Inside the AI Startup Frenzy: ‘Everyone’s Pivoting, Then Pivoting Again’  ## Metadata - Author: [[Decoder with Nilay Patel]] - Full Title: Inside the AI Startup Frenzy: ‘Everyone’s Pivoting, Then Pivoting Again’ - Category: #podcasts - URL: https://share.snipd.com/episode/0535eb50-6786-493b-ac3e-94714598be30 ## Document Notes ## Highlights - **Ellis Early Snapchat Experience** - Ellis talked about his early coverage and relationship with Snapchat and Evan Spiegel. - He was drawn to Snap by Evan's thoughtful design and unique communication approach. Transcript: Alex Heath You started covering Snap a lot. And it's really weird having this conversation because you and I have like, we've, I don't want to say mirrored each other, but we've had a lot of similar epochs in our career. Like you covered Snap a lot when it was just getting started. You did some of the first interviews with Evan Spiegel, the CEO on the verge.com. Then I went on and covered Snap for a long time after you went to Snap. But like, talk to me about that time and like why you honed in on Snap and then how that led to joining Snap very early. Ellis Hamburger I think at the time, Evan was pretty much the youngest founder I'd ever met when I was at The Verge. And he was just so unbelievably thoughtful about every design decision, which is kind of funny and maybe a little bit sad, right? Because the way the app looks, it looks like a toy and in many ways still does. And I'm not sure, you know, how much Evan really gets in public, you know, in terms of mindshare for anything except maybe stories, which in my view was never necessarily the perfect Fit for Snapchat anyway as a private messaging platform. But I mean, I remember, you know, when he explained to me that you were going to get a notification before your friend was done typing, like right when they start typing, you're going To get a notification. And I was like, that's insane. Like, there's nothing to look at yet if you go into the chat. And he's like, well, the whole point is to get into the chat together just like in real life instead of having this awkward text ping pong of people leaving each other on read. And I was like, holy crap, that's a real insight, just like ephemerality driven by real conversations with friends at a coffee shop, at a bar, at a restaurant that just makes people feel Something different. There are a lot of examples of that through Snap over history. And I think at the time, most reporters were interested in just asking them over and over again if it was a sexting app. I was the one who cared about design. So that made us fast friends. And eventually, I was not really picturing myself as an editor-in in 50 years. And so I started to think about what I was going to do next. And so I was looking at Facebook. I was looking at some others just because I knew so many folks at Facebook from being a beat reporter at the time. Evan eventually invited me to join and thus began the next chapter of going from wearing all black Brooklyn bro to jeans and white tees LA bro, Ellis, I suppose. Alex Heath Yeah. And on LinkedIn, it looks like you were obviously very early at snap and marketing and ([Time 0:07:59](https://share.snipd.com/snip/1da5e097-b0c0-4674-8f52-8ae2ffbb46d5)) - **Marketing and Positioning Importance** - Founders should invest early in marketing and positioning as key drivers of success. - Clear visionary storytelling helps define product goals beyond incremental gains. Transcript: Alex Heath Go to what you did after Browser and what you're doing now. You are, I jokingly tell you that I think you're like the Rick Rubin for AI startups, but you probably cringe at that and that's okay. But explain what you're doing, what the company is and yeah. Ellis Hamburger Yeah, it's funny. The Rick Rubin comparison is interesting to me because I do think of myself and I think he probably does think of himself as a bit of a therapist, but he doesn't know how to work the knobs And seems to be a little bit proud of that. I would like to believe I know how to work the knobs, whether that knob is a quippy social bio or your Super Bowl video script. So yeah, maybe a little bit like him, a little bit not. But yeah, after being at the browser company for a couple of years, I started meaning, which really was a leap of faith based on a gap I saw in the market, which is that most early stage companies Don't yet have a head of marketing. They don't have a head of comms. They certainly don't have a UX writer, something I did at Snap that I always thought was really underappreciated part of the product experience. And I said, what if I kind of do all those things for founders? And what I realized is that they saw something different in me, someone more like them that could speak founder in a very multidisciplinary way. And honestly, I think Evan and Josh thought of me at that same way early on at both Snap and Browser. The difference is the bigger companies get, whether it's Snap or Browser, inevitably people who wear a lot of hats get put in silos just because there's that much work to do. You know, a browser company, there are that many tweets to write, that many release notes to do, that many reporters to ping. And I think packaging it all in one, trying to create a blueprint for them that really works for every audience is what I try to do today. And, you know, I still do do some writing, like website copy, video scripts, et cetera, some of which we can talk about if you want. But at its core, I think it's about positioning. And it's funny, that brings me back to reporting. I'm so grateful that, it's funny, I think Business Insider wanted me to write five or six articles per day, which sounds like a lot until you remember that Joe Weisenthal used to do 20 A day. And ([Time 0:19:45](https://share.snipd.com/snip/056199d5-4599-44d5-a85b-88855b6b77bd)) - **AI Startup Pivot Frenzy** - AI startups rapidly pivot their products due to fast-moving technology. - Different companies frame AI agents variably, showing broad experimentation with interface design. Transcript: Alex Heath Out, like, what are you saying and what's interesting? Ellis Hamburger I think what's interesting to me is the debate about the interface. You know, I've always someone who's really cared about software design. And I think people want to feel things when they use products, when they wear clothing, when they go places, they don't always necessarily want the most streamlined minimal experience, Which is San Francisco design. You know, iOS 7 up till now with, what is it called? Liquid Ass? Your words, not mine. I'm the guest. Yeah, I can say it. And so what's interesting to me is that every time there's a really big software shift, people want to know, oh, everything's going to be voice now. All right, my app's going to be voice. And it's like, dude, people like pictures and videos a whole lot. It's not all going to be voice. And then AI comes out and everybody's like, oh, everything's going to be a chatbot conversation. And it's like, dude, do you want to make music inside of a chatbot experience? And I think this is the nature of tech. Everybody tends to overcompensate on these different user interface trends, in part because those are the kind of things that VCs like to bet on, right? I guess you've been doing this long enough to remember like the original chatbot wave that was just driven by so much FOMO. And it is, man, it is FOMO and fast motion right now. Companies that I'm working with are pivoting during our process and then pivoting again. The technology is moving so quickly that it's very, very difficult to lay out a vision. And so I think, you know, you could talk about the different vertical use cases, different industries, this and that. But one of the things that's interested me is, you know, if you look at mainframe client that's working on agents, they're kind of thinking about working with agents almost as like a To-do list and delegating things to agents. And if you look at another one I recently worked with called Strawberry, which is an AI browser, they're framing the different ways to work with agents as friendly little companions. Or if you look at Dia, the newest browser company evolution, and disclosure, I am still a shareholder in BrowserCo, they're framing them as skills that you access through a slash command. I'm very interested in how that's going to evolve and how people do want to interact with it. I'm definitely a believer that chat nor voice is really the best way to do a whole lot of things, especially if you've ever been on a New York subway. There's definitely a lot less privacy when you're speaking all of your search queries out loud, right? But I do think ChatGPT has become just so powerful that in the same way that Google was able to adopt so many use cases across images and shopping and maps and reviews, that the intent layer, If it succeeds, as I think ChatGPT is trying to do today to become an OS, whether that's hardware, software, some combination, that intent, capturing that intent is everything. And even if the OS isn't perfect, if that's what people get used to, they're willing to put up with some cruft and jump over some hurdles in order to get to the solution. So yeah, I think that's an area of experimentation that's been interesting to me. Alex Heath When you're talking to founders who are trying to figure out how to communicate what they're doing with AI, what do they say behind the scenes that's different from how they portray It publicly? How do they actually feel about AI? Ellis Hamburger I think no one wants to think that they're the one destroying the world. And there's part of that that's just like, oh, it can't be me. I'm just doing a little AI app over here, right? But collectively, we are all absolutely accelerating this revolution that honestly keeps me up at night. And so I see as part of what I do to help me rationalize my job is try and push founders toward documenting goals for society and for people, not just efficiency, saving time and money. Anybody can do that. There's no ethic or morality in that. I'll give you an example. One of my very first clients, I worked with Amo, the former Zenly folks from Snap, Antoine Martine and the crew over there. One of the things I was proud of is that after all those years at Snap, it was with Amo that I was able to document this source of truth manifesto with them with some principles, one of which Was friendship is a feeling. And we elaborated below that to being about how it isn't about metrics, it isn't about this or that. And I really do believe that as companies grow, you need those principles about your goals for society embedded in the company culture or else without it. PMs are just going to default to incremental gains, which we all know are incredibly easy to get, right? And so, I mean, ([Time 0:24:31](https://share.snipd.com/snip/2d6d3194-f747-4bba-9275-cd086f05fbd6)) - **AI Founders' Ethical Blindspot** - Many AI founders focus on efficiency and valuation over societal impact or ethics. - Without visionary principles, companies default to short-term gains driven by metrics. Transcript: Alex Heath Or whatever and help them get buzz. And I don't see a ton of morality in any of it, to be honest. And you're right. That gives cause for concern. I mean, luckily, I guess for humanity, most startups fail. So like the fact that that is so prevalent doesn't necessarily mean that this is going to be the world we live in. But at the same time, it's also indicative of the culture, right? The culture is very much like, yes, AI is really cool. There's a lot of really cool applications of it. It's also kind of like the new Wall Street. I'm curious, like, can you figure that out pretty quickly when you're meeting a founder? Like, where are their motivations actually lie? And does that affect how you work with them or don't? Ellis Hamburger Yeah, it's funny. Working with all the B2B SaaS companies that tend to be a bit more objective and money-driven has definitely made me B2B sassier of a person. I would say just thought of that on the spot. I could tell as a reporter, man, when you get an email and it's the fresh Wharton grad MBA, who's trying to work something out, but VCs see through it too. And so I think I've been fortunate to be in a space and get referrals from people who care about design. I mean, people ask me how I've gotten this or that great client. And it's like, all the people who care know each other. And I'm someone who cares. I want to talk about how product decisions affect people, make them feel better or worse about their lives. And frankly, I mean, this is kind of the beef that I've had with Facebook all this time, is that it's a whole lot of very nice, smart people who just refuse to really have a take on what a better Society looks like. You know, for a very long time, their MO was about connecting the world. And one of my favorite artists and creators, Hideo Kojima, who has a new game out called Death Stranding 2, he is known for being very prophetic with his tech ideas as predicting COVID And delivery culture and whatnot. The tagline of his new game is, should we have connected? And I mean, I don't know if that's a direct aim at Facebook or just connecting as an idea, but I think Facebook certainly was flying that flag for a very long time. As far as I can tell, the only thing they care about is giving people more of what they click and what they engage with, which leads to a very different Instagram explore page than if you Have a different set of principles. And there is no doubt that if you have ethics or morals or opinions, they isolate one audience after another. But I'm sure you would agree it's, you know, it's pretty impossible to be apolitical these days. And, you know, I hope they can start to have some ideas about what makes ([Time 0:30:50](https://share.snipd.com/snip/8af4c19c-9138-4bb4-b8f8-dc6ed257d485)) - **Cluely and Honest AI Branding** - Cluely’s branding as a tool to "cheat on life" speaks honestly to real AI user behavior. - More companies should openly recognize how AI is often used for shortcuts and cheating. Transcript: Alex Heath It's Cluely, which if people have not seen it, it's only a few months old at this point. But it started with this 21-year founder who built this AI tool to help him cheat on a coding interview for Amazon. He posted it online. It went super viral. Amazon actually rejected or rescinded its job offer. And then he turned it into a company. And it's a company that's also this creator thing. It's more than just an AI software thing. It's like a creator house. They're all super young and they're all just like super AI native. And essentially what it is, is it's this kind of semi-translucent interface that runs in the background on your Mac and listens and observes everything that you're doing and hides Itself from screen sharing and can recommend things based on the context. And that sounds kind of cool. I haven't installed it. I'm frankly terrified to install it, but I probably still will at some point. But they could say it's this new kind of assistant, right? Which is like, sure, everyone's doing an assistant. And instead their whole branding is like, this is to help you cheat on life. This is to help you cheat on everything. And like, it may not always work, but like that framing to me was really interesting because it actually just spoke to how people are quietly actually using this stuff. Like, I wouldn't be surprised if you could objectively look at how people are interacting with AI at scale. Most of it is probably developing what look like unhealthy parasocial relationships for emotional support or replacing relationships. And then stuff like this, like cheating, cheating on homework, cheating a job, stuff, whatever. And I'm curious, like when you see something like Cluey, and then they raise millions of dollars from Andreessen Horowitz, of course, like right after all this. What does that say? What does that say to you? And like, does that change anything about how you approach your other work? Ellis Hamburger There's a side of me, you know, that comes back to being a reporter that is kind of like, you do know that the more sensational headline is probably going to get more clicks or do better. And I remember being a business insider and just looking at Chartbeat and looking at the clicks you get on your headlines and it's addictive, right? And I think we all need to hold ourselves to a standard of trying to balance that cynical part of ourselves with the more honest, noble, altruistic side of ourselves. And I definitely see them more toward the cynical side of the spectrum. But I maybe have a contrarian take here, which is that I kind of think it's pretty clearly a tongue-in thing. It is certainly going to press liberal buttons, I will tell you that. Certainly going to press the buttons of the people who like to opine about this stuff in the New Yorker or the Atlantic or maybe even at The Verge. But there was a specific line of their manifesto that struck with me, which is, the world will call it cheating, but so is the calculator. And this is something Sam Altman said as well. So was Spellcheck, so was Google. What's funny about that is that they're saying it's cheating, but that it isn't going to be cheating for long. And so to that extent, I think it's kind of an interesting hook. And I would honestly much rather have companies be straight about where they think we are heading. And here's the thing. Here's my challenge to every company is to know what's ([Time 0:36:40](https://share.snipd.com/snip/2f3bb526-86ae-42d3-9a29-c08a3d46e7e9)) - **Power of Visionary Storytelling** - Craft manifestos that paint a compelling vision bigger than the company itself. - A clear vision helps people connect deeply with a product or company’s purpose. Transcript: Ellis Hamburger Every manifesto I write for a company, which is, a big part of my bread and butter is to try and paint a very desirable, but also specific vision of the future that is bigger than yourself. I think if it's only about you, it just feels like marketing, but if you're a robotics company, what vision do you literally see for our daily lives? I think if there's one thing we've seen with all the technology that we have, it's not that I am appreciative of my dishwasher and have another hour of leisure time each day. It just inevitably gets filled with something else, in part because of our culture and the guilt we all feel if we're not always being productive. And so that is really the value of visionary storytelling, I think, whether it's in tech or ([Time 0:42:20](https://share.snipd.com/snip/453841f9-cf55-4a9c-8d56-58b93a07f93a)) - **Tech's Impact on Leisure Time** - Historically, tech hasn’t increased leisure time as gains get filled with new work. - Raising people’s income has greater impact on well-being than automating small tasks. Transcript: Ellis Hamburger The light on what can be better and how to get there. Alex Heath Yeah, there's this thing with AI where I feel like everyone is saying it's going to free you up to do more. It's going to give you your time back, right? That's agents, right? It's like I was just at a podcast, a live podcast taping where Sam Altman was talking about, like, I can't wait to wake up one day and have my AI have triaged my entire day for me, pre-written Emails, like all that stuff. And just me be able to go, yes, no, yes, no, yes, no. And like, that's my morning. And like, sure, that sounds good on its face. What I feel when I'm using more AI stuff is that I just, it makes me busier. It doesn't actually like give me free time. It just makes me go like, oh, if I now have something listening to all my notes and all my calls and making nice summaries for me, like I'm just going to have more meetings and I'm just going To do more because I can. Because I know I've got this fallback. I have this crutch where all my meetings are there. And I just don't know what that does to us. Like, I feel like it probably is going to lead to more burnout, not less. I'm curious if you agree or how you think about that. Ellis Hamburger Yeah, I bet Neil, I would have a good answer to this question. Is there any evidence through the different technological revolutions that the leisure time has ever panned out? Alex Heath No. Ellis Hamburger Like, I was asking ChatGPT about it earlier today, actually. It was coming up a little empty. And I think you're right. And I mean, this isn't a surprise, but I was watching the new Johnny Harris YouTube. Not really changed all that much in decades. And more and more of what we create, the profits, the earnings are going toward the top 10 to 1%. This is the thing with tech, right? Is that we all think we're doing God's work when it's like, man, you could really have a far greater impact on people's quality of life, not by giving them a meeting no taker, but by trying To affect change so that people have more money in their pockets. Like that is a far bigger lever for living the good life than trying to automate all these little things, I would say. Alex Heath I think people, ([Time 0:43:10](https://share.snipd.com/snip/6ea84a73-db0d-4d70-8bcf-74b662fa94c0)) - **Two Founder Types in AI** - AI founders vary widely in motivations: some idealistic, others indifferent to short-term harm. - The metaphor of Batman vs. Ra's al Ghul captures differing visions for AI’s impact. Transcript: Alex Heath Headed that are not good for most people. Ellis Hamburger So can we do a tier ranking list, but only two tiers? We'll categorize all the founders, whether they are Batman or Ra's al Ghul, the head of the League of Shadows. They're both trying to do the right thing, but very different methods, right? Alex Heath You can do that. Ellis Hamburger Yeah. I mean, I think, you know, if you look at a lot of the libertarian VCs and all that, I mean, they are so in the clouds that they don't really think much about the collateral damage in the short Term, which I don't know, maybe there's an argument for if this is going to create a far better world in the long term. But yeah, certainly Raz El Ghul took it too far. Alex Heath Well, Ellis, this has been great. I feel like we could end it ([Time 0:49:20](https://share.snipd.com/snip/42b2ccad-4f20-4aeb-afa6-bcc9c6538116))