AI Hot Takes and Unusual Twitter Fundraising Strategies with Dan Siroker (Rewind AI)

Todd: Dan, welcome to the show.

Dan: Thanks for having me.

Todd: Absolutely. Um, So since you launched Rewind late last year, I think the company has seen a ton of early success. You've raised over $33 million, I believe at this point from great folks, Andreessen Horowitz, NEA, Sam Altman, and of course us here at First Round. And rewind to me is such an interesting product because its fundamental aim is to improve human memory.

And so today I'd love to zoom in on the details, of how Rewind came to be, um, to how you're thinking about go to market and building a product that your customers love. And so to kick things off, Dan, I'd love if you could rewind to maybe your mid 20s or so when I think you started to experience hearing loss for the first time, now which is which I know is a big part of rewinds kind of origin story.

Dan: Yeah well, I, I started to lose my hearing in my 20s. I have a genetic condition that caused that to happen. And at the time I was running my last startup, Optimizely. And I tried a hearing aid and it was magical, to lose a sense and gain it back again felt like gaining a superpower. And ever since that moment, I have been on a hunt for ways that technology can augment human capabilities and give us superpowers.

You know, the magic of losing your hearing and gaining it back again, wasn't just the fact that I could hear, which was awesome. It was realizing how bad my hearing had gotten over the many years before that. And so I felt there's a way to take that insight, that sort of secret into the world and apply it to other parts of our human capabilities.

That led me to memory. Memory is one of those things very similar to hearing that gradually gets worse as we get older. Most studies show that we forget about 90 percent of what happens after just one week. And so the inspiration for the company came out of this question. If there's a hearing aid for hearing and glasses for vision, what's the equivalent for memory?

And that's what started us on our journey.

Todd: Amazing. so how does Rewind work exactly? And what is kind of unique about the approach that you are taking to an AI driven product? Maybe versus some of the other products out there.

Dan: Yeah. Rewind is a truly personalized AI. So it captures everything you see, say, and hear. Stores it all locally on your devices. Today, we support Mac and iPhone and soon later this year, Windows. And by capturing everything you see, say, and hear, we're able to give you these amazing capabilities like you can ask rewind to write you an email to somebody, you know, asking to catch up and it uses the context of everything, all your experiences, interactions with them, to create, essentially a personalized, experience, that works great when you're integrating with large things models where you can ask it to do tasks for you, or it can be as simple as just rewinding time.

That's why the name of the company is rewind is that we let you rewind back and forth in your day, uh, some people describe this as opening an entirely new dimension of computing time. You know, If you think about it today, the way information is stored on your machine, it's bits that are stored into files, which is this weird kind of skeuomorphic concept from before computers.

But reality is, you know, we're constantly bombarded with new information. We're context switching more than ever. We've got information in lots of different silos, and we really want to make it so that you can capture all of that data, make it searchable, make it useful, and that's what Rewind does.

Todd: So you mentioned this interesting stat that 90 percent of memories are forgotten within a week. What are the other ways that people solve this problem? And what makes Rewind kind of a unique solution here?

Dan: You know, that's one way to think about it. If we all had perfect memory, what would be different about the world? Well, one thing would be different is that we wouldn't be chopping down trees, turning them into little rectangles called paper and taking a stick with ink at the end of it and writing things down.

You know that just if you think about it in that way, it seems ridiculous that the best way or most common way that we've thought of to remember is to write things down both physically on paper and digitally. Um, you know, If we didn't, if we had perfect memory, we wouldn't need to do that. So, you know, That is probably the most common alternative.

People writing things down, they take notes. What we found is that people start using Rewind and they get this amazing peace of mind knowing that anything they've experienced, they can go back to. It's mind insurance. You can go back to these things that you might want and for people who are knowledge workers, they're in lots of meetings back to back, they're, you're consuming lots of information, you never know what you're going to need later. That's the biggest downside to taking notes, is like, you need the foresight to know, what is it the thing I might want later? And, that's one of the things they get with Rewind, is they don't have to have that foresight. They just passively capture everything in a privacy sensitive way, it's stored locally, we don't have access to it, only you have access to it.

And if you need it, only if you need it, you can go back to it later.

Todd: So when you were building the very first iteration of rewind, what did that look like and what were kind of the most difficult parts to get right?

Dan: Before Rewind, we actually had another product that was in the same space trying to sort of solve this vision of giving humans superpowers and giving them perfect memory. It was called Scribe. It was a meeting bot. You know, we'd started this March 2020 and you know, the pandemic was happening and Zoom was taking off and we had this, you know, what we thought was a brilliant insight.

Ah, let's take Zoom meetings, the thing that we're all doing now, and use transcription, which has just become really good, and combine the two and offer a product. And we launched it, we were very excited about the uptake, and then it turned out every other day some new startup would pop up doing basically the same thing.

It's a very crowded space, lots of tools out there. And from that, we had this insight that if we could do what we were doing for Scribe, but all of the information you consume, not just zoom meetings, that would be really magical because often the thing that you're looking for when you're in this mode of racking your brain, trying to remember something it's usually trying to remember something that happened in a different channel or a different silo than you're thinking of.

You're searching your email vendor. If you've ever had this experience searching your email frantically, like. what was that note? What was that thing that Dan said or whatever? And it turns out it was a Slack message or a text message or you saw a Tweet. We're constantly consuming this information.

So something that could capture everything. And do it in a way that respects your privacy was really the insight that started Rewind. And it so, you know, co-occurred with Apple Silicon becoming really powerful and enabling this before Apple Silicon, you know, Rewind couldn't exist. So that was also very lucky in terms of the timing of the technology, the enabling technology to make this possible.

Todd: Yeah. And I mean, in doing what you just described and doing it well, like at the level that users really demand, I mean, that's a huge product right there. That's huge scope. So did you think about these are the things that we really need in V1, these are the things that we're going to wait to do later?

Dan: Yeah I'm, I'm a big believer in cutting scope and starting with something, you know, simple, easy to understand and really being ruthless in that, you know, I, I, I think there's this classic trade off in any engineering effort, which is between scope, quality and time. And I'd rather, you know, hold time.

Yeah, I'd rather hold time and quality as the kind of the goals and then cut scope ruthlessly to, uh, to achieve what we want.

So we started with, you know, some very basic goals. The first was to be imperceptible. Which is usually not a goal that people come up with when uh, when they build a product, but we want it to be so good that we could run in the background.

And you don't even notice that Rewind is running. And so that is, you know, it's, it's a hard problem to solve because what we're doing is kind of a sophisticated, we capture everything your screen is showing. We compress it, you know, 3000 times smaller so it stores, it can be stored locally. We do optical character recognition on every word that appears on your screen, and now, you know, we're doing transcription as well.

So, you know, to do all that in a way that doesn't hog system resources doesn't make your laptop overheat, was really goal number one, how to be imperceptible. And then everything we built on top of that was kind of a balloon squeeze. You know, how do we add a feature without compromising that goal of imperceptibility?

Um, and, and that's been the journey we've been on over the last uh, you know, few years is how do we keep adding to it and building more and more superpowers for people without being this sort of tax that, that feels like it's, it's hogging system resources.

Todd: So when was the first kind of public launch of Rewind? I remember it being a good launch with like with a really compelling demo and people were pretty interested from the very beginning.

Dan: yeah, we announced November 1st of last year, actually, it's crazy to think. We've been working on it sort of quietly um, from basically the December before. So just short, shy of a year, we've been working on it, really validating some of the key things. Like I said, the compression was critical. The imperceptibility was critical.

Um, and, and the quality was, it was really important. And so we launched, we announced uh, it totally blew us away. You know, We had been working on Scribe, which was doing well, it was growing, you know, it wasn't an obvious decision, by the way, to pivot from Scribe to Rewind, we were still, you know, internally we had actually written down five different strategies of which one was to persevere or Scribe.

One was this idea that we ended up calling Rewind, which wasn't, you know, even on the team, three out of my team members thought this was the worst of the five strategies to pursue. So it wasn't an obviously good idea. 

Todd: How did, so how did you decide though? How did you decide? Which direction do we go? That's a tough decision with five people. 

Dan: Yeah, it was definitely a tough decision. And at the end of the day, the main, factor was just my intuition. you know, What would I regret if we, if we persevered and, and failed. For me, the vision of giving humans perfect memory and beyond that, giving them superpowers, this felt very much the most in line with that.

And the rest were maybe better businesses or maybe less risky, but weren't as connected to the vision of what we're trying to do, um, you know, we were also very analytical and data driven. My last company was all about A/B testing. So, you know, I asked all our employees. I asked our investors. I surveyed them.

I asked them to stack rank them. We wrote up a whole notion doc with all of the trade offs and all of the ideas. And I actually think all five strategies could have been successful. You know, It's not like this is the one successful path to, you know, to building a product that works.

It's just, this is the one that I felt like, you know, given the balance of trade offs, was the likelihood, had the highest expected value of success. You know, it was riskier, uh, big risk was the fact that none of us knew how to build native Mac apps, uh, you know, that was something we, uh, we actually, I was, I was very surprised we we're able to learn and get quite good at and recruit people to join that were quite good at that. Um, but ultimately the market and the opportunity and the problem we're solving was so much more painful. And I think that was kind of the key insight for me. Scribe was kind of an, and all kinds of meeting bots are kind of nice to haves it's a vitamin, it's useful, you know. But like the idea that you could capture anything you've seen and never have to take notes again. That was a painkiller for some people. And that's partly, why this idea felt so much more worth pursuing.

Todd: Yeah. Okay. So when you have that many people who are interested in your product, you know, from day one, like how do you manage that? Did you invite tens of thousands of people or a hundred thousand people into the product and they're all giving you feedback simultaneously? Or what, how did you guys manage that?

Dan: Yeah. So we, we actually took a really uh, I, I think, useful approach. So we started, before even announcing it, we were manually onboarding people. So we were actually, we had a product, we were dog footing it ourselves. We had a list of folks that we, you know, my co-founder. Brett Bejcek he would sit with them for 15 minutes or 30 minutes, watch them go through the onboarding flow every step of the way that like didn't work or somebody was confused we'd go back and fix that. So we were pretty confident that the new user experience was magical. We wanted people to get a really fast time to value or time to magic, which is, the first magic people usually have is like, whoa, I can rewind my computer. I can go back to what I was just reading or what I was just doing.

And, um, we wanted to make that as likely to happen for as many people as possible. And then once we announced, we started getting this waitlist and we realized, you know, we saw this with Scribe as well, we had a waitlist that very quickly atrophies. Like, you know, people are excited, they sign up and then they quickly forget.

So we knew we had to get through this waitlist quickly. So we were doing uh, every week we're inviting a thousand people off the waitlist. We would look at the metrics on how people were onboarding. And then we felt, once we felt like, wow, this is actually something that can kind of self serve. We don't need to be there. For the most part, people get it. They're onboarded. And then we just, again, sort of made this leap of faith decision. You know, it's funny in hindsight, these things are so clearly the right thing. But in the moment you're like, are we ready? Should we unleash uh, the rest of the waitlist?

And I think it was probably. You know, only a month, like it was sometime in December where he just said, anyone can download it, you know, go for it. And uh, and it's only taken off from there.

Todd: Wow. So was there any feedback you were getting from these early users, you know, a thousand new users a week that surprised you?

Dan: Yeah, I mean, it's almost always user feedback surprises me because, the feedback you get very often, the things that they say they want in the product or a feature request. Maybe their feature or the solution they're proposing isn't the right thing, but they're almost always right about the problem.

They're trying to eliminate a problem in the product that we hadn't anticipated. You know, One major mistake we made early on was we assumed that people would be really comfortable with cloud transcription. We had done that with Scribe. We were familiar with cloud transcription. The product was entirely local, except for transcription.

Uh, and so when we uh, launched that, you know, we got some, the, um, All Things Tech podcast gave us some criticism on that. And we had heard from users that like, hey, I really, one of the things I love about Rewind is how privacy focused it was. But this one feature feels a little creepy, and so, again, we got lucky there with timing.

Whisper had just come out from OpenAI, and so we were able to go to all local transcription, which was another, you know, that's an insight that we, I had sort of assumed people would be comfortable with that, given that most things are already being transcribed with meeting bots. But uh, that was a lesson that we learned uh, the hard way.

Todd: You mentioned this really interesting point that users are good at sort of identifying problems, not always great at suggesting what the solution will be. Do you have a way of thinking about that when you're building products? Like how much do I listen to the users versus when, when do I sort of ignore kind of what I'm hearing from users? 

Dan: The way I think of it is kind of like the Midwit meme. Like, at some point you should just do the thing. If you, if you hear a user say a thing like a thousand times, just save yourself the time and agony of them having to write in that feedback. Just do the thing. Um, you know, so there's certain features that you should just, if they're easy enough, the effort is low, you should just do as, you know, the user is requesting.

My framework for this is is really useful, which is I view everything we do, I do, my times, uh, my team's time by the amount of impact it will have divided by the effort. And so I think of that as kind of a return on investment of our time. And I'm, I'm ruthlessly trying to prioritize my time, our team's time to really focus on that. So in that sense, the low hanging fruit, the things people just ask for that are low effort, you should just do.

The things that people are asking for that you have an incremental step toward understanding the impact. That's kind of next on my list. You know, the things that are clearly there's something here and we can make a, you know, dent toward progress on this problem that's being illuminated. I'm very in favor of it's often better just to do that than to debate should we do that? And then the hard stuff is things that are clearly hard problems are going to take more than, you know, a few days or a week of engineering time. And that's when you really have to ask yourself is the solution that, that they're asking for the right one? And I think, we view user research as a really critical part to our success, like really listening and honing in on the why's behind somebody's request.

What are the things that they're trying to solve? We have an incredible design team that they can take those why's, the problems you that's we're very problem oriented engineering and design team. It's not about, cool, sophisticated technology. It's like, what is the minimum effort, the best design, to solve the user's problem.

And if it happens to coincide with what their best guesses of what that is, awesome, they'll love it. Usually what we'll hear from users is, wow, that's even better than I've imagined was possible. Uh, you know, thank you for building the solution. It certainly solves a problem we had. And uh, you know, that's kind of we view as our job is like to take those problems, think about how technology and design. Console them uniquely well and um, and then ship them.

Todd: Is there anything that you have to do kind of from a team perspective to get everyone to understand that way of prioritization and what the right problems are to solve? Like, how do you create that culture almost?

Dan: I mean, we ship on average, you know, every engineer ships code to production every day on average, you know, like we are, we are very um, very focused on velocity. Uh, these aren't major changes, but we really view the work that we do in with this lens of the main thing is that the main thing should say the main thing. That's kind of the cultural mantra that I have for my time. It's the cultural mantra that I try to instill into the team. So, you know, the first, you know, couple of cultural things we do to instill this. Every engineer, when they start at Rewind, within 24 hours of starting, they ship to production.

It may not be a big feature, but like, it, it will validate that they have a dev environment set up. It tells them and tells us that, you know, they can ship to production. There isn't this kind of fear of, uh, you know, what would happen if I, if I broke something. Uh, it forces us to have the tooling infrastructure, continuous deployment in place that, that makes that possible, and then we, we really instill the sense of, part of our goal is to hear customers feedback and as quickly as we can incorporate it into making their product better. Um, so that's uh, you know, that's another kind of cultural mantra.

We have cultural values. We go to our, our, our website, we talk about these cultural values that break down into kind of impact and teamwork and a lot of our impact related cultural values are about prioritizing our time well, thinking about high impact versus the level of effort necessary to build something.

Todd: Yeah. And it's really consistent with the thing you said before like, I'm, I'm not going to sacrifice quality. I'm not going to sacrifice speed. So if we need to, we'll cut scope. And it's a way of maintaining that bar in the product. So back on the product, and on sort of like the product marketing or the product positioning.

So you've got a product here that is fairly futuristic, I would say. And I don't think of it as like an existing product category. This is not like a new way to do X. It's a new product category. So how did you think about that and explaining the product and how early adopters would understand it and wrap their brains around it?

Dan: It's a hard problem to solve. You know, I thought a lot about it. It would be great if we had an analog of this is X for Y, this is a DoorDash for cats or whatever, you know, something that people could understand. And I do think a lot of our success or failure, if we, we were, you know, fast forward uh, five years or 10 years from now, we'll come down to, can we tell the story of what we do well enough, that tiptoes between the concerns people have and gets them to that magical moment. Uh, you know, that's the beauty of our product is once we get somebody in the magical moment, you know, it's usually a tablet crash or an email that accidentally deleted or, uh, or just some workflow that we help to make much more efficient.

Once we get them there, we see in the data that they're kind of hooked. The problem is between first hearing about us, if somebody is hearing about us for the first time on this podcast and getting to that magical moment, we have to overcome a lot of these obstacles of I don't know, this feels creepy, why would I would capture everything I say and everything I see, you know, uh, it's a new behavior, any new behavior, regardless of whether it's creepy or not, or perceived as potentially dystopian. Any new behavior is hard. Everyone is kind of ingrained into their existing habits. And so I think part of telling the story of what we do, why we do it, why it matters is really critical.

So I can't say that we've sort of nailed it yet. You know, I think we were on to something and we certainly are attracting a certain subset of our, of the knowledge workers out there that, that find what we do is kind of a key, you know, painkiller for them, but there's still plenty of people who hear about us and they're like, I'm not sure if that's for me.

I hope over time that changes, you know, I hope we are as ubiquitous one day as a hearing aid for hearing or, or glasses for vision. But, um, but that only really depends on our ability to sort of wins over, win over the hearts and minds of people, uh, slowly over time.

Todd: You know, we've talked a lot about the product. The, one of the very interesting things to me here, Dan, is that you are a second time founder. And I think everybody's heard of your first company Optimizely. So I'm curious what feels different about starting a company the second time around? Are there new standards, new behaviors, you know, things that you've learned?

Dan: Boy, I have learned so many things. You know, I feel like now looking back at it, I'm kind of shocked at we got as far as we did with Optimizely, given my youth and inexperience and naivete. Um, you know, I would start with, focus. You know, the mantra I said earlier around the main thing is that the main thing should stay the main thing.

I was very bad at that at Optimizely. I was very attracted to the shiny new thing. Um, you know, we started off with a very clear focus. It was the product I wish we had in the Obama campaign in 2008, when I was the director of analytics to make it easy for anyone to do A/B testing. And that was a very clear focus for maybe the first three years of Optimizely.

But then we got very distracted by shiny new objects and I blame myself for that. I've just I'm, I'm, I'm very, uh, eternally optimistic, uh, about what technology can do. And, and I didn't realize my role as CEO in saying no and helping us focus and, and being decisive. That's the other thing I realized if I look back at Optimizely, the, the class of things, there's like one very clear pattern of things I regret the most and it follows this pattern. My gut says we should do A, somebody else thinks we should do B, and it turns out poorly. There are plenty of times we said, I should, we should do A, and it turns out poorly.

I don't remember any of them. The ones I remember are like these deep existential, my intuition says we're not doing the right thing. We end up doing something else, uh, because that person has 10 years of experience at, at Salesforce, or they're a board member, or they're somebody who's very smart and influential, and, um, and I didn't do the good enough job of embracing the inherent conflict in going with my gut and being decisive. And I was a little bit too hapless. You know, Paul Graham talks about the best founders are relentlessly resourceful. And the things I regret most are the moments I was hapless, where I didn't fight the good fight to go with my intuition.

And we ended up going down a different path and I regret it. So that, I'm definitely not making that mistake again, at Rewind I'm, I'm willing to- good example is the strategic choice to move toward what is Rewind versus Scribe, like that was not a consensus, clear decision. Uh, you know, could have lost some folks on the team for that decision, but I was willing to go with my intuition and, and that's the thing I didn't realize it optimizely is that you, once you've been doing something for 10 years, uh, which is I'd been doing an Optimizely, you gain an intuition that's like very hard to replace by an expert that you hire from the outside.

So, anyway, I'm not making that mistake again. There's so many other lessons I learned around hiring senior people. You know, today we have 20 people at Rewind. We had 450 and I feel like with 20 people, we're able to do more and execute faster and be more tightly aligned than and ship than we ever did at Optimizely.

And, you know, they say if you want to ship a project sooner, take people away, have fewer people. Uh, and so I definitely feel like there's some magic to that senior experience. People at Optimizely, a lot of people are doing the job that they were doing for the first time. And I didn't know any better.

And today we hire people who have done things before, who have experience. Uh, and we benefit tremendously from that experience because, you know, we avoid the mistakes of their past. So I could go on and on.

Todd: Well, you've mentioned to me before, also the psychological challenges that, that come with being a second time founder that weren't there the first time you started your first company. What are some of the psych, what's the psychology of that or some of the mental challenges that you find yourself facing now? 

Dan: Yeah, it starts with the sense of I, had some fear going into this company that it wouldn't live up to the last one, you know, and I had probably an irrationally high bar for even being able to talk about what I was working on. That, a first time founder doesn't have a first time founder just wants something to be successful. They almost don't care what it is. Uh, I, so this time I felt certainly in the beginning, given my prior experience, given that it might get some attention, uh, that I, I didn't want it to suck right off the bat and everything you do in the beginning sucks. You need to be willing to embrace that, uh, and start with that and not let it, you know, cause you to, to make something perfect uh, you know, and, and making that the enemy of good. You know, I see a lot of companies today that are ex Apple people who built amazing things and they go off and they spend years and years toiling away and never really ship anything. And, you know, I think they have this sense of like, it needs to be as good as the last thing they did.

And when you start from zero, it's never going to be that good. So that was definitely a big psychological thing. I had a, you know, you know, I'm very proud of Optimizely, but it's still growing and do you know, it's been acquired, but it's doing well. And, uh, but I also think that we never really fulfilled the true potential of what I thought Optimizely could be.

So I had a little bit of a chip on my shoulder too, starting Rewind. It's like, I want this to be the category defining company that Optimizely could have been, and uh, you know, I, I know it sounds silly because Optimizely did do well, but it's like the, it's rel- it's well relative to what it could have been. And that gap is what, that, that gives me energy and fuel and passion for this time around as well.

Todd: Was there an operating principle or an approach that you took at Optimizely that really worked and that you absolutely wanted to do again at Rewind?

Dan: Yeah. One of the operating principles we took in Optimizely was really trying to walk the walk when it comes to culture, not just talk the talk. Uh, I had, and you, you and I both were at Google for a while. Like Google has this sense of Googliness that you had to assess candidates by. But it always frustrated me that there was no clear definition of what Googliness meant, what it meant to be Googly.

And I actually found that it tended to, on occasion, become weaponized as kind of a way to, uh, sort of select out people who otherwise weren't a culture fit because they were different. Um, and so I, at minimum, at Optimizely, I decided when we were like 30 people. I at least wanted to walk the walk and have codified cultural values to find what a culture fit is.

And in every hard decision I have to make, use those cultural values. And every candidate we hire, everyone understands. So at least we're consistent. We may not be right, but we're consistent across different, people who we're hiring. 

Todd: What were some of 

them? 

Dan: Yeah, we, we described them, we had an acronym called OPTIFY ownership, passion, trust, integrity, fearlessness, and transparency with a Y. So OPTIFY, and then, you know, I think everyone knew that acronym, but I don't know if everyone really understood the meaning and nuance of each thing. Um, and then this time around, you know, we have actually, uh, far more cultural values.

There's no handy, fancy, uh, acronym, but they're comprehensive. Everything we want anyone to do at Rewind is in line with these cultural values. Or put differently, when we do, we do quarterly 360s, which most startups don't do. When we do them, the criteria we evaluate each other, and I get evaluated on by my team, is our cultural values.

There's nothing else. There's no other thing that's your performance it is do you live up to the cultural values, the things we have defined, by definition, are the things we think will make us successful. So if we think those are the things that will make us successful, then assessing each of ourselves on how we're living up to those should be the key to unlocking company success and performance.

So that is definitely something I learned. But at Rewind, I think we've done an even better job at really making that like ingrained into the way we work every single day.

Todd: What are, what are some of the most important ones at Rewind?

Dan: A big one is impact. Just generally, that's the category and that includes things like prioritization, um, quality, not making perfect the enemy of good. Um, you know, we have a whole long list of things that are really about how we do our work. Um, and, and just as important as impact is, is a set of things around teamwork, uh, being supportive, creating a sense of belonging.

One of my cultural, my favorite cultural values came from my co founder, which is a cultural value of zest. Uh, and, and that means, ,you know, bringing passion and energy to everything you do. and, and zestful teammates are fun to work with. Uh, so every quarter people are evaluated on, on how zestful they are. 

Todd: On their level of 

Dan: zest 

yeah, that can mean different things. It doesn't mean that you're sort of a rah rah cheerleader. Some of our most zestful people are, you know, mild mannered, sort of quiet engineers who, you know, they'll add a little flare emoji to something, or they'll, they'll, they'll make a comment that just sort of gives everyone else collective energy.

Um, so that, that's, that's, uh, you know, that's why, that's one of the things I've learned a lot is how important that is and getting that right up front makes a hundred other decisions much easier. It just goes away. I don't have to be in the room for other decisions. If I know we're holding ourselves accountable to the cultural values that I think will make us successful.

Todd: Okay, let's uh, shift gears a little, Dan. You know, I'm, I'm super interested just in the topic of product market fit and, and how founders go about that journey of finding product market fit. It's so hard. But it is the most important thing, you know, in the first few years of a startup, do you, do you have like a personal bar that you like to hit for product market fit or that makes you feel like we have it or we don't?

Dan: For me, product market fit really comes down to a defined market that is highly retained of users who actually love what you do. It's not a nice to have for them. It is a painkiller, not a vitamin. Likely they're, they're referring others, you know, they're, it's, they love it so much that they can't help themselves, uh, from, from sharing it with the world.

Uh, we have that today for sure for a subset of users, you know, we have, there's a billion knowledge workers on this planet. We are not a fit for everyone today, but there's some subset of that 1 billion knowledge workers for whom they couldn't imagine living, uh, their life without a product like Rewind.

And so that is, you know, and I think people, you know, at least I certainly have wrestled with. we have product market fit or not? And when you use a broad definition of what your market is, you get kind of ambiguous signals like, you know, technically, you know of, of Marc Andreessen's, list of seven from his blog post long ago.

We have four outta seven. You know, technically by Rahul's framework of, you know, it's greater than 40% of people, very disappointed they could longer use your product. Yeah, we exceed that threshold. But like, do you have, like, what is that core set of users for whom you are a must have a painkiller? that is kind of the threshold I use.

And, and I think we have that. And our goal right now is just to grow that, you know, it's, it's kind of a movement. There's some set of people who, you know, you see this on Twitter, like they try it, they have this magical moment, or maybe they'll, they, they're able to get information that would have otherwise lost.

and same way when you have, you know, homeowners insurance and that first leak, uh, you're really glad you have it. Same thing with rewind. So think over time, as people have those magic moments, they tell others about it. What we do, we become more and more normal and eventually people look back at, you know, today and be like, Oh, how funny is that we used to like do all of our work and like expect our memories that are flawed to just remember things, you know, and, and, uh, that's our hope is that one day people look back at this and think it's ridiculous that not everyone on the planet was using something like rewind.

Todd: So one of the things I'm curious about is like how you operationalize that day to day, because, you know, by some standards you have it, you have product market fit, but by other standards, you might not. Um, and it's not like you just sort of like flip a light switch one day and now you've got a painkiller.

Now you've got a must have product. It is more of a journey than that. So how do you translate sort of like moving up that scale or moving up that ladder to what you're doing kind of day to day?

Dan: Yeah, I mean I go back to our framework on prioritization, which is to maximize impact per level of effort. That impact usually is reflective of how many of the people who currently use us or will imminently use us get value out of that feature, and how much effort it would take to build it. And so that's how we operationalize it.

We look at it with that lens of what's the highest impact, with the sort of lens of the people who are currently using our product and could be using it.

Uh, you know, good example is, you know, one of the most popular features we launched recently was, as soon as you finish a Zoom meeting, we give you a little popup that say, do you want to draft an email and share a summary of that meeting with the people you're attending, you know, that's an example of a, uh, you know, it was not super hard to build and we had already the data, we were capturing everything that was being said in the meeting where everything that's on your screen and, uh, you know, if you integrate with your calendar, we know all the other attendees, so we know their email addresses.

So it was just sort of taking, okay, what's the next incremental step toward giving users? Uh, who want this kind of superpower, that capability. And, um, you know, that's, I think, a good example of, of sort of our key users, uh, the people who love us already. That's like a, now, yet another superpower that they can't live without.

You know, if they're in back to back meetings, you don't have time to write up notes. You're in the next meeting, you want one click, send a summary, look good in front of your colleagues, uh, look like you took, you know, very, uh, diligent notes, but really you just had the AI do it for you. And, uh, those are the kinds of things that this framework help us prioritize.

Todd: You mentioned kind of, uh, segments of users. And, you know, sort of like, I tend to think of these things sometimes as like concentric circles. And the biggest user is you've got a billion knowledge workers, you said that before, right? And then a circle inside that is like, uh, knowledge workers who have an Apple device, you know, with an M1 or M2 chip.

And then you could maybe even go inside of that and get narrower and narrower in terms of who is your target customer. Do you think about that more narrowly than just like M1 and M2 owners or, or do you, do you sort of like to keep it general?

Dan: This is something I'm probably most embarrassed about in terms of our decisiveness. Uh, it's just we don't have a super specific who. Uh, you know, we have almost by constraints of the technology M1 and M2 users and app-, you know, Apple uh, or iPhone users. But, uh, you know, we've even gotten praise on Twitter, like, Hey, look how great, how is, you know, how focused there're being at Rewind.

Um, and I almost felt like an imposter when I saw that. Uh, Gokal put out this great tweet saying, this is a great example of a company being really focused. I still think we can do more work there. Uh, the challenge we have is it's hard to know, um, how to classify this super specific who, and, and in a way that doesn't exclude other very clear, obvious use cases.

You know, it is kind of a universal problem. Like if there was a subset of people who have perfect memory and a subset who don't, um, then just focus on who people don't. But we all, you know, with exception of a few people who have like this amazing eidetic memory there, most people really struggle with this.

And so we're constantly torn into opportunities with, uh, for broadening our use case. And, and I feel guilty, and I sort of ashamed of this because I look back at Optimizely and with perfect clarity, I can see, with the benefit of hindsight, moments where if we had just been decisive, we had just chosen a super specific who, a good example is at Optimizely we started with the tool for marketers, then we started building things for engineers, and we were always constantly fighting back and forth and resources and trade offs and messaging and positioning and even what the name of the company should be. Because of this tension, and had I chosen one or the other, it didn't even matter which one it was, we would have been massively successful one or the other.

I mean, there's now huge companies that have been built and been successful by doing that. So I know, uh, we would have been, uh, much better off by being decisive. And even with that, I have this cognitive dissonance. I know that was the right thing to do with perfect hindsight at Optimizely. Even today at Rewind, I still don't have the courage of my conviction to be that decisive.

So I hope, you know, we'll get there, but we're not there yet. It's certainly a journey that I hope we, uh, we improve on. But I can tell you that it feels bad to know that if I had to give myself advice, uh, this would be the number one piece of advice and I'm still not doing it.

Todd: Okay. Last question on, on product market fit. Are there any things that you have read or resources that you like, um, kind of on this topic that have really inspired you and that sort of stick with you?

Dan: You know, I, I certainly think, uh, Rahul's, uh, framework that he shared on the first round review, uh, by surveying your users and understanding what percentage of people very disappointed not to use your product. I think that's a good place to start. It's very actionable, which I love. You can just go in and ask your users.

You can then segment your users to see who actually, you know, it responds more favorably. I think another area to look at, and I didn't really read this, but I do think looking at how viral people are, how willing they are to share others, you know, today, people willing to put their reputation behind a product or behind an idea is a pretty high signal that you have product market fit.

So just, and, and, and Twitter has been one of our best sources of users. And, and so I think that's an also a good sign if nobody's talking about your product, uh, on Twitter, um, that's probably a sign that there isn't a painful enough problem you're solving for them.

We found, interestingly enough, that every time we launch a new feature, we just get a wave of excitement. So our best marketing, still to this day probably, is just listen to a user, build something that solves a problem they're talking about, and then announce it on Twitter.

And we just rinse, lather, repeat. You just get more and more users that way. And so we do look at that and sort of the traction of new users.

Cohort based retention and you want that to plateau at some point, you know, instead of users at some point, you know, starts plateauing and ideally smiles starts growing over time. Um, and that's a great sign of it's, it's not just a flash in the pan. I think a lot of AI tools today are this sort of flash in the pan, exciting, cool demo.

Wow, cool. Isn't it amazing the technology can do this, but then nobody sticks around. Um, and I think that's a important metric to look at. That's probably much better than just total aggregate. Uh, ARR, Total Aggregate Number of Active Users, is that cohort based retention after, you know, ideally after four or eight weeks, you know, are you getting a plateau of people who are sticking with your product?

Todd: Make sense. So I love that you mentioned, uh, kind of AI, you know, flash in the pan, cool demos that let's pivot to that. Cause I, I think you might have some hot takes Dan on kind of the current AI market. Um, I have some specific questions, but if you want to just start out, what are your general kind of thoughts on, on what's going on in AI? 

Dan: Yeah, it's, it's an amazing, uh, it's an amazing time to be alive, frankly. Like that, that technology has gotten so good, in particular, large language, Large Language Models, that the distance between having an idea and seeing it come to life is, uh, is shorter than ever. Uh, and not only that, but you're able to, I think show technology can do things people didn't think was possible before. So there are some really exciting demos. I think you have to combine that with this realization that the market for entrepreneurs and just generally what entrepreneurs spend their time on has gone through some changes. You know, there's this whole crypto thing that people were into for a while.

Um, many, in fact, uh, one of the folks that competed with us on Scribe, Uh, you know, we met up there like, Oh yeah, the Scribe was so good, we just decided to give up and we moved to this crypto idea. And then three months later I saw they're now doing a generative AI idea. So there is just a group of people out there for, and I totally admire that they're willing to sort of be flexible.

I think more commonly the mistake is just sort of bullishly persevering on an idea that's not working. So there are a lot of entrepreneurs out there who just like putting out cool demos of things that large language models can do, you know, the cynical view is there thin wrappers on top of open, on top of GPT4.

So I, you know, I do think that's likely going to go away. And there's some, you know, you need some differentiation because it's easy for you. It'll make it easy for your competitors. And, um, so if it is a good idea, somebody is going to do it and it could even be Microsoft or open AI that does it. So you do need differentiation in this market.

So, I think those over, over time will go away. I have a lot of other hot takes on superintelligence where that's going, but I'll leave that to later in the interview.

Todd: Okay. What is the, uh, we use the word thin wrapper a lot. What, what is the difference between a thin wrapper versus something that is sufficiently like thick or not a thin wrapper to you?

Dan: For us, the biggest difference is the data. You know, we capture everything you see saying here, we store it in a way that's private, local to just you. And then we use that, just subsets of that, to serve you in the asks that you have. You know, example, you ask Rewind to write you an email. You know, if I asked rewind, Hey, write me an email to Sam Altman asking to catch up.

The data of the fact that we were introduced by Paul Graham in 2010, the data of the fact that he's an investor in our company and that, you know, he just launched a new feature. All of that is data that no large language model company or provider could offer. And that gives us like a unique differentiated approach.

You know, we use the power of the reasoning that large language models provide with the data that we have that no one else has. So I think that's a good example of what the, what, what the opposite of a thin wrapper is. Um, you know, the other part is owning that user experience. And we have, you know, the users are using us, it's, it's a native app they've installed, they've gone through that friction. Uh, so we, we have a fervent user base of people who use us every day. That's also something that if you think about it from a competitive lens, even if you built, you know, a web based version, um, the switching costs are very high. And we also have everything you've seen, said, or heard up until you start, since you start Rewind.

So even if tomorrow some big tech company launches what we do, you know, we've got a lot of users who say, well, why'd I switch? Because I already have all the data with Rewind. It's not like Dropbox you could just, know, copy the files from one to the other. So I think we have, you know, in that sense, a huge switching cost for new products.

So those are all things I look at, you know, proprietary data, high switching costs, owning the user relationship. And doing more than just what OpenAI could do if they wanted to.

Todd: Dan this might be a hard question to answer because I know you're so focused on Rewind. But if you were a founder, like let's say you were starting a new company in 2023 again, and you were super interested in AI, but you couldn't do the Rewind idea, but you could do something, you know, at the foundation model level, at the infrastructure level or at the application level, where do you think is the most interesting kind of hunting ground or, or places where new founders should look?

Dan: I absolutely think it's the application level. And there's a couple of reasons why. I think the infrastructure, foundational level is incredibly hard to win unless you have a ton of capital. You're also playing from behind. I think there's a lot of reasons to suggest that, you know, OpenAI, probably Google and others, and maybe Microsoft, are going to have a huge, huge, hard, entrenched positions there.

Um, I think the infrastructure layer, I think you're going to be met with a ton of competition. You know, every failed crypto company who had sort of a developer mindset thinks, Oh, wouldn't it be cool if dot, dot, dot, and then build some technology that they think somebody will want. So there's all of these founders out there just like trying different infrastructure thinking that like each VC and I'm sorry if First Round did this, uh, has like the landscape, but like, here's what the AI landscape is going to look like. And you know, they've got like buckets and categories and, uh, you, you're competing with a lot of technology in search of a problem. I think many of those companies will fail, but the risk for you, if you're going to be a founder in that space is that as soon as you're successful, like you have to assume if you're either going to be, either what you're doing is a good idea or a bad idea.

It's a bad idea. You'll fail. If it's a good idea, you now have hundreds of other founders who are like, Oh, that's a good idea. I should go do that. So you're in this very crowded sort of B2B space, um, trying to build tools for other AI companies. So I believe the true value is in the app side. I think it's very under appreciated.

I think, um, for our benefit, you know, it's helpful that uh you know, folks like YC, who I love, really push companies for more B2B companies, like a truly consumer company is kind of out of vogue right now. Uh, you know, most investors like to invest in something that the founders, if at minimum needs to get distribution, they can go out and sell, you know, one customer at a time.

So I do think consumer app facing niche, very valuable sort of uses of large English models is probably the most ripe, um, ground for an entrepreneur today thinking how should I get into AI?

Todd: And are there, are there particular areas of AI, obviously the whole space is very, very hypey. Are there particular spaces that you see and you're like, okay, this is gonna live up to the hype versus areas that you see where you're like, eh, that's probably not gonna live up to the hype.

Dan: You know, there's a couple of ideas. I don't want to mention them because they're startups of people I know who are starting. I think are really, really good ideas. I'll tell you there are ideas out there that you probably haven't heard of that are in the app level for AI that are just really powerful and magical.

Um, and I think will change lives in society for the better. So, like, I think, you know, it's, you know, maybe the best analogy is to think about, like, you know, when the microprocessor first came out, and people are thinking very much like, oh, cool, like, computers, like, these people who hand tabulate, like, now we don't have to do trajectories with, like, ballistics. Great. Like I guess that's what computers do now. And like they could have never imagined what the computer had enabled, like the whole set of applications that you're able to do the whole, the whole internet thing that come, came later. So I think we're at that moment for AI, that there's a whole set of applications we never imagined.

And you would look back at sort of think how, how narrow minded we were about what was possible. Uh, and that's another reason why I'm so keen on the application side. Like there's whole categories of companies that are going to be cat-, you know, category defining independent, successful, big, public, multi billion dollar companies that are just applications of large language models in a specific domain for a specific user with a great user And so I think that's, uh, you know, I think there are thousands of those companies, frankly, that will emerge over the next few years. 

Todd: With all the new advances in AI and, and kind of, you know, in some ways surveillance technology, and I know that Rewind is not that, but I'm just sort of like saying broadly. I think there's just been a ton of scrutiny in the past several years about how companies handle data privacy. And I think it's, you know, Rewind was very clear from the beginning, core tenet of your product is that everything is local, everything is private, not even Rewind employees can touch it.

But how do you think the consumer point of view has kind of shifted over the years? I think for a long time, you know, when you and I were at Google, uh, and then after that I was at Facebook, I, I feel like there was a period of time, 10, 15 years ago, where, the thinking was that consumers are willing to sacrifice some of their data privacy in order to get a great free product experience.

Do you think that's changing?

Dan: I think that if you look at the history of this, it's very interesting. There are a lot of things that people would be terrified of in the moment that a computer would do that over time became so normalized and the downsides were totally overblown. Uh, you know, in, in, just in the, just in the class of memory, I'll give you a few examples.

One is like, at some point, like I had to remember people's phone numbers. When I was growing up, I had to like remember people's phone numbers. I typed them in into the phone. I wouldn't even type them. I'd like push them into the buttons . And I'll call people. At some point like I stopped having to remember phone numbers cause my pho- my computer, my phone could do it for me. Uh, I still remember my, my wife's phone number, but other than that, I don't really know anyone's phone numbers. You look at that then the next level at some point you had to like, you know, understand how to get from one place to the other using, you know, uh, you know, cross streets and like, you know, now the machine can get you from point A to point B much better than you could before. Um, you know, more maybe closer to privacy. Like at some point when, you know, Apple launched the ability to share your location with your friends, that was terrifying. Like, why would I ever share my location? But now it's like super convenient. Like I actually, when I'm meeting with somebody, I can give them my live location.

They can see where I am and we can find each other. And so the when the utility greatly outweighs the perceived concerns, I think that's when these things tip. And I haven't heard of situations, I'm sure they happen where like the worst dystopian path on things like sh- location sharing, um, happen, but more often than not, it's the perceived fear.

And it's actually the change that's the biggest thing. It's like people are just really reluctant to change any amount of technological change, whether it touches your privacy or not, uh, is very hard for people to, uh, as much as they think that they're adaptable and willing to try the new thing, it just changes hard psychologically.

And we're just rooted in, you know, and trying to think of the world in the status quo. So I think, you know, I don't think it in this case for Rewind, it's a choice. We have to be the most privacy sensitive, most privacy first company in our space. And we have to give you the convenience that gets you past all of the potential perceived concerns.

You know, for us, the test I always use is if we were subpoenaed by the government, would we hand over anything useful or meaningful? No, like we never want to be in a situation where we have data about anything you've done that then we could hand over and that could hurt you, you know, if you're in a, in a, in a, country where, uh, that may be that get, can you get used against you? And so we, we hold ourselves to that bar, never get in a situation where people's privacy are compromised by the choices we make and give you the convenience of things that you never thought were possible. Like, you know, being able to go rewind time and go back to something like you, you, you know, something you, uh, a tab that crashed or, you know, an email that you, you accidentally deleted.

So those are the things that I think change people's perception. You asked about consumer, consumer mindset we have found, it's funny, like, well, you know, but people come to our website, sometimes they're a little skeptical. They'll do some research. You know, we have a whole privacy first website. We'll get some questions from them initially. And then what happens is they start using our product. They have that first magical moment and then their entire worldview changes. Their questions are no longer about privacy. They're about, Oh, Hey, wouldn't it be great if I had this on my phone and can you actually synchronize this data across my Mac and my phone? Uh, you know, and like all the questions that would have been the last thing they're thinking about now become the forefront cause they've sort of been sort of you think about Maslow's hierarchy of needs, you know, they're thinking more about self actualization, not about just food and shelter. 

​Todd: So this might give you an opportunity to share some of these far out views that you mentioned that you had. Because I'm curious, where does this all go, let's say, 10 years from now? Um, you know, as, as human beings, our relationship with AI, the way that we use AI, our relationship with data privacy. What do you think this looks like in 10 years?

Dan: Well, I think a lot of jobs will be gone in 10 years, but they'll likely be replaced with new ones. I think that's probably an obvious conclusion. I think on the order of probably 50% of work will go away or be fundamentally changed, some much sooner than later. I think, you know, people used to think of AI taking away the blue collar jobs.

I think it's more likely to get rid of the white collar jobs. Like if you're a lawyer today or if you're in law school today and you're listening to this, I would suggest pivoting to being something else. You know, I think entire professions might get massively transformed. Um, but I also think that, um, our lives will be massively better.

You know, there's things that computers can do so much better than we can do for ourselves. Uh, an example they gave earlier of like remembering phone numbers, like it'd be kind of a pain to have to remember that remembering cross streets, it's kind of a pain. Think about that for so many parts of your job that a computer can do it much, much better for you.

Um, and so I think our lives will be massively better. I think we will imagine, our kids will think, you know, it was so amazing that we could live our lives, you know, to be the equivalent of like living before fire, you know, like, how did you do it? Yeah, we did it. It was kind of hard, you know, fire was really helpful.

But, uh, you know, but we survived and lots will change. Um, I, the thing I'm probably the most contrarian on is I don't believe in this sort of ever accelerating AI. Imminence, you know, like, and I think people conflate AGI with superintelligence, like, I don't, I think AGI will be amazing, Artificial General Intelligence.

I think it's possible we already have it today. And the people who have it aren't talking about it. I also think as soon as we have it, it'll be amazing, the first press release, and then just like every other new technological innovation, three weeks later, people will assume it's, take it for granted. Uh, it'll be like wifi on airplanes.

Like the first time you try it, you're like, I can't believe it. I have wifi on an airplane. And then like a week later, if you take another flight, it's like, why isn't it super fast and why can't I stream? So I think AGI will be like that. You know, and maybe if you don't believe me, I'll give you the thought experiment of like, you know, AGI is basically a computer who do basically knowledge worker tasks, like an average human, but every day a ton of humans are born. 

Does the world dramatically change overnight just because a human being was born? Not really. And the first AGI is going to consume way more energy, way more resources, way slower than a new human being born over, you know, so I think, you know, I think the idea that all of a sudden AGI will exist and then super intelligence will take over the world and Skynet and all that is uh, there's a, there's a sort of a logical fallacy there where like, like one really incredible feat of technology will then inevitably lead to something that's completely different, um, in, in a way that will change society.

I think both will change society, but super intelligence is much further away than AGI.

Todd: Do you have a thought on, like, what will be the first useful, widespread application of AGI?

Dan: Hard to predict. Um, I do think things like, white collar knowledge work that is high paid, lawyers being one of them, is probably a good area. I think anything where today people are paying a lot of money and people are being paid by the hour, uh, that, that a machine could do as better in some cases, much better because they're not going to make the same mistakes that a human can make.

Um, I think that's likely to, to go away. Um, you know, I think programming and building things is also one where like, Uh, I'm a big believer there's a company, a set of companies in this space that I think are really promising, which is, you know, going from, you know, ticket in your, your linear, you know, uh, ticketing system to a pull request, uh, you know, taking an idea and turning it into action, uh, when you're building something, I think a machine can do actually already pretty well, and certainly AGI's could do even better.

Uh, and I imagine that's going to be gradual, you know, there'll be more and more today. If you look, you're an engineer, you're a product manager, you've got a whole backlog of tickets, uh, you never get to, but if an AI could do it for you, you'd probably ship and sort of these low, you know, You know, small incremental things, um, that space of things that an AGI can do, we'll get more and more.

And eventually, you know, I don't think it will necessarily replace engineers, but it'll certainly augment them. You'll have plenty of, you know, half the work that a engineering team does will be done by AGI and half will be done by humans. And the human stuff will probably be the fun, creative nebulous stuff.

And the AGI stuff will be the, in hindsight, the menial, boring, you know, changes that, that, uh, today engineers are saddled with.

Todd: All right, let's talk about fundraising, Dan. Because this got a lot of attention on Twitter earlier this year. Uh, you know, obviously you took an unconventional approach. You, you basically did a public fundraise on Twitter. Um, and I think this, there are not many founders, I think, who could have pulled this off the way that you did, but you did.

And, you know, you got, I think, 170 plus investment offers. You know, you ended up doing a great funding round with a $350 million valuation. So my first question, Dan, is like, why did you decide to do this?

Dan: The first, well, first of all, this was not an obvious again and not an obvious decision in the moment in hindsight turned out really well. So I'm glad we did. And the first reason and most important was building trust with users. You know, we had, um, you know, a lot of users who come to us, they think, you know, what happens if you get, um, you know, bought by big tech and despite all your promises now you're, you're, you know, it's going to turn into some data selling, harvesting ad driven business.

So we want to show them like we are a real business doing really well, great traction, great. And we're not going away anytime soon. And we wanted to do it in a way that, you know, I, I, I'm a big believer that transparency breeds trust. Um, that's why I announced and shared our investor deck unredacted.

I didn't hide any numbers. Uh, you know, you can find it on Twitter, you know, seven minute pitch, uh, you know, millions of people have seen it. Uh, it's another reason I did, by the way, more recently, I, I published my 360. I took my 360 reviews from the last five, uh, 360s about our cultural values. I published that, you know, transparency, most people wouldn't do that either.

But for the same reason, transparency breeds trust and trust is the most important thing we have, um, uh, from our users. So that was primary reason one. Number two was I was getting inundated by these investors who just want 15 minutes or 30 minutes. And some of whom I felt kind of like, I would offend that.

I mean, at most meetings, I'll tell most people, I'd say, Hey, sorry, I'm not meeting with investors right now. I meet once a quarter and schedule it. And I started to realize, you know, actually the better solution would be just to point them to a link of a video. And if they want to invest, let me know. And so that logically led to me.

Well, then if I'm just sharing with people, emailing me, why don't you just put it on Twitter and like, you know, see happens 

could 

Todd: have made it, you could have made it like a private Loom video though. 

What, What, 

what was the step where like, I'm just going to put this on, I'm just gonna put this on Twitter.

Dan: if I'm really honest. It was a little bit of just like the curiosity of what would happen. You know, we could have, I think we would've gone just as much of a great outcome if it was a private Loom video and you create a little bit of artificial scarcity. Um, and it could have backfired if we had published it publicly and it had flopped and like 10 people liked it and nobody was retreating it, then that kind of is a negative signal to investors like, Hey, they're putting it out there, but they don't seem to be getting much traction.

So it definitely was more risky because you now have this like societal mimetic desire, like do people like it or not? Regardless of whether an investor likes it. They like to think they're, uh, you know, not biased by what the, what the herd think. But, um, and, and I thought, you know, and it came down to, I think the decision about that really was down to what would users want, you know, what, if I were a early adopter Rewind and I'm trusting them to say what they're going to do and, and, and, um, not, not betray us. What would I prefer? And I thought, yeah, they would want to see it. You know, this is a deck that shows them that we're viable business. We're going to be around for a while. And, uh, that's what sort of tipped me, I think, to the edge, but some people on the team still, like they said, don't do it.

It's, it's not, there's a reason I'm also, by the way, deeply nonconformist. And the best way to motivate me to do something is to tell me that the opposite of that thing is the way we've always done it. Do they know when no investors, there's gotta be a reason why founders don't do this. Like, don't be silly.

Like, you know, you're, and that's the best way for me to prove you wrong. Like, no, there's, this is not a good idea. Like the best way to analogize this is that like, today, if you wanted to go find love, you could go to a bar and meet somebody in person who happened to be at the bar at the same time. Or you could use a dating app and find lots of potential people who you might want to spend the rest of your life with.

And that's how I view this relationship with my investor. It's somebody I'm going to spend the rest of my life with. I'm going to be doing this for decades. I don't want to just meet the person I find at a bar. I want to cast a wide net and potentially find an investor who I might not otherwise happen to co locate at the, you know, the metaphorical bar.

So with that analogy, hopefully it makes sense. Like, yeah, it makes sense to cast wide nets. See who's out there. See who's interested, where do you have an alignment of values? Uh, you know, not just who's giving you the highest valuation. And, um, and I think that the outcome speaks for itself.

Todd: Okay. So let's say there are other founders out there listening to this, who want to try this, try to do what you did. Can you walk us through mechanically how it worked? Like, okay. So you've decided to do this. What do you post on Twitter? What do you post on LinkedIn? And then what happens like in the days after that? In the weeks after that?

Dan: Yeah. So mechanically step one is come up with a really good short pitch. Tactical advice, number one, make the very first frame of what you share on Twitter really compelling. Uh, because that when people are auto scrolling, that stays on there for awhile. If you go back to our video, you'll see it's this hockey stick of our ARR growth.

That first frame is really important. Uh, don't, don't under appreciate that. Step number two is, is create, I created a type form of just like, you know, your name, your firm, what valuation you want to invest at, uh, which was very helpful by the way. Cause then I got like an actual distribution of what the market thought I was worth.

And how much you'd be willing to invest. Then, um, step three is, uh, I used AngelList for this, but, uh, there's, they have a roll up vehicle mechanism. That's really, really handy. Um, because if this is a public fundraise, there's a SEC exception to allow you to do this. If and only if every single investor is an accredited investor.

So this is not a crowdfunding round that, that's when you can get around that. I very much wanted this to be by the book. I didn't want actually people who weren't accredited. I kind of felt like I didn't want to, you know, I didn't want, I wanted people to sophisticated enough to know what they're doing to invest.

Um, that is important is make sure everyone is an accredited investor. and so I created an RUV for that. And we ended up having, you know, hundreds of people co invest with our lead investor through that mechanism. And then the last and I think most important step is don't treat this transactionally.

And they may be perceived as transactional cause it's public and there's a form. But I meant, if you look, I even tweeted this, my calendar, like I met with dozens of investors, uh, the week after. Uh, who are the most, you know, interesting and compelling people on that list. So, and I interviewed them, I did references.

I grilled them. Like I, I, again, I view this, the person who's gonna lead this round at NEA is what the, the firm I chose as a relationship I'm gonna have for decades. So I think that would be one mistake you could make is I think that this is gonna save you a bunch of time in terms of diligence and picking the investor and you just pick the highest valuation.

In fact, we didn't, we had 20 offers or more of a billion dollar valuation. We turned down. Um, in fact, some of them, we, some of them, we actually let them invest, but at the $350 price. So they're thrilled about that. Uh, you know, so we, we tried to do right by investors. We didn't give people different prices.

Um, so that just some tactical advice. If you're going to pursue this. I would say that if you're going to do this, you know, you want a pitch that's good enough that you feel like is going to get traction because the risk here is you put it out there and it's a big dud. The thing I did do that I didn't mention that, that I think led to that pitch being pretty good is I had met with a lot of investors, you know, I, and that was kind of the solut- this was a solution to having to avoid doing that. I'd already honed my pitch if I have like a keynote folder or folder of keynote files where like I met every time I meet with investor depending on how it went, I would slightly tweak the presentation of proactively bring up something. So by the time I had made it public, it was so refined and so well kind of rehearsed in my mind and with people, I knew it was pretty good, so I wouldn't just do it your first version. I would hone it and use, uh, I hope I this doesn't offend anyone, but use like associates and VPs at firms that you don't want to take money on as practice, because you meet these people that their whole job is to meet with you and get deal flow.

But like, actually you know, Paul Graham says never meet with these people. I say meet with them, but as a way to get practice because their questions and their kind of perspective will be the kind of the, the, the, if you can make it easy enough for them to understand, you're absolutely going to make it easy enough and compelling enough for the general partner who has more experience and perspective.

So, my advice there is practice on associates, even if you're never going to take their money, uh, and, uh, you'll, you'll make the pitch better each time you do it.

Todd: So if you, if you decided to do this again for your next round, is there anything you'd do differently?

Dan: I'd probably make it shorter. You know, the process didn't, it kind of dragged on cause I had so many people that were part of the round. Uh, I would have maybe picked the valuation and then just said, this is it. Yes or no. The part of the auction dynamic, I think made people feel like, um, they were part of an auction and investors don't like that feeling. They like to think they got a, you know, that they did their job to get a good price at the current price. So I think maybe that would have been better is to sort of make it less feel transactional, and make expectations clear upfront.

I think a lot of people were disappointed that they couldn't invest. Like there were people who wanted to invest and couldn't. And I think I probably. Could have had more investors. That's the other thing I ran up to is there is actually a number, a maximum number of investors are allowed. I knew that there were like limits on the credit.

I didn't realize there was a number in particular rounds. So that was something I should have maybe been more picky on is like, how do I make sure I just pick the right people of, you know, I really diligenced the lead, the rest who participated, I didn't do as much homework on because there's hundreds of people.

So I probably missed out on like, you know, I now get every so often some DM on Twitter from some really incredible founder who, who stumbled upon this presentation and who just, the stars didn't align and we couldn't make it work. So those are some things I would do differently.

Todd: You know, what's, what's next for Rewind? What, what are like you, like over the next two years, let's say at Rewind, what, what's most important to you? What are you most focused on?

Dan: So we are really focused on this idea of giving people superpowers and that might seem hyperbolic and bold, but it's the idea of what would you do if you had a superpower? What are the things that the machine can do for you? So in the space of that includes things like, I mentioned earlier, giving you an automatic summary of the meeting and sharing it with the participants afterwards.

What are things you're doing today that the machine can do for you? Um, so that's a big class of things. Um, so I would say that would be on the category of depth. Being more in your daily workflow, more immersed, taking things off your plate that are, things you'd be doing otherwise. And then we're also pursuing breadth, you know, it's, as I said earlier, we're going to Windows, we're going to launch that later this year, uh, likely Android next year. Today we're on iPhone and we're on, on Mac. Um, I'll tell you also a little secret that, you know, we are hiring for hardware, uh, people, uh, so that's possible that down the line, uh, if we cannot, and by the way, I'm very much a big fan of software, I don't know if you're, uh, if you've ever done anything hardware related, but software has many, many benefits, uh, of which I'm the biggest fan.

But there's limitations of what hardware that that Apple provide and you can buy off the shelf. And to the extent that we can deliver on our vision of giving human superpowers with hardware, that might be coming too. So, um, so that's our, that's our goal. And every day we use this ruthless framework of prioritization of highest impact for level effort, we're deeply customer centric and empathetic. You know, I'm not the kind of person who sits down and writes a six year roadmap and with quarterly milestones, like we have, you know, uh, adapted and listen to users in ways that I never would have anticipated. And so we keep maximum flexibility around new technology that becomes available, that can help us deliver on our vision, and to the extent that we need to build our own, we will.

Todd: Okay, Dan. So wrapping up, you know, I love to ask these questions of every guest on the show. are there people, are there mentors that you really learn from in your career and, and what did they teach you?

Dan: The most influential mentor I had was Paul Graham. I did a Y Combinator in winter 2010. And I had previously worked at Google. I had seen how it would take what it takes to build a product there, and I didn't, even, having worked at Google and, and worked on products like Chrome and AdWords, I didn't understand the mantra of make something people want.

Uh, it seems like the most basic thing. Uh, but at Google, you know, they value sophistication of the technology more than they value solving a customer problem. And so that was a cultural thing that started from the early days that served us well. It served Google well for certain products like search and probably will for large language models.

Where like the sophistication of technology is the thing that matters. But you look at a long list of Google products like Google Wave and Google Glasses and Google Buzz. Things that were technology in search of a problem. So I had to learn the hard way what it takes to build a product people want and it starts with, um, you know, deep customer empathy and solving a problem, spending 80% of your time obsessed with a problem and 20% of your time obsessed with the solution.

Um, so that's, that's one mentor who, uh, yeah, that, that, that's a lesson that I've had to repeat to myself over and over again.

Todd: And so this, this might be a piece of advice from Paul Graham or, or anyone that you could think of, but like, what's, what's a great piece of advice, uh, that you've received that, that you've really, you know, stuck with you and that you've used?

Dan: The main thing is that the main thing should stay the main thing. Uh, I think this was an old cultural, uh, HP aphorism. There's somebody else who'd come up with it before that. But this one sentence has stuck with me for many years now. And as a sentence I remind myself of repeatedly. And it's a lesson I learned also in hindsight from Optimizely.

Where in Optimizely, you know, I started it in my 20s. I, you know, I didn't have a family and I was just, I made up for lack of prioritization with sheer number of hours. Like I just did a lot of stuff and hopefully some of it worked. And in hindsight, very few of the things I spent my time on, the decisions I made actually made a difference on the success or failure of the company.

I thought a lot of the activity I was doing was helpful, but if in hindsight, really only a few things mattered. And now I realize this time around, I have three little kids, a five year old, a three year old, a one year old. I'm, you know, I'm, I'm there for dinner with them every night. I, I realized with the constraint of the limited hours in the day, I have to be ruthlessly focused on what matters.

And so that mantra of the main thing is the main thing should say the main thing is both something I treat personally. Like, how do I make sure in my life I don't spend time on things that aren't the main thing. Uh, it's a mantra I treat my company with. I try to make sure everyone's working on the thing that is the main thing.

And it's, it's just sort of a lesson I had from my last company that at the end of the day, most of it, well, at the end of the day, probably nothing matters if you're being a Buddhist about it, but if something to the extent that something matters to the success of your company, uh, it's likely not the sheer number of activities you do, the number of networking events or how quickly you're responding to emails.

It's like one or two critical things, maybe three critical things you get right. And if you knew what those were in the moment, you'd do them, but in hindsight, you realize most of the stuff you do doesn't matter.

Todd: Are there any great books or resources that you would recommend in particular for other founders?

Dan: The most influential book I recommend to other founders is The Mom Test. Uh, it is a book about how to talk to your customers when they're lying to you. Uh, and it, the whole premise of it is basically like most founders when they talk to customers, they lead the witness. They talk with, they start with a solution.

Hey, I've got this demo. What do you think? Would you use it? Worst possible way to have a conversation with users. So I highly recommend it. It's a short book. It's actually the highest ROI time. You can listen to it in a single setting of an audio book. I did that on a flight to San Francisco from Denver.

Great, great book. The Mom Test. Highly recommend it.

Todd: Love it. Dan, this has been such a great interview. we covered so many things, you know, your journey to product market fit, your hot takes on AI, the unique way that you raise funding. Um, and so I think this is going to be a great listen. Thanks for being here, Dan.