Interface Design for Custom AI Agents

We're into the thick of the season now and so's Coral's reward shaker with even bigger rewards and offers. The half and half Scarfers shaking decisively for once. They backed the super boost. Next to them the prawn sandwich brigade. Oh they,'ve shaken up a free bet after that free lunch. Play Coral's free reward shaker to win bigger guaranteed daily rewards and offers this season. Coral, we're here for it. 18 plus UK max one reward or offer per player per day. 8 a.m. to eight pm. Reward restrictions and ts and c's applied. Gamblerware.org. Getting instant insights is amazing. But if there are too many data points, it can be hard to see what works. So I'll ask my AI Aistssant for recommendations. And with PDF Spaces in Acrobat Studio, it's easy to remix documents and transform insights into standout content so you can go from idea to creation in record time, all within an AI-powered workflow. Do that with Acrobat. Learn more and try it out on Adobe. com. Support for the show comes from MongoDB. If you're a developers stuck fixing bottlenecks? Instead of building the next big thing, then you need MongoDB. MongoDB is the flexible, unified platform that gets out of your way. It's acid compliant, enterprise-ready, and built to ship AI apps fast. It's trusted by so many of the Fortune 500 for a reason. Ask any developer. It's a great freaking database. Start building at mongoDb.com slash bu ild. Hello and welcome to Decoder. I'm Neilai Patel, editor-in-chief of the Verge, and Decoder is my show about big ideas and other problems. Today I'm talking with Shashir Mahotra, the CEO of Superhuman. That's the company formerly known as Grammarly, which is still its flagship product. Shashir also used to be the chief product officer at YouTube, and he's on the board of directors at Spotify. He's a fascinating guy, and we actually scheduled this interview a month or so ago, thinking we'd talk broadly about AI and what it's doing to software, platforms, and creativity. Then things took a turn. There's a feature in Grammarly called Expert Review, which allows people to get AI writing suggestions from quote-unquote experts, and reporters at The Verge and other outlets discovered that those experts included us. Included me. No one had ever asked us permission to use our names in this way, and a lot of reporters and other authors were outraged by this. The talented investigative journalist Julie Angwin was so upset she filed a class action lawsuit. Superhuman responded to all this controversy by first offering up an email-based opt-out and then killing the feature entirely. Shashir apologized, and you'll hear him apologize again in this conversation. We'll put links to all this backstory in the show notes if you really want to dive in. Throughout all of this, the decoder team and I kept wondering if Shashir was still going to show up and do an interview, because my questions about decision making and AI and platforms suddenly seemed a lot harder than before. To his credit, he showed up, and he stuck it out for the entire conversation, which got tense at times. It's clear that Shashir and I disagree about how extractive AI feels for people and the value that these platforms can actually provide. I'm not going to stretch this out any longer. I'm dying for you to listen to this and I'm dying for your feedback. We really do read all the emails. Here' Shsashir Mahotra, CEO of Superhume. Here we go. Shashir Mahotra, you're the CEO of Superhuman. Welcome to Decoder. Thanks for having me. Uh I'm happy you're here. I'm a little surprised you're here. I think you you know what some of the questions are gonna be, but but but I'm really happy you made it. I have a lot of questions about AI, how people feel about AI, and then a feature you launched in Grammarly, which is one of your products that made people feel a lot of feelings about AI. So we're gonna get into it. Let's start at the start. Superhuman owns Grammarly, you own Coda, you own a bunch of companies. Just quickly describe the structure of Superhuman and all your products. Oh yeah, sure. Uh so Superhuman is the AI native productivity suite. We bring AI to wherever people work. Late last year we changed the name of our corporate entity from Grammarly to superhuman. Did that the scope of what we what we do broaden quite a bit. And so in addition to Grammarly, which is everyone's favorite writing assistant, we now have a document space called Coda, a very popular email client uh called Mail, and we launched a new product called Superhuman Go. Uh Go is the platform that brings you a network of proactive and personal AI assistance directly to wherever you work. So for people familiar with Grammarly, you can think about Go as taking that core idea and allowing anybody to write agents that work just like Grammarly does. Your sales agent, your support agent, so on can all help work with you write where you work. And the core idea is that most AI tools require a big change in behavior. We bring AI where to where you work across our products. We see about a million different apps and agents every day. Uh we seamlessly blend AI right into your uh experience. So you don't have to think about AI. So that's what we've doing with Grammarly for years. And now we are opening that up. So anyone can build on that with Supreme and Go. So you and I hung out a few weeks ago and one of the things we talked about was the fact that Grammarly for most people is expressed as a keyboard, right? It it shows up on your phone and your documents. You spend a lot of time figuring out how to make sure you work with things like Google Docs. All of those products are integrating AI in exactly the same way as you're describing. Are they going to put AI right next to the insertion point, right next to your cursor? What's the big difference Actually, first off, I think that uh I think very few of them actually are doing that particularly well. Um a handful do. But as I mentioned, we we see a million unique apps a day. So the way to think about Grammarly is it's it's your assistant that lives everywhere. So you might be in a web app. So it could be uh Gmail, it could be Google Docs, it could be Coda, it could be Notion. Um, you could be in a desktop app, that could be Apple Notes, that could be Slack, that could be you know, whatever, whatever app you're using, could be every mobile application. We have for every one of those applications, we figured out the right way to observe what you're doing, uh annotate it in a way that is unobtrusive to you and to the application, and to make changes on your behalf. And doing that everywhere is the proposition. So as you jump from tool to tool, yeah, there's different types of AI in each one. Most of them actually don't have that. Like I said, we've we see a million unique surfaces a day. And the ones that do don't feel like one integrated experience. That's why we have about 40 million daily active users, and that's what they that that's what they use us for. It feels like the the promise there is by looking at all the places you work, your tool will be more intelligent than disparate tools you might encounter in all those places. Aaron Powell Yeah. I mean becoming more intelligent is certainly part of it. I mean I think I think for uh many people it's just that one familiar experience that really feels like that virtual human working right next to you. So is it consistency of experience or is it better and more useful results? I I mean it's both. So I mean uh and if we think about Grammarly, I think Grammarly is both the it's uh ever-present is very important and very high quality grammar results. As we split the product into parts and we said we're going to take the platform layer of Grammarly and we're going to turn it into a platform. That's what we call Go. That's about allowing other people to create agents and experiences that provide a high you know high quality experience that we can make ubiquitous for them. All right. I wanted to understand what you what you think that the sell of the tools is. I think it's very important for my next set of questions. The other thing that I really want to ask, which is a question I ask everybody, but I think the stakes are a little bit higher here. Uh it's about decisions. How do you make decisions? What's your framework? Yeah, I mean I think uh we have a lot of different thoughts on on how to make good decisions. I wrote a piece a long time ago called Eigenquestions, which is about framing uh not only the right solution, but how do you frame the right question? Um, in terms of rituals, we use the the most canonical one is something we do called Dorian Pulse, which is a way to solicit um uh feedback and opinions so that you get uh get rid of groupthink in the decision making process. Uh but those are probably the two that get mentioned the most if you were to ask uh ask teams here at Grammarly or previously at Coda or before that, you know, when I worked at YouTube or Google or so on. All right, you can see where this is going. Let's put this into practice. You launched a feature in Grammarly called Expert Review that generated suggestions on how to improve text. It synthesized advice from experts. You used my name among many other names, Casey Newton, Julie Angwin, down the line. Bell hooks was in there, which is hilarious in its way. You did not have our permission to use our names to do this. You had little check marks next to our name that indicated that it was somehow official people did not like this. I did not like this. And you remove the feature. Tell me about the decision to launch this feature with names you didn't have permission for, and the decision to unlaunch the feature. Yeah. So let's uh uh I expected we talk a bit about this. So let's uh uh lots of lots of different thoughts on it. You know, first off I'd say um you know I understand and respect how challenging a world it is for experts and idea generators these days. You know, I've made a long career out of being a partner to folks like you, to folks uh like the ones you mentioned, it deeply pained me to to feel that we underdelivered for them. And I really like to apologize for that. That was not our intention. Um on the specific feature you're talking about, um, you know, we I'm sure we'll talk more about it, but just to give the high level view, the my view of it is the feature was not a good feature. It wasn't good for experts, it wasn't good for users, it was a fairly buried feature, it had very little usage. You mentioned it last week and talked about it. Um took all it took months for anybody to even uh sort of find it. Um all that doesn't really matter. We can do much, much better. I believe we can and we will do better. Um, we decided to kill it pretty quickly. You know, notably we decided to kill it while there was some feedback well before there was a lawsuit and so on. It was just not a good feature. It was misaligned to our strategy. That was it wasn't to go after it. And we have a much better view on how we think experts should participate in our platform and I'm I'm a lot more excited about that. How many people work at Superhuman? About fifteen hundred. So fifteen hundred people, how many people decided to launch this feature ? Oh it's a small team. It's probably a PM and a couple of engineers. Okay. Inside your decision making process where you you you described a way of making sure you solicited the right feedback and then have groupthink, it never came up that using people's names without permission would make them mad? You know, maybe we should step back and talk about what inspired this team and where what they were trying to do and what fell short. Um let's start with what they were trying to do. So you know they're they're they're they were heavily influenced by both, by what what we view users to want and what we want experts to want. So start with users. You know, we think of grammarly. A lot of people describe us as uh they talk about us as the last mile of AI. They talk about us like it feels like having your grammar teacher right next to you everywhere you work. And so many of our users will say things like, what would it feel like if instead of your grammar teacher, it was all the rest of the people in my life could be with me as well. I want my head of sales to sit next to me and tell me I'm about to recommend the wrong product. I want my support person to sit next to me and say, I'm about to email this person and you should know they had a big support issue last week and, you should acknowledge that before you talk to them. So that's the core ethos of what we're building is taking Grammarly, expanding it so that many of these other experiences come uh come along with you. Um for some of those people, the people that want feedback from are the people they admire. It's the experts in the world. It's the people that they're trying to look up to and trying to model. Um and they they try to do that today with LLMs. They go to ChatGPT and Claude and they say, what would uh Nile think about my my uh my writing uh and so on. That was the inspiration for where what the user was trying to do. I think on the other side was what the experts are trying to do. And I think as we formed our strategy here, turning Grammarly into platform, actually the first people I called when think about this, were a set of experts. I talked to some prominent YouTubers. I talked to a really prominent book author. And they all told me the same thing. It's a really hard world for experts out there right now. It's really hard to drive connection. If you're a book author, you know, your path to getting people uh to getting to your fans is you keep publishing more and more books. And they all heard what we were doing and said, boy, it'd be really amazing to develop an ongoing connection with my fans. You know, what happens when they put my book down? Can I still be with them and help them along the way? And you know, it feels like the world shifted against them. You know, AI overviews are to stealing a bunch of their traffic and so on. And boy, this seems like a much better way uh to go after it. So that was the the inspiration behind it, the the the team you know the feature didn't deliver. Um and I think it didn't deliver uh on either side of it really. We ended up with an experience that um was pretty suboptimal for the user and obviously suboptimal to the expert. Um, I think the reason, actually the fundamental reason is something you said last week that it's really hard to distill what you would do as an editor based off the outcome of your published work. I just think it's like really hard for AI to do that. And we need your engagement for that to be a good feature. So I think they I think they launched something that wasn't wasn't particularly good. You know I think doing that and learning from it is part of the process. But that that's what they thought they were doing. Sure. How much do you think you should pay me to use my n ame? So I think I think it's really important to think about attribution and think about impersonation and so on. I think that the you know, as an expert, you have a trade you make on on the internet. And I think the uh the idea that when you put content out there, myself included, you hope people use it. You you want to refer to other people's content. You want to uh you want people to link to you. You really, really hope they attribute you when they do. So I think the idea of when somebody uses your content, should they attribute you, of course. And to to attribute you, you have to use your name. I think there's a different line w which is should people be able to impersonate you? And I think that is a uh that is a you know very different standard. And you know, we we saw the the lawsuit, you know, respectfully, we we believed the the claims are without merit. I think the idea that the feature was impersonation is quite a big stretch. not only by this person, inspired by a specific work from the specific person with a clear attributed link to get back to them. Far from uh far from that test. Should you if your work is used, should you be attributed? Yes, I think you should. I I I think that that that would be the nice contract. It doesn't always happen. I mean that I think there there's many products that will use your work and not attribute. We we thought it was very important to attribute. Um but I think that would be the view. Let me flip around the other way. I let me ask you that question again. If you use my likeness, how much should you have to pay me ? or any uh product at all, you should your work, they should attribute you. Um and they should link back to you. And I think that's a that's a human contract we have of how the internet is supposed to work. And I think it's a really important one is that we should we should make sure and I think it should be the standard you're looking for from LLMs too. I think it's a very different question you're asking here, which I think is a more important one. I'm not really here to defend this feature. I don't I don't think it's a good feature. I don't really want to I I'm not trying to be close to this line. I I think our main goal is to build a platform a lot like YouTube. You should choose to be on our platform. You should be able to choose and build an experience you trust and you should choose your business model. And when you choose your business model, you should get paid for your contributions to it. That's the model we're working on. That's really what where I want to be. Aaron Powell I I hear that you're saying you're not here to defend the feature. I just want to put you in the chronology for one second. The feature was launched. It is true, it took a while before we even discovered it, wrote the story about it. It blew up. Many other people wrote stories about it. Your first response to the negative publicity was to offer people an email opt-out where if you didn't want to be in the feature if if I didn't want my name to be used, I could email superhuman and say please take me out. Only after the lawsuit did you discontinue the feature. The what we heard the first complaints from a handful of experts. They said I'd like to be I'd like to opt out of the feature and we addressed what they asked for. We then sat down and looked hard at the feature. And to be honest, I hadn't spent any time on it. I came and looked at it and I said this is all strategy for us. And we announced we're taking it down well before there was a lawsuit. The the reason we took it down is it's all strategy. It's not what we want to do. It's not how we're that's not how we want to work with uh with creators. We think we're we're building a platform you should want to be on. We think we're hopefully part of the solution for how you can uh take your work and make sure it's present for people everywhere. It wasn't our goal to be anywhere close to that line. But you know, the the feature wasn't good. We took it down. Tell me you said it's off strategy for you. The feature obviously shipped. What made it on strategy at the time it shipped? You know, I think at the time they were they they believed they were doing that. They were looking at users and they were set they were focused on a user need, which is I wish an expert could give me feedback in this moment. I wish my salesperson could give me feedback. I wish my support person could give me f feedback. I wish my idol could give me feedback. I wish this expert could give me feedback. In itself, I think that motivation that users have is a really good one. And I think one that I I would encourage experts, creators, I would lean into it. I I think it's a I think it's I think a b itig's a big opportunity. Aaron Powell Why would they lean into it if the value for that is zero doll ars? No, I mean I think it should be our job to make sure the value is not zero dollars. We want you to do that. So how much so how how much do you think you should pay me? To be clear, when you when you do the work to bring an agent, craft it, put it on our platform, then you should get paid for it. Just like just like how uh platforms like YouTube work. Yeah, just walk me through the economics. If if you launch a platform that lets me say, okay, Neil Eye Patel can give you advice inside of Grammarly. What are the economics of that platform? How much will I get paid to do that? Yeah, so we're building this business model now. The way the the uh our store currently has a payment model for this that has a 70-30 rev split. It's very similar to how a lot of other products do. So if you want to go build an agent like that, you can do that today. There are a number of experts that already have, and that's the core part of our strateg y. If you already had that system, why build another system that used my Aaron Powell We didn't we didn't have the system at the time. And it are very different features. How many times did you use my n ame? You know, I can't I can't get because it's a legal case, I really can't get into details of of those types of things. But the you know it was it was a very small number for basically everybody. W was there a set group of names? Was it just picking names out of the ether? Was it randomly hallucinating names? It came right from uh it came right from the popular LLMs. So it's exact exactly the same experience you would have if you came to Claude or Gemini or ChatGPT and said, Can you take this piece of writing, recommend the people who would be most useful to give feedback on it, take their most interesting works and use that to try to give me feedback. I you know by the way, I think that's a really hard feature to to make good for users. And and it's not it's gonna take work with the with the with people like you to make that to actually deliver on that need. Did you track how many times you're using people's nam es? I mean we've certainly logged all the all the different interactions, yes. So you you do have a record of like how many times my name showed up or Casey Newton's name showed up or anything. way. Um but we will we'll have to produce it obviously for a lawsuit. Yeah. When you think about that lawsuit, you know, uh Julia's filed it. She's filed a class action lawsuit. There's a lot of ways that could go. You've said the claims are without merit. What did your lawyers say to convince you that the claims were without merit. Trevor Burrus The what did the lawyers say? I mean I think it's actually quite clear. I think that I think it's a layman's test, it's pretty obvious. It's just not impersonation. When you can you'll look at the feature, there's a disclosure next to every single link at the top and the bottom of the of the panel, um, very clearly stating these are uh inspired by these people. It clearly states we have no relationship with these people that that's the feature. I don't, by the way, I'm not trying to defend it as a good feature. I don't want to be on this line. Uh maybe maybe I could step back for a second and say, you know, this is not the first time I've send seen a situation like this. I I used to run the team uh at Google, I used to run the YouTube team. Um when I got to YouTube we, had a big lawsuit from uh Viacom at the time. And the uh, you know, very, very heavily watched uh lawsuit. And we won. Uh we won on summary judgment, actually. Uh we completely crossed the legal bar, but that's not the standard we held ourselves to. We looked at that and we said the the law doesn't require us to do this, but we chose to do a lot more. We launched content ID as a way to make sure that creators could find content that other people uploaded on their behalf. We launched an open creator program, which as far as I know is still the only platform with an open Rev share that's out there. So I don't think the legal standard is the right standard to be looking at. I'm not trying to get close to it. It's fairly clear to me that we didn't cross below it. But that doesn't matter. We're not trying to be close to that standard. We think that creators, we need creators to work, we need their business models to work for our platform to work and it's very similar to what happened at YouTube . We need to take a quick break. We'll be right back. Support for this show comes from Upwork. In order to scale a business, it takes the right expertise at the right time. And with Upwork, you can have a team built quickly. Upwork brings in specialized freelancers so you can move faster and take your business to the next level. Upwork is a one-stop platform to find, hire, and pay expert freelancers across web and software development, data and analytics, marketing, and business operations. Upwork can help grow your business by giving you fast access to specialized talent across 125 plus categories. You can fill skill gaps, launch projects faster, and scale support up or down without committing to full-time headcount. And with Upwork Business Plus, you can access the top 1% of talent on Upwork. You can visit Upwork.com right now to post your job for free and connect with top talent ready to help your business grow. That's UPWORK.com Upwork. com. Support for the show comes from Shopify. Starting a new business, it could be a lonely endeavor, especially in the beginning. And when you're starting out, it's more important than ever to make sure you have the right tools at hand. Shopify is the commerce platform that millions of businesses around the world rely on to sell their products online. If you're asking yourself, what if people haven't heard about my brand? Shopify helps you find your customers with easy-to-run email and social media campaigns. And if you get stuck, Shopify is always around to share advice with their award-winning 24-7 customer support. Best yet, Shopify is your commerce expert with world-class expertise and everything from managing inventory to international shipping to processing returns and beyond. It's time to turn those what-ifs into with Shopify today. You can sign up for your $1 per month trial and start selling today at Shopify.com slash decoder. Go to Shopify.com slash decoder. That's Shopify.com slash decod er. Support for the show comes from Rippling. If your company's all-in-one system has you jumping between a dozen different apps, one for onboarding, another for payroll, and yet another for benefits, then you ought to reconsider your definition of all-in-one. Rippling actually is all-in-one. It's the unified platform for global HR, payroll, IT, and finance. With Rippling, workflows that normally bounce across various tools and departments all just happen in one place automatically. Here's an example. You have an employee who gets promoted or moves. Ripley can update their payroll taxes, manage any new app permissions, ship them a new laptop, issue a new corporate card, and assign any required training all in one place. So if you or your company want to run the backbone of your business on one unified platform with people at the center, head to rippling.com slash decoder and sign up today. That's R-I-P-P-L-I-N- Support for decoder comes from Granola. Have you ever been caught in endless meetings when you also have a major deadline to hit that same da y? Granola AI can help. Granola is an AI-powered notepad built for the way real people actually meet. You simply take rough notes like you normally would, and in the background, granola securely transcribes the meeting. Then it turns everything into clean, structured, actually useful notes when the meeting ends. And you want to know 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 need to do your job better. So if meetings are eating up your day, granola is a no-brainer. You could try it totally free for three months. Just head to granola.ai slash decoder. That's granola.ai slash decoder to get your time back. Get three months for free at granola.ai slash decod er We're back with superhuman CEO Shashir Mahot ra. I have a lot of thoughts about YouTube. I'm gonna ask you at YouTube. I have a lot of thoughts about the Vatican case. I uh I a lot of what happened with Google and YouTube is the foundation for the internet and the policy of the internet as we know it today. That is changing because of AI. So I do want to ask you about that stuff because I think your history uh will will shed a lot of light on how people feel about particularly AI today. Sure. I just want to stay on this for one more turn. You're saying impersonation. That's not the claim in the lawsuit. The claim in the lawsuit is the law in New York and California that bars companies from using names and identities of people for commercial purposes without their consent. And so here you did have a commercial purpose. You were selling the software and names were appearing as inspired by our names. I'm not in this lawsuit, I haven't class hasn't been certified. I promise I haven't sued you yet. But the bar is very different than straightforward impersonation. It is the use of likeness for commercial purposes. And I I'm just you're saying without merit, and I haven't seen you address that specifically anywhere. Aaron Powell You know, I think I'll have to leave the legal arguments for for our uh for the lawsuit and for the court case. I think the our view of it is that uh the the set of work that was there was a fairly standard attribution that w was well above the bar that any other product would uh would do, uh what every LLM on the planet is doing and and uh and so on, and didn't come close to us ing name and likeness in any way that was beyond attributing the sour ce. Aaron Powell I mean the thing I would say to you, and I you've already said this feature is bad, so I I won't hammer you on this too much, is I'm reading the edit that was generated with my name on it, uh which is just bad.. Yeah Like I I would literally never I would literally never give this edit. It says I should r raise the stakes of a headline by adding emotional or stakes based words that could underscore why this launch matters right now. I've been an editor for over fifteen years. I've literally never said anything like that to ask. By the way, you pinned the reason why, right? I mean the idea that you can uncover your editing style from the end work. I just think it's it's it's not possible. I mean it's it's very very hard to come back from that, from that end work and say, what would what would what was the editing pass before that? To do that well, you have to do it. You have to sit down and say, here's how I would edit these things. And I think you can provide that service and you you can get paid for it. And hopefully we're one of the platforms where you choose to do that. And so you don't you you don't have a a like an annotated list of whose names are used in the feature, but you have the logs of everybody who used the feature, presumably those logs have the names in it, and you presume you'll be able to to provide that if you get to the phone I'm sure we'll be asked. Do you think you'll be able to provide that list? I'm sure we'll be asked. We'll see. Because it strikes me that one way you could get around this lawsuit is by just saying, actually, we never use Julia's name until she went asking for it. In the same way that you know OpenAI when they respond to the New York Times lawsuit says this never happened until you prompted us specifically to do the things you said are illegal. And here you you have the same out. You could say, actually we we hadn't until you asked us, we never generated your name. Has that come up? The core of the argument is the what we did is normal attribution of content on the internet. Aaron Powell The reason I'm asking this question is harshly as I'm asking it is I don't think the defense is whether or not people use the product or whether or not the names ever showed up. I think those are just sort of like clear-cut, like binary on or off. Your name never showed up, you can't sue us. You're saying the defense is: hey, that's not how attribution should work. And you used to be the chief product officer at YouTube, and YouTube is defined by creator attribution scandals. Like every year, there's another scandal about React videos. Every year there's another scandal about like the usage of copyright, about whether or not you can make an AI creator out of Marquez Brownley and just run a million videos of him and steal his views. That is it's like the essence of the YouTube creator ecosystem. Do you know how YouTube reacted to this feature when we wrote the story? They invited me to an early preview of their AI likeness detection system because they knew that would be good press for them. If you were still running YouTube, would you have ever allowed a feature like this to go out? Aaron Powell You know, it's interesting the way the way you just described it. First off, some of the ones you described, um describing React videos as scandals is a very interesting way to describe it. They're also incredibly popular and have led to a whole genre of content being created. Likeness detection, uh, content ID, they were all fantastic tools for creators. And they're great tools for for creators. We've we I you know my team built the content ID tool with the same idea. If if Marquez Brownley, if somebody does that to 'em and they copy his videos and put them up, then you can use that tool and he can not only go claim them, he can go make money on them. And that is a that is a tool we built for for YouTube. And I think it's been incredibly popular. And we took what looked like a scandal and well went well beyond it. To be super clear, it's not what the law requires. No, I understand it's not what the law requires, but the use of content ID and the issuing of con copyright strikes, which is something I've experienced. If you issue a copyright strike as a creator against another creator, that is a nuclear move. That comes with severe like social and community consequences. Trevor Burrus, that to be clear, if you use content ID and you and you use it for monetization, you're not issuing content, you're not uh issuing strikes. Trevor Burrus, Right. But I'm saying the YouTube economy, writ large, is sort of defined in in many ways the products are built around issues of attribution and payment and monetization and where the views flow and where the money flows. And content ID is a brilliant innovation because it allows people to get some views and the right people to get paid. YouTube doesn't exist without music. The music is everywhere on YouTube, and the publishers get paid because content ID can identify the music and get them paid. I understand that. But that is a system that tracks attribution and delivers monetization. And I'm just saying I don't I don't see how YouTube could have ever said, we're gonna let Marquez Brownley edit your video without paying Marquez Brownley. Like it would it wouldn't exist in that ecosystem. And I'm just wondering if you can see the distinction. You just said it. But you're not preventing someone from doing it. It's a very different standard.aron Powell But you're making sure that the the people get paid. If I built a video that said, hey, I really like Marquez Brownley, and here's what I think he would say, let me tell some jokes about Millet, like that's a it's a very different standard. The the standard for YouTube was about copyright. That's that's a set of regulations that are covered by totally different part of the law. And in that case, you have a claim that there's a DMCA uh um uh statute that uh allows you to go and enforce your copyright. That's not actually what we're we're talking about here. But I the the the principle of what uh is similar in that in both cases there's a law, the law does not really meet the creator bar. I think the the goal of the community, the goal of our products like ours, working with people like you, is not to use the law as the test. The the goal is to get well beyond that, to align our interests, uh such that your success is our success, and that should be our goal. Are we required to do it? No. I I don't think that's the I don't think that's the requirement. We choose to do it because it's the best way to build the right products for our for our customers. Yeah. Look, I used to be a copyright lawyer. I'm saying I was the world's best copyright lawyer. In fact, I will happily admit that I was not the world's best copyright lawyer. I I understand that people don't understand the the difference between copyrights and trademarks and names and likeness. I'm saying that AI is collapsing those differences faster than ever before. And there are European countries that are just openly suggesting you should expand copyright law to include likeness. I should be able to copyright my face, and then that means I can slide in under the existing legal regime instead of hoping that the United States Congress in 2026 can reach a resolution on expanded likeness protections, right? Like this is a thing that is being suggested because copyright law is more or less the dominant regulatory framework that exists on the internet. And I look at the big social platforms like YouTube, like Instagram, like TikTok, and they have built all these systems to respond to copyright law, specifically copyright, things that can be protected by copyright law, that can be monetized in different ways by copyright law. And then our likenesses are not one of them. Our names and faces are not one of them. Ye ah. And this seems like the place where the things you're allowed to do and the things you should do are gonna be ever more divergent. And you are the one who's experienced it the most loudly recently. And I'm kind of curious if you've learned anything other than there's what the law says I should do and there's what I should do, and I'm we're gonna find the line down the middle. And we'll see we'll see if the laws find a uh a ground on that. I do think it's a catch twenty-two as a creator. I mean the copyright law has been around for hundreds of years now and it's and uh in its various forms started like the way music composition was licensed started with actually Mozart and Bach um and it's and has grown since then uh in almost every in almost every country in the world has reached a very similar standard. I think there's a very thin line between taking publicly available work and being able to refer to it and copying it. And I think the idea that defining all references to work as being uses of names and likenesses, I that I it would break it would break the inner it would break your business. You wouldn't be able to refer to me. How'd you get on a show last week and talk about I mean j uh just to be clear, and I I don't want to be in the inside baseball banking podcast. We made you sign an appearance release. No no no you to come on the show. Of course you talked about you before I came on the show, but in order to run in order to be a a real media company and not fly by night and then to use clips of your face talking, our lawyers need a release. And if you don't sign it, they won't let me use the show because they need to be protected against you showing up you know tomorrow and saying I didn't give you permission to use my face. No, I understand that. The uh my point is is broader than that. You talk about lots of people and that's part of discourse. That's part of how we work. You your articles will link to people, you attribute them. I think that's really important. And if you if you drew a line that attributing something is using their name and likeness, then it's like a v it's a very hard line to draw. Uh but I just want to again I this wasn't an attribution. You just made something up and put my name on it. There's no attribution here. This isn't anything I ever said. It's not something I ever would say. I'm not even sure how you would get to the idea that based on my work that I would ever say anything like this. Right? There there isn't an attribution here. Oh the future. There's no work that existed that would lead you to this outcome with my name attached to it . But I'm sorry, you you you think in my role as editor in chief of the verge and co-host of the VergeCast, I emphasize the importance of crafting compelling headlines that convey urgency. You're not that's not what you're talking about. I'm just saying you're you're ta you're telling me there's attribution, and I'm just wondering what the attribution is. It says b based off of this piece this work from you, we asked. That doesn't it just says this suggestion is inspired by Neil I Patel's The VergeCast. I promise you on the Verge Cast, I've hosted that show for a long time. I have never said what emotional or stakes based words could underscore why this launch matters right now. The VergeCast is not a show about editing headlines about smartwatches. time and say I read this person's work and here's now my conclusion from it. You should decide whether that is a suggestion generated from attribution or not. I told you I think it's a bad quality suggestion. I'm not trying to defend it. I don't think that's what we what what we want to talk about there. Uh but the question of can the internet when you publish work, can humans and A I use it to generate other suggestions, other impressions of uh they can and you would like for them to attribute it. But it's it's not work that that person made. Like hallucinating a thing that you thought I would make and then saying you're attributing it to me doesn't provide me any benefit. It might actually detract from the benefits I could provide to other people. That's the disconnect that's in my brain. I'm not sure why this is an attribution. Like if I'm like I talk to Shashir and I think here's what he would say, that's very different than saying like I'm I read all of his work and I've I've asked you know, whatever quick version of Claude or G HBT to just make something Like th there's something meaningfully different there. And it doesn't it doesn't seem like you're willing to conc ede that. No, I'm not. I think this is uh I I think it's fairly clear that generating a suggestion based on somebody else's work. Just use the simple test of a human doing it. If you did that on your show and you said, I read this person's work and here's my impression from that, this is what I think they meant, you build a whole show based on that. So you don't always get it right. You don't always say things about the people that you're commenting on that are that are correct. That's not right. But I'm not attributing that idea to them. That idea is clearly mine. Clearly stated that this is a suggestion developed by this feature based off this work. Aaron Powell Let me ask you a different question. And I'm curious about this across the whole sweep, right? From YouTube to now. There's an NBC news poll that just came out about how people feel about AI. And uh the answer is bad. People feel badly about AI. AI is pulling I think this is tough. AI is pulling behind IC and only slightly above the Democratic Party. This is a tough spot to be in. It's a negative 20 uh uh perception. I think the reason for that is because it's so extractive and the value isn't there. And I would compare this to YouTube, which a lot of people thought was pretty extractive, right? You fought a pitched copyright battle about YouTube, about whether South Park could be on YouTube without permission and Viacom was going to see you. And that case was fascinating because the public was decidedly on YouTube's s ide. Oh, that covered memory of it. I covered that case. I was in law school in school, like in studying copyright during the case, and the vast majority of people were like YouTube is really useful, we love it, and these big Hollywood companies suck. Uh Napster, the public was not on the side of the record labels. They were not on the side of larger look. They were on the side of file sharing because the utility was so high regardless of the economic or social cost. I could keep going on and on with this. Uh y you can tell people all day long about the the labor costs of Uber and they're still gonna use Uber. You can tell people all day there's a trial right now about whether social media platforms are damaging to teens health, uh whether they're defectively design products that hurt kids. The trial is ongoing as we speak. The jury is empaneled right now, and people are still gonna use those platforms because they don't care. The environmental costs of big stupid cars. You can tell people all day that trucks were in the environment, Americans buy trucks. That's what we're gonna do. AI isn't do it that isn't there. It is only perceived as extractive. It is not as b it's less beloved than ice. That's crazy to me. Do you understand that the the extractive nature of AI is causing a problem for the whole industry. Because you're sitting in the middle of one of these controversies right now. I think you're drawing a pretty broad link for why why people are afraid of AI right now. I think great consumer products that provide a lot of value overcome their social costs. There's lots of opportunity. It does meet your other tests. It has created some of the most popular products in history. And there are many people who would have you pry any of those products from their cold dead hands. I think that the challenge with AI right now is it's challeng uh I think it's challenging people's sense of future of their uh their humanity. And it's their ability to work. It's their ability to that those are really the challenges there. I I think I think the line we're talking about here, I don't think that's actually where what you're reading into that pol l. People are scared for their jobs. I think so. I think that's extraction, right? You've taken the sum total of everyone's work on the internet, and now you're going to use it to replace human beings and their jobs without any economic recompense. But by the way, I think they're wrong about it. I don't actually think it's going to replace as many jobs as it's going to create. I think actually one of the reasons why I think our our model for thinking about AI is about bringing it to people and expanding their work. We like to call it the the product that helps you become a superhuman. So I think they're wrong about it. But if you're asking me why does it pull so low, it's because the copywriter feels like maybe I'm not gonna I'm not gonna meet it anymore. It's the salesperson who says, I think uh or a support person who says, I I wonder if an agent's gonna be able to do my job. Um I think the idea that that has something to do with name and likeness, I think is a pretty big stret ch. Well, I mean you're again you're sitting in the middle of a controversy where a lot of people are mad at you for appropriating their work. If you're a copywriter at an ad agency, I know a lot of copywriters at ad agencies, they're saying you took all of my work. The AI, not you, the AI companies have ingested all of my work for train ing. And now they're gonna replace me and no one got paid. Uh Hollywood is basically like no one's no one's paying us for this. I think people who write Tumblr are saying now OpenAI is gonna make uh porny fanfic for people. That was our job. Why didn't you pay us? I don't think it's caused just by this feature or just by the latest advance AI, they're facing a hard future for a lot of different reasons. I think the poll you're referring to of the broad population, the broad population is not creators. The broad population has jobs that they are afraid may not be available to them. Whether they're a truck driver, whether they're uh whether they're a support person, that's what they're afraid of. I under I'm not I'm not diminishing the fact that creators also have an issue with AI. I'm just pointing out that the the broad impression of AI, the challenge we have with it is that I think the whole entire industry has done a really bad job of helping people understand why technology like this can help them and not prevent their job from de being taken away. And most people just aren't creators. I'm not I'm not objecting to what you're saying about creators. Just saying most people aren't stressed about that 'cause that's not their job. That that's not what they're they're individually afraid of. Aaron Powell No, I I understand what you're saying. I'm just pointing out that almost every major technological shift has been extracted in some way. Google copied all the books in the world without permission, and then we had a Google Books case, and Google had to win that case. And they did, they were able to do it, and Google had to win the Viacom case with YouTube. Google had to win the Google Images case against Perfect Ten, which was maybe the least sympathetic plaintiff of all time because it was a porn company and Google was doing Google image thumbnails of softcore porn. And it was obvious that Google was going to win that case, but they still had to win that case. All of this stuff got litigated at pretty intense levels in ways that are precedent still to this day. And it doesn't feel like we're spending the time to litigate, hey, you can just make a deep fake of my face and use it to sell headphones and Alibaba. Hey, you can just start a company and say, well, it's attribution. So I'm just gonna use the names of famous people on my product to say these are the edits. Like there's a link there that seems very direct to me, maybe just as a creator, but also I would submit to everyone else who says there's a pretty extractive cost here and the consumer benefits are not legally expensive. I think it's a good way when I talk to our team about why the legal standard shouldn't be the minimal standard we try to hit. I will also tell you that what we're doing here at Superhuman, I don't expect to be very close to this line. I think there are other products that are very close to this line. Um I I I think our core strategy is about building a platform that you can choose to participate in or not. And I don't think we're gonna be I I don't think it's gonna be a fine line for us. Uh I know in this case we we built a bad feature, it was not received well by either users or experts. I don't like that. Uh I killed it for that reason. Um, but I don't think our I I don't expect to be sitting here, you know, the the YouTube analogy, you're right, the Viacom case had to get litigated for YouTube to exist. And if it gotten litigated the other way, YouTube wouldn't exist. Actually most of the internet wouldn't exist. Um and so the idea that it got litigated that way, I think it's a I think it's a win for everybody. I think it was a win for society. I do think it was a win for YouTube. I don't expect that to be our case here. There's not a line I'm gonna be cl ose to. What happens if there's a bunch of copyright cases against the AI companies? Um I feel like I should disclose that our our company has sued Google over ad tech. It has nothing to do with with AI or copyright, but I feel like I need to disclose it because I disclose everything all the time. Labs over copyright infringement. The New York Times has sued OpenAI. There's a million of these copyright cases, Floyd. There's more every day. One of them could go the other way, right? And this industry could face plan. What do you think happens if that goes if what do you think happens if one of the big AI labs loses a copyright case? Are you asking me as as uh as someone watching the industry or are you asking me in my superhuman role? My superhuman role is straightforward. I mean the models the the uh whatever the models do is what we'll use. And so if the if the models end up uh needing to restrict that behavior, then that that is what it is. We sit on top of the models. Uh I don't think we'll be the ones in the middle of those cases. If I look from an industry perspective, I I think it's a really hard case, both directions. And I have I have real empathy for both sides. I mean copyright law is like you said it's the it's you know it's some of the it's some of the work that has allowed the internet to work and not everybody is happy with how the law draws the line. And I think you're right that YouTube tested that line in a new way, Viacom K so on. I think what OpenAI and Claude Gemini are doing are going to test it in a new way. And I think I hope they find a good line for it. Um I don't think that's where we're gonna be. I mean we're it's not the we're not gonna be the ones in the middle of those uh those lawsuits or those figuring out where that line is. We need to take another quick break. We'll be back in just a minute . The world moves fast. Your work day even faster pitching products drafting reports analyzing data microsoft 365 copilot is your ai assistant for work built into word excel powerpoint and other microsoft three hundred sixty five apps you use, helping you quickly write, analyze, create, and summarize. So you can cut through clutter and clear a path to your best work. Learn more at Microsoft.com slash N three sixty five cop il Support for this show comes from LinkedIn ads. Sometimes even the strongest B2B marketing ends up in front of the wrong audience. If someone's seeing ads for high-end cookware, but can barely make instant ramen, that's a sign your strategy needs a reset. So when you're ready to reach the right professionals, you should check out LinkedIn ads. LinkedIn has grown to a network of over 1 billion professionals, including 130 million decision makers. And that's where it stands apart from other ad buys. You could target your buyers by job title, industry, company, role, seniority, skills, and company revenue, so you can stop wasting budget on the wrong audience. It's why LinkedIn Ads generates the highest B2B return on ad spend of all major ad networks. Seriously, all of them. If B2B growth is the goal, LinkedIn Ads is one of the most efficient ways to put your message in front of the people who can actually say yes. Spend $250 on your first campaign on LinkedIn ads and get a $250 credit for the next one. Just go to LinkedIn.com slash decoder. That's LinkedIn.com slash decoder. Terms and conditions app ly. Support for this show comes from Vanta . If you're a business owner, you know that risk and regulation are rising, and customers now expect proof of security from the get-go That's why Vanta says they can be a game changer for you and your business. Vanta automates compliance by bringing risk, trust, and security into one AI-powered platform. They support over 35 security and privacy frameworks, including SOC 2, ISO 27001, and HIPAA, helping companies get compliant faster, stay compliant, and unlock new growth opportunities while saving valuable time. Vanta says that companies like Ramp and Writer spend 82% less time on audits with Vanta. That's not just faster compliance, it's more time for growth. You can get started at vanta.com slash decoder. That's V anta dot com slash decoder. Vanta dot com slash decod er. Support for the show comes from MongoDB. If you're tired of database limitations and architectures that break when you scale, it's time to think outside of rows and columns. Because let's be honest, you didn't get into tech to babysit a broken database. You got into it to actually build something. MongoDB lets you do that. It's flexible, developer-first, asset compliant, enterprise ready, and built for the AI era. Say goodbye to bottlenecks and legacy code. Start innovating with MongoDB. There's a reason it's trusted by so many of the Fortune 500, and that's because it's a platform built by developers for developers. MongoDB. It's a great freaking database. Start building at MongoDB.com slash build . We're back with superhuman CEO Shashir Mahot ra. If the incremental cost of a token skyrockets because suddenly the AI companies have to pay massive licensing fees to copyright owners downstream. What happens to your busin ess? I don't think it really matters to us, because it'll all happen in the models underneath us. But I don't mean to say it doesn't it doesn't matter to us as a as our own entity. It matters to me as a citizen. I mean I think it's I think it's really important. But I would also remember like for us, the primary agents people are trying to build on superhuman have nothing to do with this. I mean the expert case is one case. I mean the our the what people are doing with our p our product is they're go they're going and taking their sales methodology and turning it into agents for their for their salespeople to be able to use. They're taking their um uh their support tools, they're taking their calendars and making sure that as you're writing an email and saying I can meet tomorrow at six p.m., please make sure that I'm actually free then. I don't like I said, I we're just not this is not No, I'm I'm not saying it the expert review part. I'm saying you're describing take all of my sales literature, take my calendar. Yeah. That gets loaded in a context for a model that you call. Yeah. Yeah. If the incremental cost of a token in that model goes up because the AI companies suddenly have to pay a bunch of copyright licensing fees, what happens to your business? Aaron Powell I mean if I were those companies, the solution I would have isn't to go distribute that cost across all users, I would charge users a subscription for using that for using that information. I think that's the business model they should have. Um my personal view is what should happen is I should come to ChatGPT or Gemini or Claude and I should prove that I'm a New York Times subscriber and then it should give me answers for the New York Times. Um and the New York Times is gonna have to make a choice of do I only want I have my content to be used for my subscribers or not But the the a lot of these all these cases are different. So I I I I'm gonna generalize here and you can attack me for generalizing, that's fine. But broadly they they split into two lines. There's one, the thing you're describing, which is you spit out content that I've already made. Right? Like Suno can make a Beyonce song, that's copyright infringement on output. The other set of cases, right? Is on input, is on training. Right. And saying you actually ingested all my material without permission, that's also copyright infringement. And if that goes the wrong way for the model companies, their cost structures change in retrospect. Right. Because the model itself is the same thing. And I think if you actually produce something that could be mistaken for the work of another person, then you can file a claim, you can get you can get it taken down, and then you can get if they choose to leave it up, you can you can choose to negotiate a revenue share agreement or whatever you might want to do with that. That that's out output is cleared. Input is not cleared, like you said, and the the cases haven't been resolved in a particularly, you know, uh clear way. The point I was making is if I were them, I wouldn't take the cost of input and distribute it across all users. I would split the model. If it really went that way. And I would say, fine, you don't want you don't want your content there. I will build a version of the model that is just for New York Times subscribers and charge them. Your your particular question was will that cost get passed along to the other used I mean that is what's happening right now. I mean they are paying that content. It is being passed to us. You know, does it matter to us? I mean, frankly speaking, the pace of innovation in that category is so high that the the uh in the profits being generated there are so high that no, it hasn't mattered to the upstream users, to us, to chat GPT users, Gemini users, so on. It hasn't stopped their growth at all. Um will it someday maybe? I I I don't know. But I my my point was more in this world of like output copy output copyright is fairly clear, law covers it pretty well, input copyright is not that clear. You know, and it's not clear for good reason, right? If a if you're a human and you read a book and then you f and then you learn something and then you talk about that thing, like you know, what should happen? And that that's that's a a legitimate question that hasn't been well tested in in uh in the courts. I don't I don't think the industry is going to take that cost and just pass it along to all users. Um but we'll see. If it does, then then it does. And we'll have to deal with it. Everybody will. The other dynamic here, by the way, uh most humans cannot infinitely scale to create trillions of dollars of enterprise value by reading one book. That's the difference. Right? Is you to get that value at that scale, usually lots of people have to buy copies of the book. And the the economics spread out. Right. I mean this is the the scale is the difference. It's I I think it's a I understand that's the that is a very fair argument that hey the this is not the same as a human reading the book. And obviously that's that's the the line being taken there. I I would pro postulate that w whatever way that case ends up, the correct answer for experts is it's time for a new business model. And I think the idea that that's gonna that's gonna get into the exactly right spot. Uh and you know you're gonna get pennies for every for every query coming through Gemini. I mean that's certainly one path. Uh I think when you know when I went and talked to people about what we're doing here at Superhuman, what they told me is actually I don't really want to be fishing for pennies whenever my work gets used. I want to build connection with people. I didn't build content to put it out there and get paid a fraction of every use. I I I want to go build a product that actually connects with people. I want to do I want to do this. And I think you know I think YouTube offered a great way to do that. I think what we're doing with superhuman should offer a great way to do that as well. Let me let me ask you about that specifically. I was in Southwest Southwest. We have a little baby and it travel this year. But I you know I watched Instagram. I I experienced Southwest Southwest through magic of Instagram and TikTok. You had a suite there at South by Southwest. I looked at some of the videos. The caption on one of the carousels. I'm just gonna read you the caption. This is from the superhuman suite at South by Southwest. A lot of talks there. The summary of the talks was: AI can't replace human creativity, empathy, or emotion. It won't take all of our jobs, but it will shape reshape how we work. And in the AI era, taste and judgment are more valuable than ever. Valuable on what metric? Is it dollars ? Valuable on every metric. I mean I think it's specifically dollars. Dollars are what I pay my mortgage in. Is it dollars? If my taste and judgment are more valuable than ever, but it's also infinitely replicable and you think I need a new business model or every creator needs How do I make more dollars? If my taste and judgment are more valuable than ever, where do the extra dollars come from? So the the just to be clear on the tagline for Superhuman, what we believe is that we can help all our users become superhuman by bringing them tools that allow them to expand their work. The main way we think about people is that Grammarly doesn't do your work for you. Grammarly helps make you a better writer. And you still publish your essay, you still put post your article. It's our job to turn you into a superhuman. That's our promise to our users. So that's what that b uh what the banner is about. I think your question I think is a very good question. What what wait that's how the ban the banner says taste and judgment are more valuable than ever. And I'm just asking you to define the value and what value is going up and what value is going down. If you're if you're using Grammarly and you're a student and you're a person who's a uh a salesperson are it is your taste and judgment, that is actually what gets valued in the end. We're here to help make sure you don't make a mistake. We're here to help make sure that you present yourself the best possible way. That's what that banner is about. We're addressing, we have 40 million users who use our product. The vast majority of them work in professional industries, they're they're they're salespeople, they're uh support people. That that that's that's who that's addressing. And we're trying to tell them, don't worry about losing your job when you use our products because we're here to help you scale more. We're here to help you be a better version of you. That's what that banner is about. That's what our promise is about. I do think we have a proposition for you as for you, Nile, as well, which is that you can now become one of those assistants to all those people. And many of them have no idea that they they could use your help. But if you can build that relationship with them like Grammarly does, and people personify Grammarly all the time as my high school teeth of my high school English teacher sitting next to me everywhere I work, that makes me better. If I can have my high school industry with me everywhere I work, it makes me better. It makes my trust and judgment shine through. I would like your agent to for the people that matter to the that for for people for whom you matter, you should be able to build an agent that sits right next to them and you can actually feel like their editor. Now you gotta do some work to make that a to make that a good experience. You're gonna have to figure out how to document your editing style in a way that actually produces a good result, not like the one you quoted earlier. But if you can do that, you should be able to build that relationship, you should be able to construct it the way you want, you should control it, and you should be able to make money on it. And you should hold hold on. I'm just g onna You understand that you're saying I have to do that because all of the work I've produced in my career to date has been taken without compensation by AI compan ies. I I I like I need to invent some new business model as an expert and upload an agent of myself to your tool and then advertise it to get a seventy thirty revenue split from however many people use Grammarly, because my my actual body of work has been reduced to zero value. Okay. That's a pretty hard sell. I mean I'm not I'm not here to tell you uh how to answer every question for what's changed in the creator economy. I think it's one way to look at it is that uh the path to being a creator has become harder. Um I think there's other ways that you know the the I assume this is gonna end up on YouTube and Spotify and so on. There's paths to becoming a creator that become easier. And I think there were there were folks that when YouTube came out, they told us all the same things. And they said, we don't understand. Our business model is screwed over there. And so why should we work on YouTube? And they the ones that looked at it that way and saw it as replacement ended up not moving forward to the future. Obviously you did. You run a show in the on all these platforms and you figured out a way to turn that into a business. And I think that's you saw that opportunity and you expanded what you could do. I think there's about to be I think if we look at AI from that perspective and say AI is here uh and it's reducing the number of people who need to traffic to my current experiences, I think that's one way to look at it. And you know, I and I think there will be some creators that look at that way. I would hope we look at it the other way and say some of these platforms are going to give you a way to participate, are going to give you a way to take your expertise and put it in front of people in a way that actually helps them in a different way than you could connect in uh connect in the past. And I and I think that's I think that's a bright future. I'm not really trying to say you have to or you don't have to. I think it's an expansion opportunity. I'm not really here to defend what some other company is doing with with content, uh what's happening there is happening there. I'm just saying creators feel that pressure. We recognize it. I think there's an opportunity. I had one creator, I had one creator tell me that their traffic in just the last year from Google is down 50%. They said the you know AI overviews and so on, traffic is down 50%. They sell books. And my reaction to them was that really sucks. I mean, I I I understand why that really sucks. I would also tell you if, you're a book author, like waiting for people to search your name on Google has got to be like the least good way to monetize your expertise. So now let's talk about how we can take what you do well and get it in front of people in a way that creates value in a different way. And maybe we can do it in a way that isn't tons of incremental work for you and brings you a new type of uh new type of opportunity. And I think I think platforms like ours are gonna give that opportunity to people who choose to take it. Not everybody will . Can I extend this to you as uh the CEO of a software company? Sure. This is the same argument I hear about the frontier models, the AI companies, their sort of relentless expansion into every category, and then what you might call the cesspocalypse. Why would I pay your margin on tokens that you're buying from them when I can just buy their tokens directly and just talk to Cloud? Why wouldn't I just vibe code something that looks like Grammarly and run it instead of paying what you're like $160 a year? Like this is the thing that's coming for the software industry writ large. Do you feel that same pressure? I personally think it's a little overstated. Um I think that let I'll give you my view of it. So I think that there's a lot of software. The ability to build uh the ability to build software is definitely getting much, much easier. Um I think the reasons why people choose to use software is uh often because because it does a job particularly well uh and that there's often a network effect associated with it. Let me I'll give you an example and I'll just I'll just focus on CRM. So the people look at the SAS pocalypts, they they they go and try to judge Salesforce and say, why would anybody pay for Salesforce? I could just vibe code my own version of it. Um well first actually, first they say, why would anybody have a CRM? And then it's if they do need a CRM, why why would they pay for Salesforce? So maybe I'll answer both questions. So why pay for a CRM? And you know, my my view of it is that you know when you have when you have groups of humans working together, you need software for them for them to work together. So if I have if I have one salesperson, you know, I can I can keep all my sales in my head. If I have 10 salespeople, you know, maybe I can do it with a spreadsheet. When I have 100, I need, I need software to keep them together. That that software today is called CRM software. I think it's when I have a uh when I have a thousand agents selling on my behalf, I'm gonna need a way for them to coordinate with each other. It might be different, but I do think it's gonna be important. Why is it gonna be products like Salesforce? I don't know if it will be Salesforce, but the reason is it become uh I think all the powers of network effects are going to become much higher. And you're going to say these are products for which I'm going to pick the product that is plugged into the ecosystem in different ways. Why would people uh rebuild Grammarly? I mean they'll I'm sure they'll try. I mean my hope is by that point, we are the platform for all the best agents that work right where you work. And you'd have to go replicate all of them. And I'm sure there will be people that will. But I think most people won't. And I think that's a it's an important bet for how the software industry moves on. I think the I think the need for software is only going to increase. And I think the importance of network effects will only increase. Aaron Powell You don't think that OpenAI or Anthropic or any uh Google will say, well, Grammarly is pretty useful. We can build a tool that looks just like it in seconds and ship it and kill their I mean, they're just buying our tokens anyway. We can just kill them pretty easily. I mean, I think the ability to build that tool has existed for a long time. So if that were true, our business wouldn't be growing. We wouldn't have 40 million people using it every day. I think the idea is getting easier and easier. Yeah, we can't stand still. If we stand still and don't continue to innovate, if we don't build that network effect, if we don't continue to add value for people, we'll get caught. That's always true. I just wanna end on a big thing. Again, you used to run these platforms, you're on the board at Spotify. I know you think about the the economy here and how work gets produced and who gets paid as deeply as any one. I look at the shape of the media landscape right now, the information landscape that you might call the internet. And I say, boy, everything is slowly turning into QVC. Right? Like making this stuff is getting devalued every single day. Being the person who makes the stuff is getting harder and harder. It's something you've repeated several times now over the past hour. And at the end of it all, the creators all have to pivot to selling something. The Paul brothers have to sell you bottled water. Mr. Beast has to sell you energy bars. Like we are we we've devalued the work so much that unlike any other industry in the world, the internet industries, the information ecosystem pivots from bits to at oms. That's pretty rare in the history of business. Most businesses pivot from atoms to bits, right? The the margins and bits are historically much better than the margins of atoms, except on YouTube. Except every major artist has to be on tour forever because the money from selling music itself is solo. AI is bringing that at scale. Like you can feel the pressure. This whole this whole conversation has been about that pressure. And maybe the legal doctrines don't line up exactly, and maybe I'm making too many generalizations. And I hear the criticisms that you've you've you've parried me with, but that's what I feel is that all of these platforms at the end are becoming about someone trying to sell you something else. I think and AI is just accelerating that. I'm just wondering where you think the endpoint is. Actually, I think it's an interesting characterization. I mean, I think I think there are multiple business models out there. Um what you described as bits uh to Adams, I think is one way to look at it. I mean, I think the reason why YouTube creators end up with those other opportunities is b not be I mean, I'm sure some of them feel like the ad ad revenue from YouTube is not enough. It's because there's an opportunity, right? They they why would you not take an opportunity? I think have to is one way to to describe it. Get to is a different way to describe it. The other thing I'd say is I don't really think it's quite accurate to say bits versus atoms. I think it's much more advertising versus subscriptions versus purchases. Um, and I don't think the spread on that is really about the bit atom piece. It's about the connection piece. There's a there's a set of platforms that are built off eyeballs. I mean, what I built at YouTube was primarily built off eyeballs. And there is, you know, uh overall of history, the amount of advertising spend has always been some percentage GDP. It's hovered between two and four percent of GDP forever. And that gets divided up amongst all these eyeballs and that that that is one business model. And yes, the number of creators fighting for that has dramatically fragmented over the last couple of decades um in every platform. And so what can come from that is smaller. There's also the ability to sell products and every and it, you know, the uh the as a as old as time uh has been the ability to sell products, and in the middle is the ability to build connection. And I think those products tend to do uh a lot of work with subscriptions. I mean, it's interesting when we think about some of my favorite creators. Uh, you know, many of them subscribe to the the thousand fans theory that if you can get a thousand people to pay you a hundred bucks a year, you always all of a sudden have a hundred thousand dollar business. Um, and I think that there's a whole class of people for whom they've decided if I can I can either go somewhere where I get a little bit of money every time somebody happens to blink and look at me, or I can get them all the way down the funnel funnel to buy my hamburger or my water bottle. Or in the middle is I can build a deep enough connection with a person that they're willing to pay me a substantial amount of money on an ongoing basis. And I don't need a lot of them. And if I can do that, then I can I can build a real business out of it. And there's some there's some fantastic creators who have done a really good job of that. You know, I uh uh many of the ones I'm sure you know. The I think what what I'd like to do, what we're trying to do with superhuman and our agent platform is enable people to build that level of connection. I mean, a lot of them are doing newsletters. I mean, it's it's very meaningful to say, I got a newsletter, it's a hundred bucks a year. Um, here's how you can do it, thousand people get three to a hundred grand, ten thousand people get three to a million bucks a year. Feels like a feels like a a a meaningful uh uh connection. In our case, I'm saying AI is gonna allow us to do more than show up in your inbox. It's gonna allow you to show up with a red pen and a blue pen right next to the person and say, I can help you in the thing you're doing, at least the the part of it that we're working on. And I I'm willing to gamble that can you go get a thousand people to say that's worth a hundred bucks a year to me? I think you'll be able to. And I thought Do you think wait I I I come I'm just gonna ask you this as directly as I can. Do you think that feature will be good? I think it'll be as good as the peop the work that the creator puts into it. And it's a a little bit like, you know, is are all newsletters good? No. Most newsletters suck. You know, is every YouTube video good? No. They're they're mostly they're quite terrible. But does it allow but I but I do you think you can get an like I don't what your tool looks like to build an agent inside your platform. But I haven't seen an LLM that can replicate my writing, let alone my edit ing. Do you and I you know you're dependent on the capabilities of the models themselves. So I'm asking you kind of in a general way, but knowing how your you know how your tool is built, can you actually make a tool that can do that well? I I think so. And I mean I would say that I think we did a pretty good job with Grammarly, that we re we replicated a grammar teacher pretty well. You know, can we do that with with a broader spectrum of things? I believe so. Uh I think we have some good evidence of it already with with some of the agents working on our platform. Can I do can we build a good one for you or can you build a good one for you? I don't know. What what does that tool look like? How does it how does build a good tool that lets me edit look like? You know, I think the start of it is actually i it's what you said earlier. Like you have to you have to write down that that viewpoint of like what is your editing like. So that that's No, but I mean literally describe the interface that your tool provides me to do that. Oh it's uh the the big part of the interface is is a is a uh a prompt box and what we call triggers. And so you're gonna say, here's here's my instructions. Think of it like you're gonna publish your manual and here's your trigger. So here's a set of things that say, when you see this, do this. And here's my manual, here's how I think about things. And when you see this, do you know do this. When you see it, like you you gave the example of uh feedback on a headline, you didn't like the feedback it gave on the headline. It's uh uh uh uh reasonable I I wonder if you could write down what feedback would you give on a headline? Pretend for a moment you were trying to train someone else.