I have some questions for Mark Zuckerberg

Support for the show comes from CrowdStrike. It's not a huge stretch to say that AI is the next major computing platform. Every platform shift changes the way we build software, and it changes security too. CrowdStrike is defining cybersecurity in the AI era with AI detection and response, or AIDR, a solution for seeing, monitoring, and securing AI across the enterprise. Learn more about how leading companies are turning to CrowdStrike to secure AI and secure their business at crowdstrike.com slash decoder. In the AI era, you have to stop cyberattacks where they start. So security has to be fused into the infrastructure itself. That's why the network that connects everything is also your best defense. Cisco, the critical infrastructure for the AI era. Support for the show comes from Superhuman. The conversation around burnout is getting louder these days. The thing is, it's not that we just don't have enough hours in the day. It's that a lot of us have our time eaten by tedious, busy work. That's where Superhuman might be able to help. This isn't about clearing tasks off your to-do list faster. It's about finding a tool that lets you take on more ambitious work. Superhuman's suite of proactive AI products are built for people who take their work seriously and expect the same from their tools. Grammarly by Superhuman is a writing agent that supports you from blank page to hitting send. Superhuman Mail is like an email inbox that automatically filters the noise and lets you focus on what matters most. And Superhuman Go is an AI assistant that can work anywhere you do. No need to get it up to speed because it's already there. You can try many of Superhuman's AI features for free, so you can see for yourself if it'll give you the kind of support you're looking for. Start streamlining your workflows with Superhuman. Try today at superhuman.com. Hi, I'm Nilay Patel, editor-in-chief of The Verge, and this is Decoder, my show about big ideas and other problems. It's a big week for Meta. The company has its big Connect conference, and the new Muse AI agent appears to be an early hit. I've been using it. It is surprisingly good. CEO Mark Zuckerberg has obviously been out there taking a victory lap, doing a lot of interviews about all this, so of course we asked for him to appear on Decoder. Meta declined my interview request for whatever reason. But luckily, Mark just did a big interview with our friend and former colleague Alex Heath on the Sources podcast. So I figure we could just highlight some moments from that conversation, and I could point out some obvious gaps in what Mark is saying about the company right now. And hopefully the next set of people who get to talk to him can ask him some follow-up questions. So here's some of the things I would love to ask Mark Zuckerberg if he would join me here on Decoder. One thing before we get going, don't forget to subscribe to @DecoderPod on YouTube for new episodes every Monday and Thursday. Let's start with Mark's comments on Meta glasses. So we actually feel quite good about the product. I think it, if, I mean, phones don't have a light. I mean, people go around, like, recording people all the time. So I think it's very important to note that the entire tech industry really wants passive surveillance devices to fit into the same category as phones, for you to understand everything from flock cameras to Ring doorbells to Meta glasses the same way that you might understand a smartphone. Now, there are some very obvious differences here that I think are worth really pushing on in these conversations. For example, if you want to record someone with your phone, you have to take it out of your pocket, open the camera app, and literally hold it in their face. This is a physical action that is meaningfully different than having a pair of glasses on your face with a light and just passively recording someone. That physical action means that you have to announce your intent to do something with your body, and other people can react to it. Obviously, you don't do that with Meta glasses. You just start recording, and people can notice there's a light or not. But the big physical announcement that you're about to begin recording is not present. So I think one of the reasons the entire tech industry is so eager to fit these passive surveillance devices into the same kind of mental framework as smartphones is so that no one has to question whether there should be a new social etiquette for these devices, and instead you're like, Oh, it's just a smartphone. I already understand how to use that. But importantly, that also brushes over the fact that we did have to create a massive new set of social rules and norms around smartphone recording. We did live through the entire Black Lives Matter movement because smartphone recording allowed people to record the police, and we as a society didn't quite know what to do with that information, and maybe still don't know what to do with that information. I would just push on this over and over and over again. Do you want there to be constant recordings of everyone? Do we think that smartphones allowing everyone to record everything all the time was good? Do we think being able to livestream everything all the time was good? If you could reboot that in a new form factor that allowed for easier and more passive recording, what norms would you change? And there's already examples of this inside of Meta, where they know the Meta glasses are being used to used to harass retail workers and women, and they have started downranking those videos on Instagram to provide less incentive to make them. So I think a version of this question you could ask very directly is, if you know some of the content that is being made with these glasses is so bad that you've downranked it on your own distribution platform, what is so good about these glasses that make it worth having these debates over and over again? All right, let's get back to Mark's comments on Meta glasses. When we launched the glasses a few years back, we communicated pretty clearly about the steps that we put into it. But like you said, there's now, like, many, many, many millions of people who have gotten the glasses more than, you know, had them when we just started launching the product. So I think some of that communication that we did at the beginning, a lot of people either, you know, maybe forgot about or, you know, they didn't see at the beginning or they just weren't paying attention to it because the glasses weren't a big thing. Now that I think we've achieved a level of mainstream, well, at a minimum, I think we need to kind of go and make sure that we communicate about what we're doing and that, yes, we think that this is important and in fact so important that we designed it into the product from the very first version that we shipped multiple years ago. So I would ask a very simple question here: what is important? Is it the privacy of the user? Is it the privacy of the person who might be being recorded? And what is it that you thought you were communicating clearly at the beginning that got lost? What I think Mark is trying to say is at the very beginning, they announced that the glasses would have a light, that the light would be on when you're recording, that they thought it was obvious that there was a camera in the temple. And over time, as more people got the glasses and more people experienced the glasses on other people's faces, it was not clear what that light was or what it meant. And so now Meta, Mark, thinks they have to do more communication about specifically the light existing. And my response to all of that is it doesn't matter what Meta says. This is the essential truth of every piece of mainstream consumer technology. The users define the product. If you sell millions of gadgets to people, they will use them in ways you never expected. And it's those emergent behaviors that often define the next version of the product. Probably one of the most famous examples of this is Twitter. Twitter didn't invent the hashtag or the @ reply or any of a million features that we associate with Twitter. Users on Twitter invented those features, and Twitter adopted them because the users had figured out a way to use their service in ways that they had never even contemplated. This is broadly true of everything. It's true of computers. It's true of smartphones. It's true of social platforms. And it is true of Meta glasses. The ways people are using them have nothing to do with Meta's intentions or its communication about recording lights at the beginning or now. So I think the natural follow-up question we should all be asking is how much control does Meta have over bad actors who are using these glasses to harass people? Because just telling more people that the light indicates recording isn't going to stop it. Are they going to do more things like downranking videos of people harassing retail workers? And do they think the other platforms will follow their lead? And does Meta think that creating negative incentives for how people use a camera with content moderation will actually solve the problem? Is there anything about the history of content moderation that suggests that that's true? Given Mark's own history with content moderation. Support for the show comes from Granola. Have you ever left a meeting thinking, That was productive, and then ten minutes later, when it's time to take next steps, you're at a loss? Or worse, you've been heads down the entire meeting furiously typing notes instead of actually listening, and now your brain is fried. Meetings are a mess. Granola fixes that. Granola is an AI-powered notepad that makes your meeting notes useful. It captures what happens in your meetings and turns it into clean, structured notes with decisions and action items pulled out and made easy to find. And the best part? Granola works through your device's audio, which means it integrates seamlessly into the video conferencing tools you already use. No setup, no awkward bots. It's just your normal meeting with superpowers. You get to actually listen instead of frantically typing every word, and still walk away knowing exactly what was decided, who's doing what, and what comes next. Once you try it on your first meeting, it's hard to go without. You could try it totally free. Just head to granola.ai slash decoder. That's granola.ai slash decoder to get your time back. Go to granola.ai slash decoder and try it on your next call. You're listening to Deepgram Flux TTS. Different voices, same model, all ready to speak. Flux TTS is a streaming text-to-speech model built for voice agents. Flux reads the room. It holds context across the conversation, turn after turn. With consistent tone, interruption handling, and plenty of personality, so your agents keep it flowing and customers can just keep talking. Try Deepgram Flux TTS free now until September 12th. Visit deepgram.com slash keep talking. Terms apply. Support for the show comes from CrowdStrike. AI is already fundamentally changing how companies build products and run their businesses. If you listen to the show, you probably already know this. But while AI drives innovation, every new model and agent also creates new cybersecurity vulnerabilities. That's why CrowdStrike built AI Detection and Response, or AIDR. AIDR helps organizations discover where AI is being used, monitors AI activity at runtime, and helps teams detect, investigate, and respond to threats involving AI systems. As companies adopt AI, security has to be built in from the start. More than 70% of the Fortune 100 trust CrowdStrike to protect their businesses. Now CrowdStrike is helping secure the AI era. You can learn more about AI Detection and Response at crowdstrike.com slash decoder. Next up, Mark talks about teen safety and Meta's recent multibillion-dollar state settlement. I think it would have been possible to have kind of explained along the way how seriously we took those issues, and we— Just didn't because we thought, okay, well, people are showing that by using the product so much, they still like the product. Let me stop you here. The claim in many of the social media lawsuits is that the product is addictive. The reason people keep comparing social media to big tobacco is that smoking is fun and people enjoyed it, and it was killing them. And so the usage rates had nothing to do with the harm that might have been occurring. So from the jump here, I would have wanted to push on whether simply measuring usage rates actually lets you measure whether or not something is doing good for the world or good for an individual user. Where we've gotten with the social media products, which is both people liking it and understanding how seriously we take all of those issues. So that's what I aspire to. Is there a through line from that to the recent settlement on all the youth safety stuff, and it's a thing a lot of people are talking about? Is there any connection from what you just said to that? And I would just be curious to hear you reflect on that and what you've learned from this process, or if— Yeah, no, I think it's another good example like this. I mean, we've taken a lot of those safety issues seriously for a while, and we've been working on this teen account work with Instagram for a long time, and I think have done some leading work there. So it is true that Meta has done some of the leading work on designing features that would keep kids safer on its platforms. They've done some of the leading research in the space on what is actually happening to people on our platforms. And one of the dynamics that is particularly galling throughout all of this is that Meta gets punished for it. So they publish their safety numbers, they publish their research, they publish the amount of bad things that are happening. They do transparency reports. And even if you don't think that that's enough, a real thing that continually happens to Meta is they get punished for revealing that bad things happen on their platforms. And companies like Apple, which don't know what bad things are happening on an encrypted service like iCloud, don't publish those numbers and don't get punished for it. What I would have asked Mark here is, how do you solve that dynamic specifically? Because in these lawsuits, it's been revealed in internal email. Emails that Mark himself has questioned the value of doing the research and building the features because Meta gets punished for it. So there's a really strong negative feedback loop here, where Meta tries hard and gets punished for it because they're transparent about it, that in fact you would want a government regulator to solve. If all the social media companies were required to be transparent about various metrics around teen safety, then Meta would be judged in comparison to its peers, not for the information existing at all. And that doesn't appear in this settlement. Instead, what we have is a bunch of behavior controls, product restrictions, and Meta seems very keen on making sure those product restrictions connect to TikTok and YouTube, but not on whether we should be able to measure the harms themselves. So that's the first thing that I think is worth pushing on. Does this whole industry need to be required to be more transparent about the things happening to children on their platforms? Because if you're going to make rules and regulations, you should probably start from a common set of information. Now, if you basically said, you know, limit usage to an hour a day for teens, you know, everyone I think would be okay with that, except if, like, you have to unilaterally do it. Then you're saying, okay, like, if people don't use Instagram for more than an hour a day, but then their usage goes to TikTok, have we really, like, helped anyone? And we've just, like, hurt ourselves to not help anyone? What Mark is describing here is a prisoner's dilemma, where if they take action to make teens safer and TikTok doesn't, the teens will just go use TikTok and no one will be safe. This idea, of course, is going to come up again when we talk about AI safety. But let's just talk about social platforms for a minute. If you believe that there's a market for safety, right, if your initial premise was people are using the platforms so they must be happy with it, well, then you should make the platforms safer for kids. And parents might say to their children, you're allowed to use Instagram and not TikTok because Instagram has these controls. It is very telling to me that Mark doesn't think there's a market for safety, that instead he thinks that everyone must be forced to unilaterally disarm by the government in a regulatory framework that solves his problem. This is prisoner's dilemma. So my big question here is, why is this a problem for the government to solve in social media in a way that might implicate the First Amendment and the speech rights of teenagers, and not a problem for the government to solve in AI safety, where the frontier labs keep talking about killing everyone? By the way, when I say this connects directly to AI safety, what I mean is that Mark is arguing the opposite side of the prisoner's dilemma when it comes to AI safety. After all the calls for slowdowns from the various frontier labs, Mark posted two threads on X that he thought the market would take care of it. Here, I'll just read it to you. My view is that trust and alignment are quickly becoming the most important capabilities that will differentiate agents and models. Any lab that doesn't focus on alignment will fall behind. Labs face significant liability if their models cause harm, so they have a strong incentive to prevent this as well. Meta delayed shipping Muse for several months to focus on safety and security. We didn't call for For WeWood, we just did it as part of our day-to-day work because it was clearly the right thing for people and for us. I'm proud of the security foundations we've built. So, straightforwardly, I think the question is, why is the go-it-alone approach appropriate when it comes to AI safety, but getting the government to pull the other social media companies into a framework for teen safety appropriate in that market? The other question, I'm just going to steal from my friend Casey Newton, who pointed out that in the most recent teen safety case, Meta estimated its own liability at a trillion dollars in damages, which does not appear to be enough liability for Meta to take unilateral action to prevent teens from being harmed on its platforms. If a trillion dollars in potential liability isn't enough, how much is enough for Meta to take AI safety seriously? And is there any number in a liability scheme that would stop everyone all at once? Maybe that's an urgent email from your CEO, or maybe it's a deepfake targeting your business. Doppel is the AI-native social engineering defense platform fighting back against impersonation and manipulation. As attackers use AI to power sophisticated strikes, Doppel defends your business across three powerful layers: digital risk protection dismantles attacker infrastructure, human risk management builds team resilience through simulation and training, and email security provides agentic email protection beyond the inbox. Together, this unified stack executes automated takedowns and learns from every decision to protect people, brands, and trust. Doppel, outpacing what's next in social engineering. Learn more at doppel.com. That's D-O-P-P-E-L dot com. Support for the show comes from Pipedrive. Sales teams spend up to 50% of their time on admin work rather than selling, building relationships, and closing deals. They could really use a tool that covers all the admin stuff. That way, sales teams could actually focus on the work they were hired to do. That's where Pipedrive comes in. It's an intelligent, AI-powered sales CRM loved by growing sales teams. Their new meeting intelligence features, like an AI note taker, are built right into the CRM. Pipedrive automatically pulls deal history, email records, and previous conversations. You can walk into a call already briefed. During meetings, Pipedrive's AI note taker records and takes notes. Then it turns those notes into accurate, auto-drafted CRM updates ready for you to review and approve. The end result? Less admin, more selling. Switch to a CRM built by salespeople for salespeople and join the over 100,000 companies already using Pipedrive. Our link gets you an exclusive 30 days free instead of the usual 14-day trial. No credit card or payment needed. Just head to pipedrive.com/decoder to get started. That's pipedrive.com/decoder, and you could be up and running in minutes. You know that feeling. Too many updates, too many meetings, too many things falling through the cracks. Monday.com was built for that gap. The AI work platform designed from the ground up for people and agents to work side by side and deliver more together. Drafting the brief, flagging risks, running a competitive scan, your agent can do it all. And the best part? Anyone can build one. Even you. Try it. Create your first Monday agent today at monday.com. Mark then got into how he plans to run Muse as a business. It, I think, is just going to be able to, like, make people money and save people money. The reason for why we're doing this is we basically are confident in standing behind the fact that we think that this is going to, effectively, for people who are going to use it for running a small business or making money or transactions or commerce in some way, we actually just think it's going to make so much money for people that the business model over time that we expect is to effectively just take a very small cut of whatever the transaction is. Take rate business. Small. Yeah. Yeah, and not even necessarily the person paying for it. If you're a Ben Thompson fan or you've just listened to enough Decoder, you know that what Zuck is describing here is aggregation theory. He's going to have all the consumer demand, and it'll be so big and so lucrative that businesses who want customers will pay Meta a small fee to be on the other side of transactions that happen in Muse. Being one of the big aggregators is a very lucrative thing. All the big tech giants aggregate demand in one way or another. The problem is the businesses on the other side of the aggregators hate them. And Mark knows this better than most because he's on the other side of one of the biggest aggregators of all, the Apple App Store. If you want to reach customers on iPhones, you have to go through the Apple App Store and importantly agree to a bunch of privacy rules and pay 30% fees on most of the transactions that happen on an iPhone. This hits Meta in two ways. One, Apple nearly killed their ad business with something called App Tracking Transparency when they issued a rules change that really favored Apple's ad businesses over Meta's. And two, if you want to boost a post on Instagram from your phone, well, you basically have to pay Apple 30%, a fee that Meta breaks out. Mark hates this so much that he has explicitly talked about building new platforms to get away from Apple for years now. One of the reasons he chased the metaverse was to build the platform that came after the iPhone and escape Apple. One of the reasons he's chasing Meta glasses is to build the platform that comes after the iPhone and get away from Apple. And I am dead certain one of the reasons he's invested so heavily into AI and now Muse is to build the platform that comes after the iPhone and get away from Apple. So the question I would ask Mark here is very simple. How will you make it so all of the businesses downstream of your huge aggregator AI agent feel good about this arrangement? Because when you've been in this arrangement, you have felt very, very bad. Mark also spent some time talking about Meta's data center investments. So if you look at the data centers, what we've actually found is that when a company like Meta goes into a community and makes a commitment that we're going to invest there for decades, right, which is really what we're doing when we're building up a data center, we're able to make it so that it's very good for the community. I mean, the kind of the tax revenue that they bring from that. I mean, in Louisiana, we have this example where, like, the tax revenue funded these $50,000 bonuses for teachers in the community. Okay, so this data center in Louisiana, Hyperion, Meta is obviously very proud of it, and Mark is talking about the positive impacts it might have for that community. But the actual process of landing that data center deal was basically a bunch of secrecy and backroom deals and huge asks for property tax exemptions from the state government that residents really didn't know about. There's a huge New York Times piece about this project that really goes into detail about the kinds of backroom negotiations that Meta secured in order to go fast and build this data center. So I would simply ask if companies like Meta should lay out these supposed benefits in a much more public and transparent way and let people actually decide for themselves. Of course, you have to design it in a way that can be helpful to people in all these ways, which is also part of the reason why, like, for people who are skeptical or have so much doom about the whole thing, I mean, one of the things that I basically think is that if we don't end up building it in a way that's positive, it just— So this is another market argument from Zuck. He's saying if people don't like the AI that's being built or find it helpful, it simply won't be allowed to exist. The market will kill the products. You can feel a lot of different ways about that, especially if you've experienced AI being force-fed to you in your apps and services right now. But I would connect that to social media, where a lot of people felt that social media was doing harm and the industry raced ahead anyway until the government required some regulations. So I'm once again very curious about Mark's different postures to government regulation, market forces when it comes to social media, where he thinks there should be government regulation, and AI, where he thinks the market should be able to run however it wants. Secondly, and I think most importantly, I spent a lot of time on the show asking various AI executives if consumer AI products are good enough to overcome doomerism and skepticism, because right now they're not. And even the way Mark describes Muse is being helpful for people running businesses. or helping people save money or doing shopping. They're fundamentally enterprise applications. They're not big consumer applications that make people overcome their skepticism. For example, Instagram is a beloved consumer application. It just entertains you, and that means all of the downstream harms of Instagram fade away because people like being entertained. It is not clear to me whether people are going to love letting Muse run their small businesses so much that they will accept data centers in their community. And it's not clear to me what consumer AI products will ever exist such that people overcome their skepticism of the enormous cost of data center buildouts. And so I would want to know if Mark thinks Muse is that great consumer product or if there's something yet to come. So those are some of the questions I would have asked Mark Zuckerberg if he had agreed to come on Decoder. But there's one more clip, and I do want to give the man props. We're also just making it so that you can get a very large amount of usage for free. I think we're offering something, I think to start, it's like 100 million tokens a week for free, and you get this virtual machine. So it's like a lot of kind of computer. That's a good meme. It's a lot of computer. It's a lot of computer. Yeah, it is a good meme. It's a lot of computer. As I mentioned at the top, I've invited Mark to be on Decoder, but he declined the offer. That offer still stands, though, and I would love the chance to talk about all these things directly rather than through a YouTube reaction video. I'm also dying to know what you think. Anything I got wrong here, anything Alex missed in his questions, or anything else you wish Mark would get asked. Drop your thoughts in the YouTube comments, and if you like this piece and want more of it, please subscribe to the @DecoderPod YouTube channel and share this episode with a friend. And if you're like me and you're an audio listener of the show, reminder that you can get ad-free episodes by subscribing to TheVerge.com. So we actually feel quite good about the product. I think if, I mean, phones don't have a light. God. Your heart can tell you a lot about your health. Apple Watch Series 12 measures your heart rate every five seconds with the most accurate heart rate sensing in a wearable. So your Vitals app, now with heart rate variability, can tell you when something is off. And your Readiness score can let you know when to rest and when to push. Hear the story in every heartbeat with Apple Watch Series 12. The features described are for wellness purposes only and not for medical use. iPhone 11 or later required, based on Apple-conducted study of heart rate accuracy, August 2026. Visit apple.com/apple-watch-series-12. Booking.com is the easiest way from a day surrounded by noise to a stay surrounded by nature. That’s nice. Go on, book it. It’s easy. Booking.com. Booking.yeah. For a limited time, you can get a Big Mac meal for just eight dollars. That’s a burger, fries, and a drink. They don’t call it an extra value meal for nothing. Get a Big Mac meal only at McDonald’s. Price and participation may vary. Promotion pricing may be lower than meal pricing.