Jeff Seibert: Why OpenAI Will Become an Infrastructure Play | E1085

Jeff Seibert 0:00 I think Google's by far the most vulnerable. They need to go all in on it. I don't think they have a choice. (instrumental music) I view OpenAI probably evolving more into an infrastructure company like AWS. The road ahead for OpenAI is not easy. What very few people, I think, are paying attention to is Apple, because again, they control the silicon. Imagine they're able to pioneer small models that run on-device, and then they do custom silicon to make 'em run. The performance could be outlandish.

Harry Stebbings 0:27 Jeff, I am so excited for this. What people don't

Harry Stebbings 0:30 know is I will always remember, I remember being 18, maybe 19, and being in your office at Twitter in San Francisco. I was so nervous. I was like, "This is so cool. I'm with, like, the head of product at Twitter. This is so cool. Mom's not gonna believe this." Anyway, that was a while ago. So first, thank you so much for joining me today, Jeff.

Jeff Seibert 0:49 Harry, no, it is so great to be here. Thanks for having me on. You know, the whole world only makes sense going backwards because, like, who would've guessed from that meeting you'd become one of my largest investors, like, five years later? Just incredible what you've done.

Harry Stebbings 1:01 I mean, uh, yeah, definitely not me, to be honest. (laughs)

Jeff Seibert 1:05 (laughs)

Harry Stebbings 1:05 Everyone always thinks that things are so strategic and you're like, "Well, you know, sometimes you have to go to the party to meet cool people." Um, that's what I always say. But I wanna start,

Harry Stebbings 1:14 I find actually, like, one's childhood aspirations quite revealing. What did you wanna be when you were a child, when you pictured yourself growing up?

Jeff Seibert 1:21 Oh, man. So I loved building things since I was little, and I was completely obsessed with LEGOs. And so my dream was honestly, literally to be a LEGO Master Builder, until my mom did some research, and she found out that actually, like, it's not that great of a career. They're paid something like minimum wage. And that was in middle school. And so I forget if it was that Christmas or the next year, but she gave me a programming book for Christmas, and I was, and that was the end of the story. I was like, "Okay, computers are next."

Harry Stebbings 1:49 (laughs)

Jeff Seibert 1:50 (laughs)

Harry Stebbings 1:50 Dammit, I think you'd have made a magnificent Master LEGO Builder. Um, it's a little bit like journalism. It's like, oh, a wonderful creative pursuit and then you're like, "Really? That's the pay?"

Jeff Seibert 1:59 Right, exactly.

Harry Stebbings 2:01 Um, okay. So when I met you, you were at Twitter, and it is a incredibly formative experience, I think, being at Twitter, especially in the role that you were. How did it shape your operating mindset and approach today, do you think?

Jeff Seibert 2:15 Yeah, the biggest lesson I learned was empathy, honestly. So when you're in consumer software, you can't possibly begin to understand how many different people, personas, use cases, mindsets, like the human experience all comes to bear on your product. And what I saw was actually a trap. So the product managers who were super data-driven started designing and building features for the average user, 'cause that's what the data told them. And they actually believed that there was something such as an average Twitter user. It's such a huge mistake, right? Like, you're conflating all of these different populations. You have sports fans who want a live, like, chronological timeline during the game. You have celebrities who want to maximize their reach. You have Japanese users who, by and large, want to remain anonymous. None of them is average. And so what I really learned is you have to deeply understand each population, and design and build a feature for them. Don't, like, let the data lie to you.

Harry Stebbings 3:09 Can I ask you a bit of a weird one? But often we're told, you know, solve a problem that you personally know and experience and care about, like, so you fully understand and have that empathy.

Jeff Seibert 3:18 Yeah.

Harry Stebbings 3:19 But then other people have said to me before, "Don't, because you can be too emotionally attached to it, that you don't almost think rationally." Would you agree with that? Which side are you on?

Jeff Seibert 3:29 I'm definitely on the side of build it for yourself. Like, you deeply understand the problem, that gives you superpowers in terms of solving it the right way. You definitely need to understand, are you sort of, uh, repeated by many other people around the world? Do they share your problem? Do they share how you think it should be solved? But I think it's so much easier to build something that you personally feel than have to sort of try to interpret other people's thoughts and beliefs about it. Um, so no, I would not be worried about getting too emotionally involved. I think that's a superpower.

Harry Stebbings 4:01 Uh, another thing is, like, speed,

Harry Stebbings 4:03 which is often, you know, like... (laughs) This wasn't in the schedule, but I'm just too fascinated. Like, the Twitter's product, uh, kind of cadence now, regardless what one thinks of Elon, and we won't get into that, but, like, the cadence of release is very impressive in terms of what they're pushing out.

Jeff Seibert 4:17 Yeah.

Harry Stebbings 4:17 How important is speed, bluntly, when it comes to product, product cadence, do you think?

Jeff Seibert 4:24 I think it is critical. Um, and it is too easy for companies to fall into like, "Oh, we don't know. We need more data. We need to run a survey for that." Particularly at Twitter, we need to get StatSig data, which always took two to three weeks.

Harry Stebbings 4:38 Huh.

Jeff Seibert 4:38 And then you would have to analyze that and then see what happened, and it's a, it's honestly a disaster. And so yes, I agree with you, like, agree or disagree with some of Elon's decisions, but he is moving quickly in a direction that is way better than sort of standing still.

Harry Stebbings 4:54 I, I totally agree with you there. Um, I do have to ask this, which is like, you know, when we think about non-obvious things, we mentioned that two commonly said tropes, obviously speed and then solve for a problem that you know deeply. I, I think there's a lot of things that aren't well-known about entrepreneurship. Given the fact that you've done Crashlytics, you've now been in Twitter, you've now founded Digits. What do you think is the most, like, misunderstood or non-obvious element of entrepreneurship?

Jeff Seibert 5:20 This sounds silly, but honestly, pure execution. And what I mean by that is people know the vast majority of managers are terrible managers, right? The vast majority of founders are simply bad at running companies. And, and I'm sorry, but it's true. And so what I mean by that is, like, most founders aren't intentional about how they go through and operate the business, intentional with their time, with their decisions, with who they hire, with what they say no to. And so if you don't have conviction, you're really gonna struggle as a founder, because you need this, like, deep-seated obsession of what's right and what's wrong and what you believe in, and how that informs every decision you make. And your decisions may still be right or wrong, they're not gonna be perfect, but if you are intentional about them, at least you can trace that back and learn from it...... versus I see too many founders just sort of going on a random walk, and then, when it turns out they were wrong, what do they have to learn? There's noth- there's nothing to trace it back to.

Harry Stebbings 6:15 Can I ask you s- unpacking that. You just gave me, like, gold dust there. Why are m- most managers ma- bad, do you think?

Jeff Seibert 6:22 P- Peter Principle, they get promoted into management because they were good at a former job, and so, they- their passion, right, their experience isn't managing, and they're- th- their feedback cycle is slow. They're the boss, so they don't get the raw feedback. I think there's a ton of challenges, and it's even worse for CEOs, because who in your company is gonna give you really crisp, blunt feedback on what you did right and wrong? And your investors aren't involved enough in the day-to-day to really know. So they can give strategy advice, but they don't know how you're behaving in meetings.

Harry Stebbings 6:52 Okay, so help me out here. W- we're both like CEOs as well. How do you think about promoting people then who are great ICs? Do you not promote them to managers? What's the right way to do that challenge, then?

Jeff Seibert 7:03 That is a great question, so actually, every time I've promoted an IC to a manager in a startup, this was not at Twitter, but in a startup scenario, we've done it as sort of a trial period. And so it's like, "Hey, so-and-so, we have this opening for this new role. The team's growing, we need some structure. Uh, we are going to have so-and-so take on the role as a trial over the next two months, and let's see how they do." And honestly, there's been both outcomes. In some circumstances, it's been great and everyone's rallied around them and it's like, "Great, okay, now they're the manager for that team," and I've had circumstances where they haven't, and it was sort of widely recognized that they weren't excelling in that role, and we decided to move them back to an IC. And that was okay as well. I think there's a f- a really bad perception that being a manager is better than being an IC. I think it is different. And you can be an exceptional IC, and you should be comped appropriately for that, or if your career passion is to mentor and guide folks, like then you move into management.

Harry Stebbings 7:58 Okay. I- I appreciate that, and I think it's also important for people to understand that you can be comped appropriately. I think there's this like barrier in one's head that to break that barrier on comp, you have to become a manager.

Jeff Seibert 8:08 A hundred percent.

Harry Stebbings 8:09 Yeah.

Jeff Seibert 8:09 And this is something I felt strongly in for a long, long time. Google sort of pioneered this well in the early days and created this whole track for engineers to sort of keep climbing in- in comp and title and recognition and so on without taking on management roles, and I've tried to mimic that at all my companies.

Harry Stebbings 8:25 You mention

Harry Stebbings 8:26 accountability within CEO-ship and why like no one really can who could do it with the visibility they have, and then, you know, th- those that could won't because they don't have the visibility. So how do you create that accountability as a CEO?

Jeff Seibert 8:40 Yeah, it is- it's certainly not easy because you can constantly fall into a trap of thinking you're getting feedback and you're not.

Harry Stebbings 8:46 Mm-hmm.

Jeff Seibert 8:46 It's really how you set the culture of the company. So one of the things we do at Digits is we run the entire company on a weekly sprint. As part of that, every Friday, we end every week with a full team retro. And so- and we call it anchors and breezes. Anchors are what slowed you down, what didn't go well, what you need help with, like sort of feedback on the week. And then breezes are what went well, shout-outs to people who helped you, things you learned, et cetera, et cetera, and you create this culture of just constant iterative improvement which then allows sort of feedback conversations and one-on-ones and so on to be widely recognized by the company as like that's what we want. The whole mindset is just, how do we get 1% better each week?

Harry Stebbings 9:24 Can you do that at scale? I love the idea, but when you get to 100 people, does that still work?

Jeff Seibert 9:29 Yeah, it fractalizes. Uh, so what happens is each team will run their retro on Friday, and then surface sort of the highlights, like the biggest anchors or breezes to the full company-wide retro. And so everyone has sort of two opportunities. You do a team-wide thing, and then you do your own small team, and that's where you get into more detail.

Harry Stebbings 9:50 Do you like celebrating wins? I worry that it creates complacency. We've never won. I'm always chasing someone, we both are always paranoid. I hate this-

Jeff Seibert 9:57 No.

Harry Stebbings 9:57 ... like tap on the back. Uh, do you celebrate wins?

Jeff Seibert 10:01 We do. It's also- it's important to do it correctly, so I agree with your mentality. Crashlytics, I never thought was like successful in any one moment, not when we were acquired, not when we hit a billion MAU, et cetera, et cetera, like there's always the bigger goal. But if you have that mentality with the team, it's very demotivating, right? It's like what are we- what are we trying (laughs) to go to, like when are we gonna get somewhere? And so it's really important to celebrate small wins. And so we use this Friday, uh, show-and-tell we call it basically to show off what we did each week and champion who did what, and celebrate all the small wins of the week so people feel really connected to the company and what's happening.

Harry Stebbings 10:37 Yep. No, I know, I need to get better at it. All right. Um, (laughs) I love the way I still use this show. Despite its size, I still use it as like this like merciless testing ground for my own ideas. (laughs)

Jeff Seibert 10:47 (laughs) I love it. I love it.

Harry Stebbings 10:48 Yeah, um, tell me, obviously, I know this story being an investor, but you know, we have Crashlytics, then we have Twitter. How did Digits come through? What was that founding moment for you?

Jeff Seibert 10:57 Yeah, Digits really came out of the Crashlytics journey and so, you know, we got very lucky with market timing. We scaled from zero to 300 million phones in 12 months, got acquired by Twitter. Today, Crashlytics is on five or six billion MAU, roughly every active smartphone on Earth. It's incredible. And through that journey, I was struck by this dichotomy. So on the product side, you have real time analytics, performance monitoring, live dashboards, right? Like I knew exactly what was going on with the product and who was using it. And then on the finance side, I literally had a black and white PDF of my P&L and balance sheet once a month two to three weeks late that I didn't understand 'cause I didn't have a background in finance, I was an engineer, and so I was like, "What is happening?" And so literally that is why I started Digits, the- a simple premise, can we make accounting real time and intuitive for startup founders? And so what's crazy is it took us five years, but we finally just launched it, like it's actually here five years later.

Harry Stebbings 11:51 (laughs) And I mean, that five-year journey is one with twists and turns, and the idea that initially was Digits on founding day one is different in terms of the product that we're releasing today. What did you learn that led to your realization of the need to pivot, like why pivot and why was that enough?

Jeff Seibert 12:09 Yeah, it- it's definitely been quite a journey, and I know you've been on it with us for years now, so I've deeply appreciate it. Obviously, trying to make accounting real time is a lot easier said than done. Uh, when we started the company, we went heads down on R&D and really struggled with data quality for like three years. And that was because in 2018 when we started the tech to really automate bookkeeping, didn't fully exist. I think I was a little optimistic on how it could work, and so we have dozens of patents on it now but it was like a brick wall.Um, so in 2021, we made the decision to pivot from like the pure bookkeeping automation to collaboration tools, so better financial reporting, better client portals, better transaction review process,

Jeff Seibert 12:50 and- and that worked. We got a thousand accounting firms on the product, 5,000 downstream businesses. Like, that was sort of off and running. But what bugged me is that wasn't really why we started the company. Uh, uh, we had bigger ambitions. And so then, last year, literally all of a sudden, GPT-3 comes out, ChatGPT comes out, GPT-4, and we started experimenting, and we're like, "Whoa. Hold on. We're back." Like, "We can actually do what we set out to do." And so literally overnight, just like, we're back focused on this and spent this whole year building it.

Harry Stebbings 13:21 Can I ask you, how do you advise founders on when they have enough data to make that pivot? Because you don't wanna make it too quickly where it's like, "Whoa. Hold up, horsey." Right? "Not enough data."

Jeff Seibert 13:32 Yeah. Yeah.

Harry Stebbings 13:32 But also, you don't wanna be too slow. How do you know when you have enough data to make a decision?

Jeff Seibert 13:37 That is the million dollar question, and (laughs) uh, I'd say-

Harry Stebbings 13:41 (laughs)

Jeff Seibert 13:41 ... it's more of an art than a science. Like, you need to have a feel and an instinct as the founder of, like, do you see a path to success? And if sort of the window's closing on your path to success with your current business, like, that's to me when you have to go pivot. And a lot of super successful companies were hard pivots, right? Like, Twitter was a podcasting startup, Slack was a game, YouTube was a dating website.

Harry Stebbings 14:06 (laughs)

Jeff Seibert 14:06 Like, it is- it's totally crazy. And so to me, it's, like, impossible to say like, "Oh, no. Stop." Like, "That pivot's too far a field." Like, you have to, like, have this sort of founder instinct, and that's how great companies come to be. The one thing I'd modulate that with is, like, you need at least a year of cash left because if you don't have a year of cash, you're not gonna have time to see this pivot through. And so it's really like founder grit and enough cash. If those aren't there, return your capital. If those are there, I would go for it. That's how, like, huge opportunities come about.

Harry Stebbings 14:37 All right, a couple things that I always say, like do you have two to three experiments that you're still excited to run in this phase of the product? And if the answer's like, "No, I'm kind of out," then you-

Jeff Seibert 14:48 Yeah.

Harry Stebbings 14:48 ... you- you have a real understanding that actually that- that could be a sign. And then second, is 12 months enough? I don't mean to push you there, but I'm just intrigued on like if you think about it, you need to raise six months ahead of time.

Jeff Seibert 14:58 Yeah.

Harry Stebbings 14:58 That gives you six months to build and get enough traction to raise a- a price that's even a flat round to your loss.

Jeff Seibert 15:06 Yeah.

Harry Stebbings 15:06 Is that enough?

Jeff Seibert 15:07 It is- it is not easy-

Harry Stebbings 15:09 (laughs)

Jeff Seibert 15:09 ... but I think it's rare that you would have more than 12 months of cash because like usually you raise, uh, 18 months, and so by the time you figure out it's not going well and you need to pivot, it's 12. The big thing I would change from what you said is it's not an experiment. W- at no point where we're like, "We're gonna run two or three experiments." It is pivoting on a dime. You are all in on the new direction, and that is the only thing that matters to your success.

Harry Stebbings 15:31 So when you advise founders on the right way to pivot, what would you advise them knowing all that you do now?

Jeff Seibert 15:36 Yeah, so the key is getting your team onboard. If your team loses trust in you, you literally have no one to pivot, so it doesn't matter. And they can really sense the uncertainty, and so my- my other advice, as I've said, is like be very intentional and very decisive. And so both times we've pivoted Digits, I gathered our core leadership team, laid out, like, what are the challenges, what am I seeing, what are the options, and we knew within 24 hours what the new path was and what the priorities were. We were all in on that new direction. And so I- it comes back to conviction. Like, it is still a bet, but it's like, hey. Here is the information we have on the field. We need to make a decision right now because what kills companies is uncertainty. And if you sort of muddle your priorities and have one team try this and another team try this, no one's heart is in it, and both are gonna be mediocre. I would rather see founders take like one to the moon bet on one new direction, and it either works or doesn't.

Harry Stebbings 16:32 Do you agree with the idea of disagree and commit? I find it challenging. I don't think someone can fully commit to- to something that... and give their life to it-

Jeff Seibert 16:42 Yep.

Harry Stebbings 16:42 ... if they disagree.

Jeff Seibert 16:44 I- I agree with you. (laughs)

Harry Stebbings 16:46 (laughs)

Jeff Seibert 16:46 I will not disagree and commit. Um, I think this is the hallmark of great founders is you need to be able to convince your team and- and have the trust of your team to go all in on a new direction. Disagreeing and committing in that scenario is like, okay, you might as well step back, and like let's just have a smaller team and really focus on this 'cause it's not gonna be productive.

Harry Stebbings 17:05 I- I- I totally agree. Is there anything that you think are big mistakes that you see founders make when it comes to pivoting, either that you made or you see angel investments make when it comes to pivoting?

Jeff Seibert 17:15 Great question. Um, I think it's about this experiment thing, honestly. A bunch of angel investments I've made have tried to pivot, but I don't think they went all in on it. I think they saw it as a flyer that they'd try for a few months, and they didn't see it as life or death. And by the time they realized it wasn't really working, it was life or death 'cause they didn't have much cash left. And so you really need the conviction, like day of, to just sprint towards the new direction.

Harry Stebbings 17:39 I would also say, and this (laughs) desperately self-serving as an investor, but like as we said, the runway's crucial. If you're a great founder, if you pivot and it's unsure, most of the time your investors will back you just because they believe in you.

Jeff Seibert 17:53 Yep.

Harry Stebbings 17:53 If it's a surprise, and they're finding out through an article (laughs) or a tweet, uh, leaves a lot more doubt.

Jeff Seibert 18:01 Yes, 100%. And so yeah, next to your team is obviously keep the investors informed and up-to-date, particularly your board. And I will give a shout-out, so we're lucky to work with Peter Fenton at Benchmark. His clairvoyance on like where the large opportunity is and just relentlessly pushing us to find that and adjust, like, on a dime is super impressive and I think hard to do. Um, so definitely, like, keep your board tightly in the loop on this.

Harry Stebbings 18:28 ... gotta admit. I've got a man crush on Peter. I do.

Jeff Seibert 18:31 (laughs)

Harry Stebbings 18:31 I know. I- I- he knows it. I told him. My question to you is, what been your biggest lesson

Harry Stebbings 18:36 from working with Peter?

Jeff Seibert 18:37 The power of really deep intuition and conviction, and looking at a market from a very theoretical level. So one of the most interesting aspects when he originally agreed to do our A round, um, I sort of asked him why afterwards, like, why he committed so quickly. And he said, "Well, it had flashbacks to Uber, because when Uber was going against the taxi industry, the NPS scores on taxis were so bad." Right? Like, even if Uber was mediocre, it would still be way better. And he said accounting gave him the exact same vibes. Like, the- the status quo is just so bad that if you can make accounting, like, somewhat enjoyable, it doesn't even need to be delightful, you've already won. And so his ability to distill these markets into these, like, very high level, crisp, understandable talking points is super impressive.

Harry Stebbings 19:23 It'd be such an interesting, like, start-up selection criteria to only go after companies in sectors with the single lowest NPS's.

Jeff Seibert 19:33 Yeah.

Harry Stebbings 19:33 Like, let's you just map out the low... I'm sure you'd have insurance in there f- f- for one, health care-

Jeff Seibert 19:39 Yeah. Healthcare. Of course, there's other structural issues going after healthcare, but totally.

Harry Stebbings 19:43 But- but then you, but then you could actually isolate them and go like, "Okay, healthcare, no, too challenging," but like, actually, insurance underwriting may be way more addressable.

Jeff Seibert 19:51 Yeah. Totally, totally.

Harry Stebbings 19:53 Yeah. Um, you mentioned about kind of, uh, OpenAI and ChatGPT kind of... I don't know how to say it correctly, but, like, opening your eyes to the new possibilities that were available to you. When we think about that, I'm just intrigued, do you worry about a lot of your company being based on an external party's direction, development? It is unlike other times in that way.

Jeff Seibert 20:14 Yeah, this is such a special moment, and it... I'd say it vividly reminds me of the rise of mobile, which we wrote in Crashlytics from 2010 to 2013, the Web 2.0 transition in 2005 to 2007. It's like these platform shifts are so rare, and I love 'em because that's where all the massive opportunities arise. And so the key though, as a start-up founder, is to be in a position to rapidly adapt to that new reality, and you'd need to be faster than your competition. So, like, for five years now, I've run Digits on a weekly sprint. Every week, we get to decide what our direction is that week, and I have to say, that's been critical this year, keeping up with all (laughs) of the changes and evolution of the tech. So I'd say, I don't... No, I'm- I'm not particularly afraid of it. I love the energy of the tech world moving at a, at a huge pace. And so the key thing though is, keep in mind, this is technology. Like, we are viewing it as a tool. It's not life or death for the company. It's like databases and so on, it's just evolving faster. And so let's see how we use it, and it gives us more capability. But you ultimately need to stay focused on your customer and what problem you're solving, first and foremost.

Harry Stebbings 21:18 I mean, I'm putting the ball down for a fairway hit here. You're teeing me up, baby. You said there is kind of, like, databases or, like, a foundational technology that you build on top of. Will we see the commoditization of LLMs, do you think, Jeff?

Jeff Seibert 21:31 I certainly think we will. Um, and this may not be a popular position. Obviously, OpenAI is charging ahead, sort of leading the way right now. I think the market forces at work mean there's just immense energy to have an open source equivalent. Meta appears to be highly motivated to open source its work. Many folks wanna run these themselves and tune them themselves, and so on. That is hard and expensive today, but I can't think of another thing in time in history where something hard and expensive in tech has lasted all that long. Right? Like, it's going to be commoditized.

Harry Stebbings 22:02 Can I ask you, when we look at historical data on open versus closed systems, uh, there are many examples, whether it's, you know, your- your- your- your- kind of Linux to your, you know, Apple and Android. Traditionally, it's been the closed that wins.

Jeff Seibert 22:17 Yeah.

Harry Stebbings 22:17 Why is that potentially different today?

Jeff Seibert 22:19 I don't know if it's different. The- so the closed may win in terms of having the most advanced model, but I think the Apple versus Android comparison is exactly accurate. So you're gonna have something proprietary that might be best because it can be fully vertically integrated, they have- they control all the different variables, but I think there's gonna be an open source equivalent or more open equivalent that's a very close second. And for m- many people and for many use cases, it's just as good. Um, so what I'm really excited about is I can't wait to see who becomes the Android to OpenAI. Like, who is clearly the, like... solidifies the second tier?

Harry Stebbings 22:55 Well, what I'm kind of also interested in is, like, uh, yes, I agree with you slightly on the commoditization of LLMs, but I think also we'll see the specialization of LLMs for different things.

Jeff Seibert 23:05 Yes.

Harry Stebbings 23:05 And actually, if you're a creative tool, hallucinations are wonderful, but for digits, I don't want you hallucinating with my numbers. (laughs)

Jeff Seibert 23:12 Correct. Correct. (laughs) Yeah.

Harry Stebbings 23:13 I'm just kidding. You don't have a billion. Um, (laughs) it'd be a ch- sad day. My question to you would be like, you know, bluntly, do you think the best companies will leverage many at the same time or will it be one that we rely on?

Jeff Seibert 23:26 That is a really good question. Uh, so we built our ML team three years ago. We're training our own in-house models. You're right. Like, you can't hallucinate in finance. We've done a ton of work to make sure our math is always accurate. It's- it'll be really interesting to see. There's... Every model has different strengths. Like, if you already look at Bard and versus GPT and so on and so on, um, my sense is there'll be a very common popular open source sort of base LLM, and then tools to allow folks to fine-tune it easily. And what we've discovered is, while it crea- it takes so much time, data, energy, money to train an LLM, fine-tuning it can actually be done relatively straightforwardly with a surprisingly small amount of data, as long as your data's very high quality and focused. And so imagine you basically have your sort of default Linux operating system, right? And then every- everyone customizes their flavor of it for their product market, use case, whatever it might be.

Harry Stebbings 24:22 Got you. When you say about kind of, like, the data size not being as important as the data quality, I- I- I'm always kind of interested, like, how do we think about the importance of model size versus data size? 'Cause we are trained that data size is so important.

Jeff Seibert 24:36 It seems like...... that at the base LLM layer, the data size so far has been very correlated with performance.

Harry Stebbings 24:44 Hmm.

Jeff Seibert 24:44 And so, right, the bigger the models, the more data, the more parameters, et cetera, the better they do. Now they're starting to be this counterpush of like, "Okay, can we compress them? Can we pull that back?" Like, "How do we maintain the performance improvements without the size?" So I think that's a super interesting part of R&D. What I'm talking about is sort of the next tier of, how do you fine-tune the models? And that's where actually I think the quality of data's most important.

Harry Stebbings 25:06 I, I totally get you there. I also view it as like levers, which is like, you've got latency, you've got cost, and you have to have a trade-off like anything-

Jeff Seibert 25:13 Yep.

Harry Stebbings 25:13 ... on where you wanna perform and where you are happy to have some form of degradation.

Jeff Seibert 25:19 And look at Apple Silicon, right? They've done an amazing job. Like, yes, they're fast, but instead of pushing the bounds on pure compute, they're pushing the bounds on energy efficiency, which for Apple's use case is critical. And so I think there will be a lot of really interesting R&D on, how do you make these models maybe smaller and performant in certain use cases?

Harry Stebbings 25:37 You said about kind of, uh, individual cases where you'd fine-tune on top of a core foundation model. Uh, a lot of investors today, and, and dude, like honestly, 90% of companies that I see are like, "AI-4 Wealth Management," "AI-4 Podcast-to-Album Artwork Creation." I'm not dissing it, but like-

Jeff Seibert 25:54 Yep.

Harry Stebbings 25:55 ... it all... And it's like three weeks old and I'm like, "That can't have been that difficult to build." (laughs)

Jeff Seibert 26:00 Yes.

Harry Stebbings 26:01 Um, my question to you is, what's the difference between a thin layer on top of one of these models and a thick wrapper with inherent value?

Jeff Seibert 26:10 Yeah. This is the real problem you're hitting on. So startups are gonna get killed because they're very thin wrappers. And I agree with you, most of what I'm

Jeff Seibert 26:17 seeing on the angel side right now are these very thin wrappers on top of OpenAI. And like, if your primary product value is, is scripting GPT, that's a thin wrapper. If you've built it in two to four weeks, that's a thin wrapper. And so the, the key thing to me is like, we are in a hype cycle around AI, much like potentially the hype cycle with crypto and other tech before it. And all of these sort of fake use cases are gonna get quickly washed out and replaced and commoditized. It's really important to me that people view this as a technology. It is, it is like your MySQL database. MySQL was very popular 20 years ago, right? Like, it was super cool what it could do. And it's still cool, but if you try to go raise money on you've built a form on top of MySQL, you're not gonna be successful. And in five years, that's what a lot of these are gonna look like. You need to really focus on the market and, like, solving a core problem.

Harry Stebbings 27:09 I, I totally agree with you. So when we think about adv- like, founders today building in this environment, who's vulnerable then? Uh, like, uh, d- and what I mean by that is like, is it incumbents like Zendesk? Is it like high growth companies like, I don't know, your Notions of the world? Or is it your startups? Or is it all of them?

Jeff Seibert 27:27 It's probably more startups. The o- the one thing I would say is I view OpenAI probably evolving more into an infrastructure company like AWS. Like, they will host these models, they'll allow you to fine-tune them, they'll allow y- they'll give you all these base capabilities, uh, like you get with EC2 and S3 and so on and so on. The big companies, the big incumbents that will be able to leverage that tech, I would doubt if OpenAI goes and builds like, a Notion competitor or an HR or Salesforce competitor or so on. They probably wanna stay at the more generic level.

Harry Stebbings 27:56 Mm-hmm.

Jeff Seibert 27:56 From the startup side, a lot of startups are getting killed. And it, it's funny, I, I use the term Sherlocked. I'm an old school Mac programmer. Back in 2002, Apple killed Watson with its Sherlock tool-

Harry Stebbings 28:08 (laughs)

Jeff Seibert 28:08 ... and so the name sort of stuck. (laughs) Um, there's a lot of companies getting Sherlocked because they're pretty incremental and they're filling gaps in OpenAI's current product without realizing that like, yes, they're just on the roadmap. They haven't gotten there yet. And so if you're working on a use case that's pretty horizontal that, like, OpenAI is going to need to solve within five years in order to scale, that's not a great investment and that's not a good use of your time as a founder.

Harry Stebbings 28:33 One thing that I, I do think about, which I don't think people talk about enough, which is like, I completely understand the criticism of Google and Bard, but Google have access to compute and the ability to control compute pricing-

Jeff Seibert 28:44 Yeah.

Harry Stebbings 28:44 ... whereas OpenAI are at the whims of, you know, Jensen and NVIDIA-

Jeff Seibert 28:49 Yeah.

Harry Stebbings 28:49 ... uh, to his, you know, smiling appreciation right now. That's fundamentally challenging, no?

Jeff Seibert 28:54 It is very challenging. I... Like, the road ahead for OpenAI is not easy. And what very few people I think are paying attention to is Apple, because again, they control the silicon. And so Apple was caught flat-footed for sure. They're way behind at the moment. But imagine they're able to pioneer small models that run on-device, and then they do custom silicon to make 'em run, the performance could be outlandish compared to any other platform.

Harry Stebbings 29:20 Can you just help our-

Jeff Seibert 29:20 And so if you-

Harry Stebbings 29:21 Can you just help our listeners understand, uh, why is it important that it runs on-device and, you know, bluntly, with their own silicon?

Jeff Seibert 29:27 Yeah. So Apple of course is super focused on privacy. They don't want your data to leave the device. The only way to do that with AI is if you can fit an, a machine learning model on the device and keep all the data there. I bet Apple can and will. And so if you project forward five years, if they get to the point where they can run a sufficiently large LLM on your iPhone, then OpenAI is out of the picture. You don't, you n- don't even need to hit their servers. It's just on your phone.

Harry Stebbings 29:55 (laughs)

Jeff Seibert 29:56 (laughs)

Harry Stebbings 29:56 I hadn't thought about it. I'm like, uh, uh, "Are you my broker? Can we buy some more Apple?" (laughs)

Jeff Seibert 30:02 Yeah. (laughs)

Harry Stebbings 30:02 "I mean, can we load up on this one? Warren might have been right about them." (laughs)

Jeff Seibert 30:06 (laughs)

Harry Stebbings 30:06 I, I totally get you. Can I ask you, how do you think about enterprise adoption then? 'Cause I speak to some of the largest enterprises in the world, and they're like, "Dude, I'm not sending my customer data, my transaction data," and I'm not picking on OpenAI here, "but to a model that is outside of our bounds."

Jeff Seibert 30:23 Right.

Harry Stebbings 30:23 How do we think about enterprise control of very sensitive data in this world, where they want to get the benefits but don't wanna lose control?

Jeff Seibert 30:30 Yeah. Two different thoughts here. This is a really good question. So they had the exact same reaction to cloud. If you go back 10 years, they were like, "I would never put stuff on AWS or Google Cloud or Azure or whatever. That's ridiculous. Why would I share my data with those companies?" Right? And now it's just like, not only do they all do it, but it's actually better because those companies' core competency is running data centers. Most enterprises have no idea how to operate a data center.And so, I think this will go in the same direction. Um, and like, you'll, you'll have very clear guidelines around how the companies use the data for model training and it's off limits and so on, and sort of that trust will be overcome. The other angle though is there's also the danger of every enterprise jumping on AI because it's hot and cool. And sort of like they all jumped on blockchain for zero reason, even though it did nothing. Right? And like IBM's at fault. IBM was consulting, like charging for services to consult on how to adopt blockchain into your enterprise. That's ridiculous. And so, like again to me, focus on your customer, your market, your product need, and view this as a tool, not a panacea, and adopt it strategically on like what makes sense and where it's gonna push the product forward.

Harry Stebbings 31:42 I, I totally agree with you. So, uh, uh, you said about IBM there, you said how silly. I, I, I say how brilliant. That's called a monetization strategy in every environment.

Jeff Seibert 31:50 Yeah.

Harry Stebbings 31:50 I, I think that actually AI implementation services will be one of the biggest categories in the next few years. Do you agree with me or do you actually think that enterprises will adopt natively, it'll be fine? How do you think about that statement?

Jeff Seibert 32:04 I, I think you're probably right in that it's a large market. To me, it's not a very interesting market. Like yes, there'll be a lot of consulting to help enterprises adopt AI and the products won't be that great and they won't really make that much of a difference. What I see when you come to platform shifts is, again, this opportunity for a sort of a net new approach to really take over fast. And what gets me excited is like let's build real workflow automation for real people in massive industries that are outdated, like, for example, accounting. Like are you gonna trust Intuit to adopt this and totally disrupt their own product lines? Or do you think a new upstart's gonna come in and do it better? And that's where I'm excited to see 'em like pour energy into it.

Harry Stebbings 32:46 If we're totally honest, you have a couple more years in this game than me, Jeff. Uh, my, uh, my question to you is you've seen transitions like that with mobile.

Jeff Seibert 32:56 Yeah.

Harry Stebbings 32:56 You mentioned before databases. I'm always worried about speed of adoption and speed of transition and I always fear that it takes a lot longer than we think.

Jeff Seibert 33:04 Yes.

Harry Stebbings 33:05 But then actually I guess, you know, the rise of the iPhone and the App Store was actually really quite quick. When we think about the speed of transitions, will this be a slow transition or a faster transition than we give credit to?

Jeff Seibert 33:17 It'll be faster for a couple reasons. So if you look back at mobile, so the iPhone came out in 2007. They opened up the App Store in 2009. By 2011 to 2012, a lot of people were using and building apps. And then enterprise adoption even lagged from there, so call it five to seven years. With AI, there's no new hardware to buy. So you don't need this huge purchase price, right, locking out enterprises and people all around the world. You can just instantly benefit from it online, and there's no new UX pattern to get familiar with. Right? You don't need to be used to carrying something around in your pocket, looking at a little screen, squinting at reading things. Like there are chat bots. It's a very fluid interface. It does things for you. That makes sense. And so I actually think the adoption curve here will be radically faster and industries will be disrupted probably way quicker than prior, uh, tech waves, just because of the barriers are so low.

Harry Stebbings 34:11 You said there about kind of the disruption inherent there, one thing that I find people kind of don't understand as well is like everyone's like, "Oh, Google is so slow and behind." And it's like yes, but their golden goose is Search, which produces-

Jeff Seibert 34:24 Right.

Harry Stebbings 34:24 ... some, I think it's like 100 million a day or whatever it is. They're essentially having to cannibalize their cool golden goose. What w- what would you do if you... I, I asked Des Traynor this actually from Intercom, but what would you do if you were CEO of Google from here? You mentioned Apple's strong stance. What would you do-

Jeff Seibert 34:40 Yeah.

Harry Stebbings 34:40 ... if you were Google?

Jeff Seibert 34:41 They need to go all in on it. I don't think they have a choice. I, I agree with you. I think it's existential for them. Because if AI replaces Search, their golden goose has been killed. And so it is way more effective to kill your own golden goose than let and watch someone else do it. And again, I mean, going back to Apple, it reminds me of the iPod nano. Apple killed their most popular product, like actively killed it because they knew there was better tech coming. And I think Google needs to get bold and do the same.

Harry Stebbings 35:09 How do we feel about the cost of compute changing over time? You know, right now it's actually, you, we mentioned a kind of cannibalization, like, you know, cost of queries is significantly higher with, you know, models than it is for Search today.

Jeff Seibert 35:21 Yeah.

Harry Stebbings 35:21 How do we feel about cost of compute and cost of query in the next three years? Is it traditional Moore's Law? Is it faster? Is it slower?

Jeff Seibert 35:29 Yes. I, I'm admittedly not a hardware expert.

Harry Stebbings 35:32 Mm-hmm.

Jeff Seibert 35:32 Uh, so I don't know if it'll exactly follow Moore's Law or not. It'll certainly improve. What has been interesting to see is it's very clear the top models are memory-bound as well as CPU-bound. So you can't just make the CPUs, the GPUs faster. You need to increase memory bandwidth in, uh, sort of on par with that. And OpenAI actually shared a tech talk a few weeks ago that talks about how they were tuning their data centers like this. It's gonna be new challenges for NVIDIA, for Arm, for these chip companies to really unpack, but I mean, I would bet on the pace of technology. It will get a lot faster and a lot cheaper, uh, very, very fast.

Harry Stebbings 36:07 That's a relief 'cause it's fucking expensive now. Most of the VC dollars are just being shifted towards, you know, any of the large

Harry Stebbings 36:13 model providers.

Jeff Seibert 36:14 Yeah. Totally.

Harry Stebbings 36:15 Um, I, I do have to ask, speaking of kind of the price per query there, I think pricing is this undiscovered element of this next generation. Traditionally, we've had per-seat pricing in the world of SaaS. Will we continue in a world of per-seat pricing? Will it be consumption lab? Will it be project led? How do you see the future of pricing in an AI world?

Jeff Seibert 36:33 So this may be just me, but I, I very much see AI as a tool, not a product. And so it's a, it's a technology. It's like your database. It's like mem cache back in the day. And so because of that...... I don't think it'll change how people price in specific industries. Like if your market does per seat pricing, that'll probably stay. If your business and product does consumption pricing, that'll probably stay, and you'll have to work that into how you use the AI. I think it'll be commoditized and seen as technology within a couple years.

Harry Stebbings 37:04 Okay. So, you mentioned now the word commoditized. I'm really interested when it comes to the data itself. Sorry, my mind just jumps around. It's Friday evening, it's dark. Just-

Jeff Seibert 37:11 All good.

Harry Stebbings 37:12 ... just roll with me on this one, Jeff. You mentioned the challenge in terms of acquiring clean data earlier. How challenging do we think it is for companies today to acquire high quality, clean data? Is it as proprietary a defense mechanism as some suggest it is?

Jeff Seibert 37:28 It is extremely challenging. And what's interesting is the counterreaction, because you're seeing Reddit, Twitter, et cetera, shut off APIs, put in more strict, uh, rate limits, et cetera, et cetera. And so, the whole world is starting to lock down data, which was counter to the trends over the past 20 years, when everything was being pushed more and more open and API accessible and so on. And so, I think there's been a clear realization that the data is valuable, and that's one of the things, like, we've been really focused on at Digits is we have a proprietary dataset of 100 million financial transactions, and that's what we can train on and make sure our finance and bookkeeping AIs know what they're doing. Um, so I do think the data's really, really important.

Harry Stebbings 38:10 Can I ask, in terms of permissioning around training, for, you know, healthcare companies that, you know, have, uh, patient records for you, which has obviously financial records, do you need permissioning on an end client basis to be able to use that data for training?

Jeff Seibert 38:24 You do, and this is, uh, usually sort of broadly captured in the terms of service of ha- having access to your data to improve the product. And so, we do not share any of that data externally. We train our own models internally on the data, um, but that does allow us to improve the product for you and make your accounting better. Um, and so that's broadly how it works across the industry.

Harry Stebbings 38:45 I'm pretty sure that one day, Tim Cook's just gonna come and take one of my children-

Jeff Seibert 38:50 (laughs)

Harry Stebbings 38:50 ... and say that, "Actually, I- I- I ticked the, 'Do you agree?' in 2004, and now I've lost my kids." Um, uh, which, you never know, I might be quite grateful for, to be honest. (laughs)

Jeff Seibert 38:59 Well, let- let's step back for a moment, because it's not entirely a joke. So, Google Photos, potentially one of the most strategic products ever launched, because it allowed Google to collect the world's largest library of photos ever assembled, and that's how they were able to train their early vision models and so on.

Harry Stebbings 39:17 Huh. And they have... Sorry, I didn't, I wasn't aware of this. They're able to use the photos that you collect in Google Photos for their training models?

Jeff Seibert 39:24 Yeah, presumably. Yeah.

Harry Stebbings 39:26 Fuck. (laughs) That's impressive. Why doesn't Amazon just buy Anthropic? Like, I- I- I'm- I- I look at Amazon's play here and I'm like, it'd be quite an easy buy. I mean, especially at 4 billion, whatever it is, it's a minimal m- amount of market cap for them.

Jeff Seibert 39:39 Yeah.

Harry Stebbings 39:39 That seems like a wise acquisition, though. What would you do if you were in Amazon's place?

Jeff Seibert 39:43 That is super interesting. Uh, I- I would be very nervous about the OpenAI Azure partnership. And of course, Meta just made a big deal about partnering with Microsoft as well. So, I think you're right. I would look to aggressively move into the space and acquire something to bolster AWS.

Harry Stebbings 39:58 Can I ask, which incumbent do you think is vulnerable from the top incumbents in terms of like, you know, your Apple, Amazon, Facebook, Google, uh, of the world? Because we've just said Apple actually have a huge opportunity. They're behind, but huge opportunity.

Jeff Seibert 40:10 Yep.

Harry Stebbings 40:11 Who is vulnerable?

Jeff Seibert 40:12 I think Google's by far the most vulnerable, because again, their business model is pretty binary, right? Search is all their revenue, and so if that gets damaged, they're in a huge problem, and they've been slow to react. They've combined two different ML/AI teams. They've just punted Gemini into Q1, which tells me it's not doing very well. Uh, so I would be nervous.

Harry Stebbings 40:31 And then, sorry, final one and then I will move on from ... (laughs) I'm loving this. But like, every scale-up is introducing an AI product, whether it's your Notions, whether it's your Canvas, whether it's your Airtables.

Jeff Seibert 40:44 Yep.

Harry Stebbings 40:45 Are they all just kind of like jumping on the topic du jour?

Jeff Seibert 40:48 They really are, and one of the funnier ones to me is Dropbox. Now, Dropbox has a built-in AI, and I don't know, I mean, I- I'm sorry, Drew, but I- I just want Dropbox to store the files. I think there is a bit of a sort of bandwagon ride the hype wave aspect.

Harry Stebbings 41:04 Dropbox Dash.

Jeff Seibert 41:06 Yes, exactly, exactly. What is fun though is obviously it's a lot of experimentation, so all these companies are trying things out and seeing what works and what sticks.

Jeff Seibert 41:14 And so, the industry's gonna learn a ton over the next 12 to 18 months on like where it's appropriate to use AI and where it's sort of just a useless add-on.

Harry Stebbings 41:22 Elena Verna, who's the head of growth at Dropbox is, is in our team with 20 Growth, and so, uh, I'm gonna get absolutely killed for that. Uh, thanks, Jeff. (laughs)

Jeff Seibert 41:31 Sorry. (laughs) Uh-huh. Uh, no, listen, uh, it's great. Can I ask you, on- on the startup side, do you think 90% of investor dollars going into early stage companies today will go to zero? Oh, yes, likely. Um, isn't that always the case though? Has there been a period where that's not been the case?

Harry Stebbings 41:47 Yeah, I think so, actually. I think this is a gross over-exaggeration on the mortality rate of startups, which is like, going to zero is actually rarer than people give credit for. It could be half your money back, it could be a won X, whatever it is. But- but it is actually rarer that it's zero and a company's killed overnight.

Jeff Seibert 42:04 So, I- I actually have some data for you. I'm more than happy to share this publicly.

Harry Stebbings 42:07 Oh, yeah.

Jeff Seibert 42:07 Um, so since 2014, I have angel invested in 97 startups, uh, and I- I just did the count. So far, 30 have failed outright, so about a third. Another 19 are still at one X, so basically haven't gone anywhere. And if you look at the overall portfolio, really it's like 10 of them matter. Um, but what's crazy to me is, again, 2014, nine years of data, still the vast majority of the gains are on paper, and it depends how these 10 do on how successful the portfolio is.

Harry Stebbings 42:39 Uh, th- I mean, dude, you have just opened up treasure trove of tres- tre- questions for me here. Um, do you have cash back on many of them?

Jeff Seibert 42:47 Uh, no. Very few have returned. Uh, so a- a handful gave cash back, but... Or are you talking about on paper?

Harry Stebbings 42:53 Uh, no, I'm saying like actual cash back, like DPI.

Jeff Seibert 42:56 Oh, oh, oh, very, very few. Yeah, the, the time range on seed investing on the angel side is just, is like a decade plus.

Harry Stebbings 43:03 Okay. No, I get you there. Okay, so we have that. In terms of the top ten, don't kill me, but do you trust the book values?

Jeff Seibert 43:12 I do not, no. And I think a lot of them are basically holding at their 2021 valuation, whenever the last round they raised was, and it's nowhere close to the reality. I do get data from the secondary markets, and so I can see with some of them, like where they're trading versus where their last preferred round was, and it's a bloodbath. It's like down 80% for many of them. Um, so I think it'll be really interesting.

Harry Stebbings 43:37 What do you do if you're a founder who's sitting on a, a price that's just untenable?

Jeff Seibert 43:41 That is a really tough position to be in, and I think you, you don't wanna get there. Like my advice has always been be really disciplined about each round and what you give away and terms and so on. If you're in one of these boats, I would look to like recap a few and really focus on growth, trim expenses, lengthen runway. Um, it's, it's gonna be big shoes to fill into.

Harry Stebbings 44:04 Okay. So of those ten, would you do secondaries on any of them? Like how do you think about liquidity planning?

Jeff Seibert 44:10 I have actually. Um, so I've done secondaries over the time, and that's actually been my most successful outcome so far.

Harry Stebbings 44:18 Hmm.

Jeff Seibert 44:18 It's like waiting for them to sell. I've unfortunately had the opposite, uh, uh, happen where a couple of my early investments went on to IPO, and so were extremely successful, but then during the six-month lockup window, went almost to zero and I couldn't exit almost anything. Um, so I've actually had more success selling on the secondary market pre-IPO than I have had actually waiting the whole time.

Harry Stebbings 44:42 I mean, that is just the ultimate pain, isn't it, when you wait ten years, it IPOs-

Jeff Seibert 44:46 Yeah.

Harry Stebbings 44:46 ... and then you can't do shit, (laughs) and then it goes to zero. You're like, "Ah, ah." But-

Jeff Seibert 44:51 It, it is crazy. I, I mean, I understand why, but I don't understand why small-time angel investors should be locked up for six months. (laughs)

Harry Stebbings 44:58 I totally agree with you there on, on secondaries, and I think that's important. (laughs) Is there anything that you know now, angel investing-wise, that you wish you'd known when you started?

Jeff Seibert 45:07 Big lesson learned is no matter how great you think the company is, how great you think the market is, how great you think the founder is, it is still damn hard. Right? The odds of success are very low. And you can't get too cocky. So on two of them, actually two of my biggest failures, I was so convinced. I was like, "This is a no-brainer. This is gonna be a huge home run. I'm gonna put in way more than I usually do." Usually I try to stay pretty disciplined on check size, and both of them went to zero.

Harry Stebbings 45:36 Oh.

Jeff Seibert 45:37 And what can you do about it?

Harry Stebbings 45:39 Well, well, okay, well, okay, so what gave you the confidence in those companies?

Jeff Seibert 45:43 It was- seemed to be a very impressive founder, to me, a very obvious market. I won't name their names, but I was like, "Hey, just like solid execution here, should be a clear good outcome." And there's just too many variables at play over the course of the startup journey, and you don't know everything. Like as an angel, you don't do that much due diligence, so it's hard to fully understand. And so my advice would be just like, be very disciplined. It's like, do a bunch of deals 'cause you need a portfolio, and put the same amount in every time and average it out.

Harry Stebbings 46:11 I'm so with you. The idea that you have more conviction in one versus another at an early stage in particular, totally wrong. You also get known for writing a certain check, I find. Like, "Oh, Jeff's a 50K writer."

Jeff Seibert 46:22 Yeah.

Harry Stebbings 46:22 If you do 500 and then 50 or whatever it is, it's like, "Whoa." Um-

Jeff Seibert 46:26 Right.

Harry Stebbings 46:26 ... unless it's like an unwavering founder, like you really don't know. And so like, I mean this in the nicest way, you're in my unwavering founder box, which is like, you could be selling knives and I'd be like-

Jeff Seibert 46:36 (laughs) Nice.

Harry Stebbings 46:37 ... "I guess knife market could expand," like, you know?

Jeff Seibert 46:39 There we go. We'll make it happen.

Harry Stebbings 46:40 If Jeff believes in it, great. But when you are like you met them recently and they seem really smart, yeesh. Also, don't buy traction. And what I mean by that is like I, dude, I was in Clubhouse, I was in HOPIN.

Jeff Seibert 46:53 Yeah.

Harry Stebbings 46:53 I've been in some of the massive spikes. Traction doesn't mean sustainable.

Jeff Seibert 46:58 That is very key advice. Yes, totally agree. Um, on the consumer side, I'd say real traction on enterprise software, okay, potentially more sustainable.

Harry Stebbings 47:08 Totally. But it's generally less that.

Jeff Seibert 47:11 Yes, agree.

Harry Stebbings 47:11 You might get to 10 million ARR in 18 months, which is amazing, but it won't be like, uh, like, you know, 100 million units in a week.

Jeff Seibert 47:18 Right, it won't go vertical.

Harry Stebbings 47:18 Yeah, yeah, yeah.

Jeff Seibert 47:19 Yeah.

Harry Stebbings 47:19 What's been the biggest hit for you, and what's been the lesson from that?

Jeff Seibert 47:22 Oh, man. So okay, so this is all true. Uh, so this was ten years ago. A college friend of mine was prototyping what was gonna be a new teenage social network, and I invested 10K. I was like, "The guy is smart. Let's see what happens." He eventually pivoted that into Down to Lunch, which did well for a bit, and then failed unfortunately. Then he pivoted that into crypto, but not, uh, the coins. He started and launched Alchemy, which was like an Ethereum dev platform, dev tool thing. And at the height of the Bitcoin boom, I exited on the secondary market for 200X.

Harry Stebbings 47:58 (laughs)

Jeff Seibert 47:58 (laughs)

Harry Stebbings 48:01 Holy shit. (laughs) Um-

Jeff Seibert 48:03 So lesson learned there, nothing. It's all about (laughs) a founder's grit and mentality, and you can't back the early idea and think that that's gonna be it.

Harry Stebbings 48:12 Totally. Did that pay for all the other angel investments?

Jeff Seibert 48:15 Yeah, so that is why on paper, and like, and, and in terms of cash returned, I'm actually in pretty good shape. But yeah, it's remarkable how, how the outliers sort of outclassed the rest.

Harry Stebbings 48:26 But doesn't that also tell you one other takeaway, which is market timing?

Jeff Seibert 48:30 Yes, yes. And getting in very early at a very low valuation.

Harry Stebbings 48:34 Yeah, totally. But like greedy when others are fearful, and fearful when others are greedy.

Jeff Seibert 48:38 Yes, totally.

Harry Stebbings 48:39 I, I remember, I remember with Fost, do you remember this OneClick Checkouts company?

Jeff Seibert 48:43 Yes, of course.

Harry Stebbings 48:43 Yeah. I remember last year, uh, being, you know, being away with my family and I had this offer and I was like, "Ah." It's like, it was like a 50X on an angel check. I was like, "No, like I, I buy it. Like I'm, I'm gonna stay in here. Stripe have just come in. There's gonna be a partnership agreement there." No, I should have fucking been fearful. (laughs)

Jeff Seibert 49:02 Yep, yep. That's another great lesson learned for folks actually. So I had the opportunity to take some money off the table in a deal, so I got a 3X return. Right? Returned cash in the bank, 3X return. On paper, it's gone on to do another 10 to 15X, but......I'm worried that valuation's fake. I don't think it'll ever return the remainder. So, I only sold some of my shares, but that might be the only cash I get out of the deal.

Harry Stebbings 49:25 How do you feel about investors asking for cash back? Some people are like, "You can never do that. You believe in the founder when you invest. How dare you?" And others are like, "It's, it's entirely reasonable to say hey, Jeff, like, 'We both know this isn't working.'" How do you feel about doing that?

Jeff Seibert 49:39 Great question. I, so I am biased obviously by the fo- founder's side, but I think it's very scenario-dependent. So, if the founder is still, like, excited, the team is there, they have gri- like, I don't think you can ask for cash back. I think, like, you wanna bet on the founder, like, let's see where it goes? If the founder's clearly wavering, if there's performance concerns, if there's been a pattern of bad decisions, then maybe there's a, a clearer scenario and case for it. But in general, I would bias towards supporting the founder.

Harry Stebbings 50:08 I agree with you, for sure. I think sometimes founders actually feel like they just have to keep going. Do you know what I mean? And it's like a letdown-

Jeff Seibert 50:14 Yeah.

Harry Stebbings 50:14 ...for everyone if they don't. And so sometimes I feel like it's good to be explicit. O- one thing I have to ask about as well. Everyone has actually, I'm sure, had fraud in their portfolio in some way or another. I don't think we've seen the tip of the iceberg of it coming out. Do you agree or do you think I'm over-exaggerating?

Jeff Seibert 50:29 No. I, I think a lot of valuations will still crash from here dramatically. Um, and there's probably gonna be huge layoffs going into Q1 and Q2 as well. What'll be really interesting is the number of companies that raised and have sort of struggled through 2023, but now are down to six months or so of cash. I think next year could be very tough, 'cause they're not all gonna be able to raise follow-on rounds.

Harry Stebbings 50:51 What happens to talent migration? And what I mean by that is, like, you know, you have some amazing talent within large companies, and like high, high growth but highly valued companies.

Jeff Seibert 51:00 Yeah.

Harry Stebbings 51:00 Do they optimize for safety and stay in the well-paid job? Or do they go, "Fuck it. I'm at a company that's way overvalued. I should leave and do something." Which side do you err on?

Jeff Seibert 51:11 Yeah. This is a great question, because a lot of folks, you're right, are trapped in companies with high valuations where their options are likely under water.

Harry Stebbings 51:18 But it is safer than just starting something new.

Jeff Seibert 51:21 Right. And so it, it probably is safer than starting something new, but it's not safer than switching to a earlier stage sort of growing company where you're gonna have real options and a real impact. I actually love sort of down markets, because I think in the very hot markets, it's too easy to raise capital. And you peel off all the sort of like fantastic second engineers or product managers or designers or whoever they might be, and they go get funding and start something, and that actually sort of, uh, blurs, it stretches the talent density across the entire ecosystem. Versus in down markets, you can build higher talent density because it's harder to g- to raise your own money.

Harry Stebbings 51:58 I totally agree. So you think we'll see a migration from later stage to earlier stage?

Jeff Seibert 52:03 Yeah.

Harry Stebbings 52:03 Will comp be a line though? 'Cause that's the big problem. They, they just pay well.

Jeff Seibert 52:07 This is the problem, you're right. And that's probably the counterforce, is all of these big companies who raised huge rounds in 2021, have high salaries, and if you switch, you'll get more equity but way less cash. And so that's probably keeping folks where they are.

Harry Stebbings 52:20 And so will they stay? People like cash. People are trained that cash is now king, right?

Jeff Seibert 52:24 Exactly.

Harry Stebbings 52:24 You know, Ray Dalio's going, "Ah, cash ain't trash no more."

Jeff Seibert 52:27 It's very true. It's very true. Uh, so yeah. It, w- it's, you could argue both ways.

Harry Stebbings 52:33 Fa- final one, but like, and it's, it's a shit question. I can't believe I'm asking it. But like, you know, from the 97 angel investments, is there anything non-obvious that you see in terms of patterns from the most successful founders? And so like, I know it sounds strange, but one that I see is often the most successful founders moved a lot in early childhood. (laughs)

Jeff Seibert 52:54 Oh, fascinating. Interesting.

Harry Stebbings 52:56 And why, w- why, why does w- if you theorize around that, the theory is that actually you essentially have to constantly reinvent yourself to align to culture, society's friendship groups, and you're constantly understanding nuances, intricacies. That's the idea. But that's one that I've found.

Jeff Seibert 53:12 Oh man, no. I don't have a good of an insight as you. I'd say looking at my portfolio, it's not even necessarily the most techy founders that are successful, but the most sort of customer obsessed and just like really deeply personally understand the market and really wanna solve it. And usually have some pattern in their background, like their parents were in a sort of adjacent space or something where they've just felt this like personal identity come to bear on what they're doing. And I think that gives you like an extra push, an extra drive, an extra dr- grit to make it successful.

Harry Stebbings 53:44 I totally agree. Don't you love some founder stories though when it's like, you know, f- I don't know, they'll be doing like a food, food company. They'll be like, "And my grandparents ate, and so I feel it." (laughs)

Jeff Seibert 53:55 (laughs)

Harry Stebbings 53:55 And you're like, that's not like, you know, my grandfather was a car owner. Like, my grandfather-

Jeff Seibert 54:00 Mm-hmm.

Harry Stebbings 54:00 ...was, you know, a garage person. My father was-

Jeff Seibert 54:03 Right. (laughs)

Harry Stebbings 54:03 ...a guy. That you kind of get, but it's like-

Jeff Seibert 54:05 (laughs)

Harry Stebbings 54:06 ...really generic stories. We're like, "We're searching for some form of story." (laughs)

Jeff Seibert 54:09 Yeah, yeah. (laughs) That's the struggle of an origin story. (laughs)

Harry Stebbings 54:12 Yeah, yeah, yeah. Listen, dude. I wanna move into a quick fire. I've so enjoyed this. So I say a short statement, you give me your immediate thought. Sound okay?

Jeff Seibert 54:19 All right. Let's do this.

Harry Stebbings 54:20 So what do others not know that you know to be true?

Jeff Seibert 54:24 Runaway

Jeff Seibert 54:25 climate change is less than ten years out.

Harry Stebbings 54:27 What do you mean runaway climate change?

Jeff Seibert 54:28 Like, just a self-building cycle of higher temperatures and like, uh, cra- catastrophic storms and droughts and everything.

Harry Stebbings 54:37 Whoa. And so we just accept this new reality?

Jeff Seibert 54:41 No, I think we're too late. I funded climate change films 15 years ago, and it's still shocking to me that this is not like the singular top priority in any political campaign in the entire world.

Harry Stebbings 54:52 But, but then do we just accept this new reality of storms and unpredictable weather patterns and-

Jeff Seibert 54:57 Yeah. I don't, I (laughs) , I don't think you can't accept it. It's going to happen. And so the real question is what are we doing about it and which countries is it gonna completely impact way more than others? Um, so it'll be really challenging.

Harry Stebbings 55:09 Sorry, being a realist, I, I think it's something like if China hit its emissions goals for a year, it would be the same as Canada hitting them for 26. And might be-

Jeff Seibert 55:21 I think that's right.

Harry Stebbings 55:22 And so...... a- as much as we would like to, does it even matter if you don't get China on board?

Jeff Seibert 55:28 The impact is unfairly distributed and the cause is unfairly distributed, and that's why it's basically a tragedy of the commons. And so economically, it's like, how do you motivate that on a global scale?

Harry Stebbings 55:39 Do you not also feel slightly for merging economies in terms of like, you know, for years you've downtrodden us and beaten us to a terrible way of living, and finally we're increasing our, you know, standard of living, and now you wanna, you know, focus on other priorities, increase tariffs, make like ... F- f- we're still scaling the pole of starvation.

Jeff Seibert 56:00 Right. Exactly. And they haven't had the time to catch up and drive efficiency, and technology and so on. I ... The whole thing is unfair. Yes.

Harry Stebbings 56:09 I'm asking the tough questions. (laughs)

Jeff Seibert 56:10 (laughs) You are. It's turned into another quick fire round.

Harry Stebbings 56:13 Yeah, yeah. No, no. It's funny, man. Dude, you gave me such a good one now. Most people give me shit on that question. I'm like-

Jeff Seibert 56:18 (laughs)

Harry Stebbings 56:18 ... "All right. Well, this is good." Um, tell me, you can be CEO of any other company for a day. Which company would you be CEO of and why?

Jeff Seibert 56:25 All right. Well, th- this is for one day, so I would go be CEO of OpenAI just so I can see the roadmap, and then I'll, I'll know what to do from there.

Harry Stebbings 56:33 (laughs) Uh, next one. Most controversial view that you have today.

Jeff Seibert 56:38 Ah. So I would say actually, uh, on a good note, like A.I. won't replace many jobs, I think, actually. It'll drive productivity, not replacement, and it'll be like other technologies that have come out. Like phones didn't replace people. Now you just use 'em and you can do business from wherever. I think AI will let you get a lot more done in a lot less time.

Harry Stebbings 56:55 Why do you think the popular narrative is that it will? Is it because we like to be scared?

Jeff Seibert 57:00 Yeah. We like to be scared. We like to be terrified. So there's this concept in economics called the lump of labor fallacy. It was literally from the late-1800s. And you can look back. There's newspaper articles in The New York Times from the '30s being like, "We are gonna be replaced by robots, uh, within the next decade." And it's like, nope, that didn't happen, and it just marches forward. Every decade, there's a fear of the new technology coming in. People hated personal computers in the '70s 'cause they thought it was gonna take over and replace them. And so it's just, it's this fallacy.

Harry Stebbings 57:31 I, I also think like when you think about like seismic shifts to get your head around, can you imagine going from pen and paper to computer? I know, I know it was a 20-year period, but like, that really would seem like we are fucked. (laughs)

Jeff Seibert 57:43 Can you imagine being a math teacher back in the day with the advent of calculators? You would think your entire job is useless.

Harry Stebbings 57:50 It is useless. (laughs)

Jeff Seibert 57:53 (laughs)

Harry Stebbings 57:53 I hate to say it. I, I mean, I was taught early, those that can, do. Those that can't, teach. (laughs)

Jeff Seibert 57:58 (laughs)

Harry Stebbings 57:59 But, uh, that's why I was kicked out of most schools I went to.

Jeff Seibert 58:02 There you go. (laughs)

Harry Stebbings 58:03 Um, uh, but what's the right way to view competition, Jeff?

Jeff Seibert 58:05 Ignore them completely. Focus on the customer.

Harry Stebbings 58:08 Really? I'm always like, ah, it's helpful just to be aware of the other boat, you know?

Jeff Seibert 58:14 I don't think so. It's like, be really aware of the customer pulling you and like make sure you're going in their direction as fast as possible, a- as a startup. Maybe if you're a, a giant Goliath, you should start paying attention to some competitors. But as a startup, the market is way larger than you could possibly capture this year. So like focus on your customers. I think a lot of product managers get too distracted watching the competition, and they don't have a vision for their own product.

Harry Stebbings 58:41 What's the best piece of advice you've been given?

Jeff Seibert 58:44 24-hour hypothesis. So when your team comes to you with a question, you need a decisive answer within 24 hours. And if, if you're slower than that, you're not moving fast enough.

Harry Stebbings 58:56 What if it truly needs to be incredibly thoughtful? Like the-

Jeff Seibert 58:59 So-

Harry Stebbings 58:59 ... strategic direction of a company or really like all structure redesign?

Jeff Seibert 59:04 Yeah. So this is where I go back to one of Bezos' letters, type one versus type two decisions. The vast majority are type two decisions. They're reversible. So you can undo them. Uh, what I would say it is on net, it is way better for a startup to be moving in a direction than the absolutely perfect direction because you can keep refining that going forward. But if you pause on a decision and you're like, "Give me a week," what do you, what do you want the team to do? And this was one of the biggest challenges Twitter faced internally back in the day, was just complete indecision from leadership.

Harry Stebbings 59:37 I, I, I think we constantly underestimate the value of just activity. Do you know what I mean?

Jeff Seibert 59:42 Yeah.

Harry Stebbings 59:42 Which is like quite often, you know, I, I look back, but like the early days of doing these shows, I wasn't so good. The topics weren't so great. The questions weren't so great. (laughs) But I learned and I iterated. The shows developed, evolved, guests got better, worse, what ... But like-

Jeff Seibert 59:55 Right.

Harry Stebbings 59:55 ... activity drives progress.

Jeff Seibert 59:58 100%. And learning. If you're not moving, you're not learning. So what are you gonna do differently?

Harry Stebbings 1:00:02 Great. So if you're not moving, you're not learning. What did you believe 12 months ago that you no longer believe?

Jeff Seibert 1:00:08 (laughs) 12 months ago, I thought useful AI would be years away.

Harry Stebbings 1:00:11 (laughs)

Jeff Seibert 1:00:12 No longer think that. (laughs)

Harry Stebbings 1:00:14 Yeah. It, it is amazing to think of the speed, huh? Like it's not even been like a year since the evolution of like the next gen.

Jeff Seibert 1:00:20 Totally. Totally.

Harry Stebbings 1:00:21 Yeah. Uh-

Jeff Seibert 1:00:21 It's crazy how fast.

Harry Stebbings 1:00:23 Jeff, 10 years, you know ... Or I mean, we met like eight years ago, I think it was.

Jeff Seibert 1:00:27 Yep.

Harry Stebbings 1:00:27 So if we take like that and add two years, so 2033, where are you then?

Jeff Seibert 1:00:32 So keep in mind, there's gonna be some runaway climate change. So I'm gonna be off growing obscure wine grape varietals in a cold climate.

Harry Stebbings 1:00:40 (laughs) Where will Digits be?

Jeff Seibert 1:00:45 Uh, hopefully Digits will be the new de facto accounting platform. We're really, really excited. We've built basically a completely object-oriented, real-time, AI-driven approach to finance to make it really intuitive for startup founders to understand their finances as the business happens, not three weeks late on a black and white PDF.

Harry Stebbings 1:01:05 Jeff, I've absolutely loved this. This has gone in so many great directions. Um, thank you so much for being so brilliant, and it's been the highlight of my week.

Jeff Seibert 1:01:14 I love it, Harry. This was super fun. Thanks so much for having me on.