpredictable changes to just how we have thought about making and taking photos in quite a while.
That's true. It's going to be fun. Very excited to find out September 9th. All right, we've got a lot more to talk about. We're going into the Meta settlement when we're back.
All right, we're back. So, big news this week in the world of Meta. Meta has had a lot of lawsuits flying its way. One of the most important was a lawsuit from attorneys general across the United States who had sued Meta, claiming that its platforms had been made addictive to teenagers and children in particular and that Meta had deceived people about the design choices they were making. Lauren Feiner and Addie Robertson were on the show earlier this week to get into the weeds on that. But I really want to talk about what this actually means for Facebook and Instagram as apps because there are kind of two halves of this settlement. One is that Meta has to pay some money. The other is that they have to make a bunch of changes to their apps. David, can you kind of break down for me what some of those changes are?
Yeah, there's quite a few of these changes and honestly I would love to see some of these as changes or at least options in my app as I am not a teen. Uh, one of them, 2-hour daily time limit that's cumulative across Facebook and Instagram, uh, which can be turned off with the parents' permission. Am I my own parent? I hope so. There's a night mode where you can default block the apps from midnight to 6:00 a.m. to help you sleep. School mode, so notifications can be muted from 8:00 a.m. to 3 p.m. By the way, a lot of these exclude DMs, which is an interesting sort of idea there. I mean, obviously DMs are extremely common in all of these apps, and arguably Instagram is a DMing app that has reels. There are screen time prompts every 15 minutes to let you know you've been on the app for another 15 minutes, which is helpful. Algorithmic feed control. Teens can now choose non-algorithmic feed as their default, and parents can adjust the default. I want all of these things.
I mean, this is really like a wish list of features for people who are like, 'These apps are too addictive.' I do want to step back for a second. So the states had gone very aggressive at Meta. My question is, you're looking at this list. Who won this trial, right? No, or not this trial, this settlement. Who comes out ahead?
It depends on who you think is getting what.
Like in a very real way. This is a list of regulations that if the states had passed laws to get, they probably would lose in court.
Because a lot of these implicate the First Amendment very directly and there's not some great body of precedent or academic research that would overcome the First Amendment problems with this. These are speech regulations. You can only speak on Facebook for two hours a day, says the state. Straightforwardly a speech regulation straightforwardly implicates the First Amendment. So, if you're the state regulator and you're like, 'Look at this big addictive platform that's hurting our teens. How do we get Mark Zuckerberg to do anything about it?' You've won.
Because they have agreed to the limitations that if they had fought you on, they might have won. And in fact, almost certainly would have won something, right? That this would have been a 10-year battle that would have gone to the Supreme Court and we would have had 50 episodes of the Vergecast about whether night mode implicates the First Amendment for teenagers. You could feel a lot of different ways about that. So, in that very narrow view, the states have won here because they've accomplished a series of regulations that might help teens be less addicted to this one platform in a way that if they had literally passed the law and Meta had taken them to court, they might not have gotten as cleanly and easily as they have gotten here.
I think a lot of people who spend time thinking about platform moderation and regulation are looking at this and saying, 'Oh, Meta just ran away with this. Right? Like Meta just hoodwinked these states because the money part of the judgment only pays out over 10 years and only 70% of it actually gets distributed over those 10 years. So that's like 10 billion dollars over 10 years. And the last 30% only gets distributed if TikTok and YouTube agree to the exact same restrictions, which is bananas. Because who? Meta has every incentive in the world to keep TikTok and YouTube from agreeing to this because they'll save money if TikTok doesn't agree to this. But they've also agreed to run an ad campaign agreeing to this. And they've given a huge sign-off on these are the gold standard teen safety regulations that all of the other platforms have to contend with.'
Whether or not you are worried about the First Amendment implications of the states imposing speech regulations on these platforms, this is a big swirl. If you're a parent, you are looking at this. If you have, like me, your parents are senior citizens who are addicted to WhatsApp, you're like, 'Oh, can I get this for WhatsApp?' That'd be great. There's a lot of this that on its face is good and smart. It's the stuff we all want. Like we can see that this will disrupt the engagement cycle. This will disrupt the business model built around making the platforms more addictive, making them full of rage bait, all that stuff. This will disrupt that. This will disrupt Meta's ad revenue for young people because they're going to be using the platform less. But hidden in here is a bunch of regulations that I don't know if the states would have gotten all on their own that Facebook is now going to impose on the other platforms. And I think that's pretty wild.
So I think from that perspective, Meta has kind of run away from this. They've like hoodwinked the states into age verification at scale across the internet. They've hoodwinked the states into going to TikTok and YouTube and whoever else and Snap and saying, 'Hey, if you don't voluntarily impose these restrictions that may or may not have speech problems associated with them, we'll bring a lawsuit and we'll say to the judge, well, Facebook settled this way and we'll tie you up in knots for it.' Or you could just do age verification whether or not you agree with it. So on that perspective, you know, all of the policy people that I pay attention to are kind of like, why, what, don't do that?
So I mean there's some real back and forth here.
I think it's very funny you mentioned it, but Meta is running these full-page ads in the New York Times and the Washington Post egging on TikTok and YouTube to adopt these exact same rules. And I'm curious, is there any reason either of these platforms would voluntarily agree to any of this? Right? It seems to me the answer is no, which is why Facebook did not. Right? Like they waited until they were sued by basically every state in the US. Yeah. I don't see a reason other than garnering goodwill, but it's not enough for the amount of money that they're making from all of these things.
And I'm not sure that anybody's going to buy that Meta is coming from like a righteous position here.
No, no, no, no. They were going to lose this. The weird thing about this is, you know, Adam Mosseri was on the stand a few days ago and it was brutal.
Because he's reacting to documents that literally say, 'Don't show this data to Adam. It will create litigation risk for us.'
And he would have to say things like, 'Well, I want my team to be honest with me, but sometimes they can't be because of this.' And he's like looking at the lawyer. Like they were stuck in a very complicated trap where their own research into what the harms actually were and how to combat them was being used against them to show that they knew there were harms and they weren't doing enough. And so, you know, Zuck's response to this was we should stop doing this research.
Every platform's response to this is we should stop doing this research.
Yeah. Because if you create the knowledge that something bad is happening, then you have created the liability that you didn't do enough to stop it. This is a very bad set of incentives. On top of that, you're getting, well, the way to stop it is to impose controls on how people communicate on your platform, which you as the private company can do. But maybe you think that the government should not be able to mandate. In this country, the government usually can't mandate that stuff. Other countries around the world very different. But in this country, the First Amendment usually prohibits the government from mandating that stuff. And you end up in this place where, oh, you can concede to the government all of these things. And you can neuter your competitors who get to do those things by saying, 'Oh, we're going to make it so every attorney general in the country who participated in this can now run around with a huge incentive, monetary billions of dollar incentive to punish Google unless they agree to these rules, to punish TikTok unless they agree to these rules.' And that is pretty, you know, it's like it's pretty sharky. It might be the right outcome. I don't know that answer because I am a parent and I dread the day my children get phones with social media on them and I'm looking at two-hour daily time limit and night mode and school mode and algorithmic feed control. I'm like, 'Oh, these are the things that I want.' Right. If you don't build them into the software, I can't parent enough unless I impose draconian surveillance on my children to make them stop doing these things.
Well, it's very interesting. Like these are all features that we all believe, we think that these will make the app less addictive, right? Like there's no autoplay on video, right? That you can choose a non-algorithmic feed. We all kind of just have this belief that these things will reduce usage. I think it's probably true, but this is probably the first time we're going to find out at scale if that is true, which is a very big but important if, right? Like if these things don't work then it's just the time limit. It's just we're cutting you off after two hours which is still a lot of time.
And the other question becomes, as part of this here it says that they have to do better age verification which I feel like we've been through 100 times and teens are very good at getting around the age verification things and Meta can say we're doing more intense age verification systems. We're using more of the data we have to actually track how old you are and all this stuff and your time on other apps etc. But people are still going to get around them.
Oh, but that's built into this. The age verification systems they have to use, Meta has a choice here where they can use an accredited third-party verification service and ostensibly the idea there is to create a market for these third-party verification services or they can build their own and there are some acceptable levels of tolerance in that rule and I think it's 10%. So they can have a 10% error rate and that is acceptable. And I think it's different in the different age ranges. So there's a little bit of tolerance built into this.
It's like the 17-year-olds, you expect more of them to be able to fake their ID.
Yeah. There's like some thing in here where it's like if we detect that you're an actual child, we need to be more tolerant of that. There's some amount of we understand the technology is not perfect and they will try to defeat it. So we will allow for some error rate. But what you're really getting, and this is true in all age verification, is you have to tell a big centralized service provider what your government name is. You have to identify yourself, even if you're an adult because there's no way to only do it for kids. And that, right, that obviously creates a huge burden on adults who are not subject to these regulations. And Meta just horsepowered it through, right? And like great, like we all agree that this is a good thing. This has been up and down to the Supreme Court and the arguments against it are super real which is making adults in this country submit government-mandated identification to the internet every time they want to do anything is actually a huge burden, is actually another restriction on your speech and your liberty. And here we are. Maybe the states couldn't have gotten it if they just passed this law because of that burden, but they got it because Meta caved and there was an incentive to get YouTube and TikTok to do it, too. And there's some piece of this like again when you say who won and who lost. Boy, does that cut in 50 different directions.
Yeah. And it gets messier like the state's job is not done, right? Like because there are parts of this where the state still have to figure out where the line is. There's this very weird thing going on where Meta agreed to limit the use of what is it? Extreme makeup filters, but that it's not defined what that is. So there's some system where the government is going to determine what that is. Cosmetic procedure filter shall mean any digital filter or augmented reality effect that distorts, sculpts, redefines, or idealizes a user's face in a way that cannot be achieved without cosmetic surgery or extreme makeup techniques.
Now, for the avoidance of doubt, it goes on to say this does not include one, fantasy character effects. And I just want to be clear with specific definitions, filters that change a user's facial structure for the purpose of turning the user into a non-human or fictional character, e.g. an elf or an animal, e.g. a dragon or puppy.
Okay, just putting it out there. A filter that turns you into a puppy.
So Memoji still work. Memoji puppy is fine.
Carved out in a, makeup smoothing effects. Filters that smooth skin or alter appearance in ways that can be achieved by ordinary makeup techniques without altering the appearance of underlying physical facial structures or meaningfully altering skin tone. And then lastly, parody and exaggeration effects. So to help operationalize this definition, the states will provide Meta no illustrative examples and guidance of AR effects that are cosmetic procedure filters and AR effects that are not cosmetic procedure filters in a letter to be sent within two months. So the states, someone in some government office is going to have to sit around doing puppy filters that are okay and puppy filters that are not okay, put that all into a letter and send it off to Meta.
