Will AI Replace SaaS? The Future of Software

Foreign.

Welcome back to the Next Wave podcast. I'm Matt Wolf. I'm here with Nathan Lands. And today we're going to talk about the crossover of software and AI and we're bringing on a guest expert, Dan Shipper, to talk all about it. Dan is the CEO of Every. He has all sorts of companies underneath him that he's incubated, he's an investor, he works with Sequoia, he has all sorts of knowledge in the software space, and he's going to break down his, his exact process for deciding what products to build and what tools he used to build them and exactly how he brings them to market. It's a great discussion. So let's just go ahead and dive right in with Dan shipper.

Using only 20% of your business data is like dating someone who only texts emojis. First of all, that's annoying and second, you're missing a lot of context. But that's how most businesses operate today, using only 20% of their data. Unless you have HubSpot, where all the emails, call logs and chat messages turn into insights to grow your business. Because all that data makes all the difference. Learn more@HubSpot.com.

Thanks so much for joining us today, Dan. How are you doing?

I'm doing great. Thanks so much for having me.

Yeah, so I'm excited to really dig in with you and what you're doing over at your various companies that you're working on. Let's go ahead and like lay the groundwork. Like tell us a little bit about your company and what you're building and what is every.

Yeah, that's a good question. We've been around for five years, so it's evolved significantly over time. The way that I talk about what we do right now is I call us a multimodal media company. So we publish writing, we publish podcasts, we publish video, and we publish software. We do all that using AI and we've sort of gone in that direction because I think that there's this really interesting thing happening where the line between being a writer and being a builder is starting to like blur and they're sort of moving into this like overlapping venn diagram because now that you can write in English and turn it into code, like writing is building. And also because software is so cheap to make, I think that software is becoming content. And so as a media company, like what we focus on at every is we have a daily newser every day we write a long form essay about what comes next in tech. We've got like, about 10 regular writers who do that. I write every Friday column called Chain of Thought on the sort of intersection between AI, creativity, psychology, philosophy, all that kind of stuff. And, you know, our job is to sort of explore the future. And the best way to do that, in my opinion, is always to be using new tools and experimenting with things and building things and all that kind of stuff. And what we found is that when we experiment with something, if we release it to the audience, people love it. It makes the writing better. And sometimes those experiments become like real products and should be real products. And so what we've done is then made them into, like, real, actual products that we often start by using them internally, and then we'll publish them to the audience, and then we bundle them into the subscription. So you pay one price and you get access to all of our content and all of the software that we make. And we've done four incubations so far. The most recent one, which we just launched about three or four weeks ago right before Christmas, is called Quora. And it manages your inbox with AI. So basically what happens is as emails come in, it decides if you need to respond to them, if you do need to respond to them, if they're from a human, it makes it to your inbox and you see it immediately. And if not, it automatically archives it. And then twice a day you get a beautiful brief that summarizes all the things you need to know about what's going on in your inbox, but you don't need to actually do anything with. And it just saves like a ton of time because, like, 80 to 90% of your emails are like that. And we have a bunch of other products that are. That are sort of in that vein for different parts of your work life. And yeah, that's the basics of every.

And that was cora. Like, C O R A.

C O R A. Yeah, C. Okay, cool. Yeah, it's been going really well. We've got like 10,000 people on the waitlist. We're taking off, like 20 people a day and just sort of starting to ramp it up.

I'm on the wait list, so, you know, I can't really talk about it. I haven't tried it.

I'll get you off the waitlist. You should have told me. I know a guy.

Cool. That's fascinating thing about, like, writing is becoming software. Is that what you said?

Writing is building and software is content?

Yeah. Okay. I've been noticing that too, because it's part of this. Doing this podcast I've been trying to get more hands on with AI. Like actually started building some stuff. So recently I started using all the different AI tools to kind of develop a game prototype just for fun. Maybe it becomes a real thing, I don't know. But it is fascinating that I am now talking with AI and kind of writing my game design document with AI, using my voice, you know, using stuff like whisper flow, chatting with it, having it update the document, and then I feed that document back into Zero1Pro when I'm coding on it. And it's just amazing that, yeah, that conversation I'm having with AI then informs the actual code and software that is created that's never existed before.

Isn't that so cool? And like, so you can code while you're walking around, just like, you know, record into your voice note like Kieran who he is our GM of Core and he built Core. Like that's how he started the first Quora mvp. He built that in a single night and he did it by going to take a walk and he just like recorded himself being like, this is how I want the app to work. And he threw it into 01 and just like, it's awesome. I think, like, it's so cool. There is a big difference between you can build a prototype and then actual like production grade software. But even the production grade software, it's just changed what you can do and how fast you can do do it. And it's really, really fun.

Yeah, I really like what you said there. So writing is building. That makes sense to me, right? Like everything is kind of becoming prompting to get what you want, right? If you want to write software, you prompt it. The software can come into existence. I mean, even now, like we're talking about, you can even speak it into existence and then it actually becomes software. Can you speak more to the software is content? What do you mean by that?

Yeah, so software is much cheaper to make now. And one of the things we found is that if we release a little experiment as a piece of software, like all of the experiments we run are much better at attracting audience and emails than like any given essay or YouTube video or whatever, because it's this totally new format of thing, right? And so like in the same way that, you know, on Instagram you can take pictures and TikTok you can take videos. On YouTube you can take videos and those are different from like Hollywood films, but they're still videos. I think that we've been living in the world of like, software is sort of like a Hollywood movie. Like you need a lot of Money and time to make it. And now, like, with all these tools, it's expanding beyond people who are professional developers. And anyone can make a little prototype. And that switch is sort of like, anyone can kind of like write a blog post. Anyone can, like, you know, do a tweet. It's going to be the same thing. There aren't really platforms yet for this, but I think there's going to be like a YouTube of software development, maybe like replit. Is that to some degree. And it's not for professional coders, it's for like people who just grew up prompting. And that's happening right now.

That definitely makes a lot of sense. And I also feel like this is something we've actually talked about on the show in the past too, is like, I almost feel like software could be the sort of next SEO, right? Because I feel like just writing blog posts and writing content is so easy now. Like, anybody can do it because of tools like Claude and chatgpt and Gemini and everything that's available out there. I think the value of sort of written content is starting to go down. And when it comes to SEO, one of the goals of SEO is to bring somebody to your site and keep them on your site as long as possible. It almost feels to me like really good software that people find value in will be the new thing that keeps them around longer on your platform and help your site rank better and help you actually grow your business a little more. I'm curious if you have the same thoughts or if you would disagree with that.

That's really interesting. For one, like, I love words and I think the written word is always going to be super, super valuable. I do think there's something to what you' saying, which is that to some degree, a lot of the content that gets made right now, like, I just saw this thing about HubSpot's like, traffic, and HubSpot's traffic is just like going down. And so, like, a lot of the SEO type content that's made right now is probably better served by a chatgpt. It's like, okay, like, how do I ask for a raise? You know, it's like there's zillions of those SEO articles and like, that's much better from ChatGPT, especially, like, the more generic stuff. Right? I still think that, like, for every, for example, like the audience that we have wants to hear from me and wants to hear from other writers that they know and are not looking for like a generic fact or a generic process. And that's a sort of different thing. That's a definitely smaller audience, but it's still like, I think, incredibly important and valuable. I think on the tools front, you may actually see some of the same commoditization because you, you don't need to go use a SaaS tool when, like, artifacts just like, spins up a little UI for you and you can just like, do, do, do whatever. So that may also happen with software to some degree. And then I think you also have to, like, wonder, okay, if software tools are the new SEO, like, what is the thing that is, like, pushing people to those tools? Is it Google still? Is it ChatGPT? And if it's ChatGPT, like, how and why is ChatGPT like, letting those things surface and rank? Right. But I do think that providing people with like, rich, let's say, multimodal experiences, that's not just text, it's not just video, it's maybe it's tools or whatever. Like, that is, I think, going to be a really legitimate way to, like, build an audience and build a following.

Yeah, I totally agree too, that when it comes to, like, the world of content, right. I think as AI becomes, you know, more and more prolific and everybody's using AI to, you know, try to gain the system through SEO and things like that, I think it becomes more and more important to have your own voice and to, you know, actually be a voice that people can trust and, you know, actually put your face out there. Right. There's a lot of stuff you're seeing on YouTube now of, like, faceless video channels, and people are claiming it's all the rage right now to make faceless video channels. But in my mind, if that's the direction everybody's going, the people that actually put their face on YouTube are the ones that get a rise and, you know, stand out.

The sacrifices we make for fame, you know, just us.

You're in San Francisco, right, Dan?

No, I'm in New York.

Oh.

I go there often, but I live in New York.

I lived in San Francisco for about 10 years. I'm out in Japan now, and that's one thing. But nice being here is, you know, even if the podcast blows up in the future, I don't have to worry about fame. Like, nobody will notice what I am. That's why I wear this, like, media mogul hat that HubSpot gave me. I think it's hilarious because obviously I'm not a media mogul, but it's kind of funny in Japan, you know, I don't know if, you know, this is a HubSpot owned podcast. I'm not sure if you're aware of that. And you were talking about HubSpot's traffic.

I am aware because the emails are from HubSpot. Oh, shoot.

Yeah, yeah, yeah.

I did not put that together.

That's totally fine. That's totally fine. I mean, that is one reason to do this podcast right now. This is even. I think kind of what's part of our pitch too is like, you know, having faces. It's going to be very important in the future versus just generic content on, you know, text content that anyone can throw out. And AI is going to get better and better at that. So SEO is going to be harder in the future. But I do wonder, why would chatbots send you to other tools in the future if it gets so easy that they could just, like, create it for you on the fly with, like, what we're seeing right now? Like, I've been using O1 Pro a lot and like, yeah, seeing how good it is at software, I can't imagine in another two or three years that any idea you have AI is not just like, cool, here it is, here it is right now in the window. And just use it.

Totally.

It's interesting because when it comes to the whole HubSpot conversation, I do think maybe HubSpot's traffic is going down. But podcasts like this and what they're doing in other media spaces, I think is their sort of answer to that. Right. You've got podcasts like the Next Wave and My First Million and Marketing against the Grain and the Hustle, and they're starting to do all their newsletters now with the Hustle Daily and. And all that kind of stuff. So it's like, maybe less people are going to the HubSpot website, but HubSpot's becoming like this larger media brand.

Totally.

So I don't think there's anything wrong with saying that, you know, HubSpot's traffic is dropping because they obviously have, like, you know, backup plans that they're building as well. But just wanted to throw that out there.

I feel like I just, like, went to someone's dinner party and like, spit in soup or something like that.

It's okay.

It's okay.

Like, I knew in one part of my brain that this is a HubSpot podcast and then. But anyway, yeah, I'm not trying to imply that HubSpot is going to die or anything like that. It's just the whole SEO game is, like, changing rapidly. Right. And yeah, it seems like shows like this are a really good bulwark against those kinds of changes which have been happening for a long time. It's not just AI. And I think there is that interesting dynamic that, you know, you can see OpenAI trying to react to or trying to head off, which is like, you know, they do deals with publishers and they cite their sources and all that kind of stuff. And I do think that they would like to keep a vibrant ecosystem and app partners and all that kind of stuff. And I also think you can see with Google's example that like over time the pressure is to keep people on the platform longer and add more ads or like do the like instant answers or whatever. And I think AI, broad use case, general purpose AI, consumer apps are going to have that same pressure. Maybe it's not now, but in like 20 years, like when Sam Altman is like, you know, on his island somewhere and like there's another CEO, that's what they're going to have to do. I mean, maybe we'll just be living in a post capitalist utopia because we'll have AGI or whatever, but like assuming that that's not the case, that is a thing and that is a pressure and I think that's a really interesting thing. But also like all of these new tools open up new opportunities. So like, it seems likely to me that maybe it spins up a whole UI for you for like a specific thing. But like the kind of UI that it makes and the way that it does the UI like we've been talking about is a content format, is a perspective. And it would be surprising to me if they didn't allow brands or businesses or whatever to like make available those tools or those powers within their app in some way. Sort of like an app store type thing. And they've done that like repeatedly. They've tried to do app store type things for a long time. It hasn't quite worked yet, but it'll probably work eventually. So it's not like the death knell of like all independent tools or media or whatever. It's just like it's changing and there'll be new opportunities in the old ones. You have to like move on to the new ones.

Yeah, totally. Would you be open to like sharing a little bit of like your workflow for actually building new software? Like how do you validate an idea and then what sort of tools do you use to actually build the idea? And yeah, you know, can you walk us through some of that?

Totally. I think the really cool thing about the current AI landscape is it has completely Reset the table for what tools are possible. And so that means that there's a lot, a lot, a lot of low hanging fruit because we're all just figuring out our workflows, right? It's all new. And that wasn't the case two years ago because it was like the apex of the B2B SaaS. Boom. And every low hanging fruit had been picked, right?

Yeah. There's a clear way you do things. It was all, it's all been figured out now.

It's totally different. So what I do and like sort of the philosophy is first I use these tools every day. And by these tools I mean like the more general purpose tools like Claude and ChatGPT and cursor and Windsurf and like all that kind of stuff. And what ends up happening is you end up finding workflows that start to work pretty well for you. But once you find a workflow and a chatgpt type tool, it's kind of a pain to just do it over and over and over again. It's not built for that. It's built for this like general purpose thing. Right. And that is always an opportunity to me. So, like, a really good example of that is we have this app called Spiral. And Spiral helps you automate a lot of repetitive creative work. You know, for me, it's like doing my podcast descriptions or my podcast Tweets or my YouTube titles or like that kind of stuff where it's like, I have a format, I know basically what it should be, and it's like really important that I get it right. But it's also like fairly repetitive. And what I realized at some point, I guess in like May, that Claude was like really good at that. Like, if I gave it a few shot prompt with some examples and I was like, okay, I want you to take a podcast transcript and turn it into a tweet. And like, here's examples of me doing that in the past. Here's a new transcript. Can you like do a tweet? It would like get me 80% there. And then I would just have to like do some modifications. But like, that prompt was like really hard to do all the time. Like, I didn't want to have to redo that all the time. And if I didn't want to have to do it, like, I just knew that my team wasn't going to do it. And so I just spent like two days basically making an app called Spiral where you can like, you make a spiral, you say, like, I want to turn podcast transcripts into tweets you feed it some examples and then you have like a shareable widget that you can, any time thereafter, like generate more and more and more tweets without having to like reprompt anything. And that's going really well. We have a full time GM for that on that product. His name's Danny. He's really talented. We have almost 10,000 users for it and it's growing really well. And that's an example of like, you find this like use case in the general purpose tool and then you can like unbundle it into like its own app. And the way that I do that is it's sort of the way Nathan, you mentioned earlier is like I use Cursor or Windsurf. I kind of go back and forth between the two. They have these like really, really good agentic building environments now where you just type in a chat and it's like it just goes off and does stuff and then it's like it's made like five files and it's like, I have no idea how it all works, but it's fine.

Sometimes it's fine.

Quick question. Do you actually know how to code yourself or are you letting AI write all the code? Okay, you do know how to code.

Okay, I do know how to code, but I am letting your AI write all the code because I'm very lazy.

Yeah, yeah, but, but at least you're able to do a little bit of extra troubleshooting if the AI doesn't figure it all out for you. Yeah, okay.

And like, you know what's really nice is like O1 and O1 Pro. Sometimes these tools don't have tool use. And all the agent stuff uses tools. Honestly, when 01 or maybe 03 has tool use, which I think is coming very, very soon, it's going to completely change everything again. But I'll often, like, I'll start in 01 or 01 Pro. I'll say make the plan. I paste the plan into cursor. Cursor will do stuff. Then it will get stuck and have a bug. And then I'll just paste it back into 01 and then paste it back. And then every once in a while I'm like, I'm bored of copy pasting. Like, why don't I figure out how this actually works? But also for me, like, to be really clear, like, my job right now is I'm mostly prototyping new things and writing about them. And then sometimes they become products. But once they're actual products, we have actual engineers who like actually are, like, in the code a lot. And so I could be, you know, reading the code or whatever, but I'm not doing that right now because it's not really my job. Yeah.

So I want to ask this to you because there's been a lot of discussion around, like, SaaS companies having no moat. Right. I'm sure you've heard that a million times over the last couple years. What are your thoughts on that? So let me preface this with saying, like, I personally, like, if I have a problem I need to solve, that's like a fairly simple problem I've stopped going out and looking to find. If there's a software product for it, I will literally open up Cursor and tell Cursor what I want it to make and it'll write a little Python script for me and it solves that recurring problem for me. I know right now I'm probably, like, abnormal in doing that. I don't think most people sort of have that kind of workflow, but I'm wondering if it's going to get to a point where when somebody has a problem they need to solve, instead of turning to SaaS products and paying, you know, some sort of recurring subscription, they're just going to go to one of the AI chat bots that are out there, maybe it's ChatGPT or Claude or one of these tools, and just say, this is a problem I'm trying to solve. And it says, okay, here's a little script I wrote that solves the problem.

Yeah.

So I wonder what the future of SaaS looks like.

I mean, I think you're definitely right in the sense that there are certain little things. There are maybe like a thousand little things that come up every day that, like, maybe I would have, like, looked for a tool for before, or probably I would have just, like, suffered through, like, doing them manually or whatever that. Now ChatGPT just does it for me, and it's amazing. But I think there's also a big difference between that and, like, a fully featured, like, SaaS app, like, for a power workflow, which I just think that things that are more specialized for those kinds of users are going to win. Even when you have AI, even if you theoretically have AGI, and that there are room for those kinds of tools, whether they look like traditional B2B SaaS or whether they look a little bit more like an AI native version of that, whatever that looks like, I think it's going to be more like that. And I think there's also a really interesting question about like how much SaaS is actually the right business model. What I've been seeing a lot more of, and I think is probably a little bit more of a AI native business model is paying per completed task or paying for the value that you get basically because when you have AI, you can pay for the, the job to be done versus just like the time or by the month. And I think that's really interesting because SaaS is such a much better business model that like, if you are currently a SaaS company, you'd probably rather stay a SaaS company. And that means that there's a lot of opportunities for startups that don't necessarily need to keep that model.

Yeah, performance based SaaS would be incredible. Like, yeah, that's, that's an amazing idea.

Yeah, I think that's going to happen. I mean you see it already in certain places, like the LLM APIs are all pay per use or whatever, but I think that'll filter up through the stack. I feel like I'm your worst guest ever. I'm like, SEO is dying and SaaS is dead.

I'm so sorry.

No, no, I mean we've got other tools that I think are cool. We have a pool called Sparkle, which is a file organizer. So basically it just sits on your desktop and just automatically organizes your desktop, your downloads and your documents with AI, which is like really cool and very useful.

Everything a mess for me. So like, but now it's not. And I didn't have to do anything that's am.

I'm always like trying to think about like how to organize my files. I'm like, oh, one day I'm going to organize them all they're going to be. And I was like, wait a minute, AI is going to eventually do that for me. So it's like, it's awesome.

Yeah, it's just a matter of time. Another thing that we've been working on is actually one of our writers, Michael Taylor has been working on and we're, we're kind of doing it with him is AI is quite good at acting like a particular kind of person, like simulating a particular kind of person. And Michael built this tool called Rally that we've been using internally. I think we're going to make a, like every column out of it and then maybe launch it as a product that's still tbd. And it really depends on if Michael wants to do it with us, but I think it'll be really cool. And basically what it does is like, for example, if we're coming up with headlines for a podcast. What we can do is we spin up an audience of like a thousand every people or a thousand hacker news people. And then we can be like, okay, here's my three titles. Which ones do you like? And it will just like, survey the crowd and then tell us what the crowd of LLMs thinks. And it's really cool. It works.

I thought you were going the opposite side. So actually I thought it was gonna be like, the person who's helping you create the taglines or whatever. You fed it a bunch of books from some famous person who's an expert at that, but you're doing the opposite. Like the audience side.

Yeah.

That's incredible.

Exactly. So it's sort of like wisdom of the crowds, like, ask the audience type stuff, which, like, now you just have to try stuff where you have to do AB testing and it takes a long time. You need a big audience or whatever. And if you use this, you don't have to. And it's still very early and it remains to be seen, but, like, I feel quite confident that the results will align to some degree with, like, what you would actually find out in the world. And that's really cool.

Wow.

So I actually do have one other sort of little rabbit hole I do want to go down with you. So I believe you do some investing outside of every as well. Right. So you have some investments. I believe you work with Sequoia Capital as well. Is that correct? What are you looking for in investments? I'm assuming you're probably focused more on AI. I could be totally wrong. I'm just sort of guessing. But what are you looking for in investments? Like, how do you find something that stands out to you?

The investments that I've done usually are places where I've had a particular thesis I've been writing about and thinking about, or I just know the founder and I know they're really awesome, and I just, like, want to get into whatever they do. Yeah, I do invest in AI. I have found, honestly, that a lot of AI companies are just raising at such gigantic, gigantic valuations that, like, I haven't actually pulled the trigger on that many deals, even if I really like the person or really like the thing they're building, because it's just like 40, $50 million seed round is just, like, crazy to me. But, like, generally it's like, is it in a thesis, in a space that I care about, or, like an idea that I care about? And I have a bunch of articles that sort of lay out all that stuff and I can give you a little preview about the things I care about right now. Like, for example, one is the future of science. Like how AI might change how we do science is a big one for me. Another one is, yeah, just do I really like the founder? And then the last one is just, do I use the product? Do I like it? I think the proof's in the pudding a lot in those situations. And you can tell immediately when someone's built something awesome because you just feel it.

Yeah, yeah. So when you say, like you have a thesis on it, you're kind of saying like, this is something I wish was in the world, but then you actually find something that does that, that's something that you'd want to invest in.

Yeah, exactly.

Very cool.

Or I assume some way the world's changing and then how a product would fit into that changing world.

Yeah, I love that stuff. I especially get really excited about a lot of the science stuff. Like, I was really nerding out about like that matter gen that just came out from Microsoft the other day. I think that's something that's really cool and really exciting. And I think there's a lot more stuff that's going to pop up sort of in that world, you know, this year. I think we're going to see a lot of sort of materials science sort of things come out of AI this year. And so that, to me is super, super exciting. But on that note, I really, really appreciate you hanging out with us today. Where should people go after listening to this episode? Is there a specific social media account you want to shout out your website? Let us know where to check you out.

Thanks so much for having me. You can find me on X at danshipper S H, I P P E R or at every also on X. And then the website is every to. You can find our newsletter, you can find all of our products, the courses we offer, all that kind of stuff. And it's been awesome chatting with you guys.

Yeah, it's been great.

Awesome. Yeah. Thanks so much, Dan. Really, really appreciate you. And thanks everybody to listening to this episode. If you haven't already, make sure you subscribe wherever you listen to podcasts. We're also available on YouTube if you want to see our faces actually talk to each other. And thanks so much for tuning in. We'll see you in the next.