Is investing REALLY the hardest job in tech? A TWIST VC Roundtable (ft. Deedy Das and Jay Eum) | E2204

27:09And that, I think, is the highest multiple fund on cash that Sequoia ever did. And now with secondary markets and the ability to swap investors out in years 12, 13, 14, like SpaceX and Stripe are doing, it is creating some liquidity for venture, but you would be much better served to have it actually be public companies and distribute to your LPs the public shares. That's what LPs really want is to get those public shares and then make their own decision of what to do with them. So it's an industry in transition. Great founders are coming from anywhere, to Jay's point. It really is a global industry.

27:51And yeah, these industries are being jump-started. But if you look at the Scandinavian countries, Sweden specifically, Israel obviously, and then you go to Australia with Canva and Atlassian, now you're starting to see DECACORN. come out of these regions. It's not just like the occasional unicorn and like, oh, Spotify's worth a billion dollars back in the day. It's now Spotify's, I believe, worth over a hundred billion dollars. 128. Yeah, so you got a hundred billion dollar company coming out of a country of 30 million people. Atlassian is also up in that range. I don't know off the top of my head where they're at, but then Canvas at 40 billion, I think, in the private markets.

28:34These are not small companies anymore. Talent can come from anywhere. Great founders come from everywhere specifically. They used to land on our shores. We used to welcome them here a lot more than we do at the current moment, even with Trump saying he's going to put those green cards. I'm still going to follow up with him about that, put those green cards on the diplomas. but yeah these companies are being formed all around the world and the secrets are out you know just being here in Saudi for the last three days they're listening to our podcasts they're reading our news sources they're on the sub stacks they know the techniques they're asking me about Raul's product market fit engine blog post and the talk he gave three years ago on this podcast Like, they're up on it.

29:19So the secrets are out, and talent comes from anywhere. So there's a very bright future here. We did have a four-year pause under Lena Kahn where there was no M &A. And I think when that indigestion gets cleared and we can kind of move some of the companies that should have been bought and consolidated and get more DPI, I think people feel a lot different about the industry. But I know a lot of venture capitalists now who are just saying, this is too hard. This is too hard. I'm quitting. And that's not the case with Ruloff, I don't think. But and they were succeeding. But we all know on this call, many folks who are not going to raise their next fund.

30:02They're just like, this has been too hard. There's no DPI. In fact, if you want to get rich, it's not the perfect way to get rich. It used to be in people's mind. Oh, I'll become a venture capitalist. I'll be a billionaire. No, venture capitalists are not becoming billionaires. Hey startup founders. Are you tired of losing deals to endless procurement cycles and complex contract negotiations? There's a better way to sell your SaaS product. AWS Marketplace isn't just another sales channel. It's your shortcut to enterprise customers. When you list on AWS Marketplace, your software charges appear directly on your customer's existing AWS bill.

30:36No separate invoicing, no procurement headaches, no waiting months for approvals. But here's the best part. AWS actually incentivizes buyers to purchase from your listing with credits, helping you land your first 20 customers. And after that, AWS provides you with a budget to fund even more deals. Plus, you get one-click deployment, free trial capabilities, and access to committed spend budgets that enterprises are already planning to use. So stop watching your competitors close deals faster. Join thousands of successful startups already selling on AWS Marketplace. Ready to accelerate your sales cycle?

31:12Visit AWS Marketplace today and turn your biggest procurement obstacle into your biggest advantage.

31:23Their funds do two or three acts. They got to chop up the, you know, funds, you know, with four partners, five partners. They're not actually bringing down as much money as you might think because they own such a small percentage of startups. It used to be the Series A, the seed round, they would get 15%, 20 % of a company. Now the seed round is 5%, 6%, 7 % of the company. The Series A might be only 10 % of the company. It's an industry in transition. Yeah, for sure. I mean, one thing I would add, though, is despite there's more competition, there's more challenges for sure. Some of these companies are growing at a rate that I've never seen in my 20 years.

32:02I mean, it's just the potential outcomes of some of these companies where everybody knows that Lovable hit 100 million ARR in less than 12 months. We have other companies that are trying to beat that record right now. And so if you extrapolate some of these companies and the growth rates that they have, and again, the good news is a lot of these AI, especially AI application, AI platform companies, they're growing on existing infrastructure that has been built for decades. Connectivity is there, mobile app distribution, all of those things are in place. And on top of that, more recently, there's actually a willingness to pay, shockingly.

32:44People understand that there is actually a cost for every inference that goes into it. And, you know, thanks to OpenAI and others, they've actually been able to be trained to actually pay for the utility that they get from some of these applications. So I've never seen anything like it in my career. I think that last point Jay made is I've never heard anybody make it actually so crisply, but I think it's so important to think about. We all think, oh, cloud computing, the infrastructures, the rails, the bandwidth, sure, all that's set up. But what's also been set up is that businesses had such a great experience paying$100 a month or$500 a month for a piece of SaaS software that they're like, yeah, this thing's as good as Salesforce or Slack or HubSpot.

33:27It is so worth it. Yeah. It's so worth it. And so when all these AI tools came out and there was like$40 a person to be on Claude for the enterprise account, I was like, yeah. And, oh, Grok wants to charge us too? Bring it on. Oh, there's a$200 version? absolutely i'll pay for some people on the team to have the 200 version if it makes them one percent more efficient if a developer wants cursor and co-pilot and five versions of it and they use two a month businesses are just very frisky about that and they understand i think that making existing high performers higher performers is better than hiring more people right so that's just to build on your exceptional point jay that is also part of it that goes into why these startups are doing so well.

34:12Do we hire another designer or is there a design tool that this designer can use that makes them go faster on Figma? Do we hire the sixth developer or just give the other five another tool? It's a pretty easy decision for management. I think Jason's talking about the B2B case here, Jay, but when I was thinking about the company you might be discussing that's trying to race lovable to 100 million ARR, I was thinking of Higgsfield, which recently raised and is growing quite quickly. That's B2C, so a slightly different point, I think. Are you seeing it the same? It's not. So certainly there's a B2B customer base, B2C customer base for sure.

34:47They fall into the category of some very similar to Canva. They actually sell for short-form video creation. And this is our co-investment with Menlo, obviously. They actually sell to creators and they do short-form video generation for social media, right? So if you think about who the major customers are, including the Mr. Beast of the world and whatnot these are not individuals per se right these are actually small enterprises some of them are quite big and they are looking for platforms that actually generate you know high quality realistic video so that they can actually supplement their own efforts to actually have these studio setups camera setups and whatnot and they've done a fantastic job they have grown and this is since they tweeted themselves public information they've grown 60x in the last six months i've never seen anything like it it's crazy 60x so one million arr to over 70 plus in a little over six months so jay one thing jason likes to talk about though is there is sampling there is testing people have to play with things i'm curious if you can share anything about the revenue health at that company?

36:01Is that ARR sticking around like we would expect traditional enterprise SaaS ARR? Is it a little bit more loose? What's it like? Yeah, it's too early to say, but it's certainly somewhere in between. So it's not like in a B2C revenue where people try something out and the minute they get a better deal or a better feature from somewhere else, they'll drop it and move on. It's not as sticky, obviously, as enterprise revenue, but the company is actually moving into that category as well. But again, the canvas segment of what we call prosumers is a massive segment. And they are sticky. They are loyal.

36:37If they find a platform that works, why would they switch over? So, yeah, they're actually accelerating revenue to a rate that I've never seen in my career before. Didi, I want to get you to weigh in on the willingness to pay in the AI era, because I think this is a major tailwind for startups of all sizes. I mean, look, I think Jay said exactly what's going on and Jason said it too. There is immense willingness to pay and try new things. When it comes to thinking about growth versus retention is often a thing that people in venture talk about. And I think the point that's really missing here is you can grow today in a product-led motion faster than ever before.

37:19And if you actually compare those retention cohorts to enterprise businesses, you're not going to see the same kind of retention. But what you actually see in great businesses is you end up seeing a smile curve. And my favorite graph for this is, you know, I've treated this multiple, multiple times, is you look at the ChatGPT retention cohorts. If you looked at, there's this great article, funnily enough, this is a great article Sequoia put out a long time ago, which is talking about generative AI Act 1, or I think it was Act 2, I'm not sure. It's they said, you know, the growth is here, but the retention is not.

37:53If you compare one month retention on ChatGPT to any other product in the market, say like classic YouTube or a WhatsApp, it's far, far from that retention. It was something like a 40 percent, maybe less, whereas other products are at 80 to 90. And then if you look at the cohort retention curves over time, holy hell, has that changed? Like it has become a 90 percent retention product over time. Yeah. So, you know, the two common things I read and I see all the time is people, they whine about retention and they whine about margins. Growth, I think, is way, way, way harder than either fixing retention or margin.

38:29So as an investor, I am very comfortable, you know, backing companies who are able to grow and then think about how we can figure out the retention and the margin story. And we're seeing that at these companies, like people complain about cursor margins all the time. But you know what? They're doing some interesting stuff. The likelihood that Cursor fixes their margin structures, the likelihood that Anthropic fixes their margin structures is far more than another company growing into tens and hundreds of billions of dollars of revenue scale. So I'll put it that way. Also, I think the retention numbers matter a little bit less when the top of your funnel is 800 million weekly active users.

39:11I mean, you're not selling to 15 people, so you need to keep 10 of them. You have the whole world, essentially, as a potential customer. And OpenAI has done a lot of work in India, Didi, to build out ChatGPT Go. And then they also launched a Indian language model benchmark recently, which I felt was indicative of where we're seeing a lot of AI demand, which would be, in this case, India. India is an interesting story, especially when it relates to ChatGPT. I'm really, really curious how this plays out because, you know, I'd worked at Google prior to this and we were focused on a lot of the India efforts.

39:41And one thing that's really, really surprising, was surprising to me then is the India Internet view is very different from what the American view is. So, for example, it was kind of like I took it for granted that search is always the biggest Google product. And then there's, you know, maybe a YouTube, the Maps, things like that. in India, YouTube was far, far ahead of search as a product. In fact, people didn't see search as the same kind of valuable utility, even though it was a really high number, as people in the West would. And so I think what's really going to be interesting for ChatGPT is people are adopting it, but language and typing out queries is not fundamentally as widespread in that market as it is in, say, an American market.

40:27And so it'll be super interesting to see how that plays out. I think video, audio, these sorts of multiple languages are going to be a really interesting phenomenon when it comes to AI in India. And also, willingness to pay for software in India is far, far less than in America. So we'll see. Well, I mean, that'll change with time, but I do have the chart you were talking about. Here is the smile curve from ChatGPT. And Didi, just be a VC for me for a second. When you see this, what's the first thing you notice? And talk to people as if they can't see the chart because a lot of folks are on audio as they hear this.

40:57Okay, so let me try to describe this. when I had tweeted this out a while ago, I call this the product manager's wet dream. I think this is absolutely, when you're a product manager, you look at things like cohort retention charts. What a cohort retention chart is, and I'm not going to put all the caveats in here, is if I have 100 % of users on day zero for a given week, how many of those users return to pay me or return to the product, there's two kinds, on week two, three, four, and down the line. Now, when you look at those weeks stacked up over time, what you're really looking for, both as a PM and as an investor, is you want the curves to look like they're going up.

41:38You want more retention. So for month one, if you're looking at this graph, it used to be a 60 % at month one. And you see this go up over time for every incremental week. And now it's about 90%. And this entire graph, not only does it sort of decay over time, but then you see what's called a smile curve, which is people actually come back and pay more. So after they may have churned, they may have canceled, they come back and they go like the same user comes back in maybe month 10 and goes like, you know what? I missed that. I need ChatGPT again. I'm going to go and pay more. And a smile curve is the most beautiful one line description of I have product market fit.

42:19people will turn and go like nah just kidding I'm back um and that's what's beautiful about um that graph it goes up into the right up into the left in this case actually and that was for Gemini product this was for chat GPT um we have a Gemini right chart here just no no it's fine I'm just yeah it's just impressive to see that the brand is so strong it's this would go back to the sampling thing we were talking about earlier people like will start using a product and then they forget they used it uh and then six months later a year later they go to it and it's like oh you already have an account oh yeah there was the project i worked on yet now i'm ready to pay for it because the product has proven itself that is the one strange thing about this moment in time in this technology wave is that so many of the um products are providing so much value that people will pay for them.

43:13And to think like, you know, prosumers, you know, content creators are willing to just buy 10 tools a month for 15 to$150 each, like it's just 10 years ago, it was like pulling teeth to get them to pay for anything. They were just like so cheap, they you know, I'll just figure out a way to do it for free. I'll use open source software if I have to whatever. And now they're just like, yeah, I'm making money, I need to I need the best possible tool. I'm not trying to save money. I'm trying to make better product.

43:59I want to bring up something that we're hearing from a lot of launch founders in the accelerator or at founder you. There's a lot of reports that the bar for pre-seat is just getting higher and higher and higher. And so people are kind of curious for you guys to answer what is a reasonable set of venture expectations for pre-seed companies. And people are particularly asking why is there so much of a revenue expectation for companies that used to be at that stage, a couple of founders and an idea. And Jay, I want to start with you on this one, current expectations and why things have changed. Yeah, it's interesting.

44:32I think this whole AI boom has masked the fact that it's actually a tough, tough environment out there for everybody else. And so if you think about kind of a lot of the capital that kind of rushed in into the industry, and these include non-traditional sources from private equity firms to sovereign wealth funds and whatnot that were trying to get into the act, and that kind of culminated at the peak of the bubble in 2021, a lot of that money has actually kind of gone away. And so there's actually, as Jason pointed out earlier, a lot of the seed funds or pre-seed funds that were actually funding those kind of opportunities, they're having trouble raising their funds, right?

45:17So they're having their challenges. Not a lot of pre-seed seed funds are actually actively investing as they were just two or three years ago. And so obviously the funding environment for those startups that are just coming out and pre-seed is obviously very, very challenging. And so, you know, even for us, when we launched in 2021, the quality of the companies that we're seeing then versus what we're seeing today and the quality of the deals in terms of bang for the buck valuations that we can get, it's so much better now. I mean, we have invested alongside with, you know, you guys at launch in a profitable 1 million AR company at sub 20 million post-money valuation, right?

46:00And this company has grown 5X in the last several months. So there's some really, really good companies out there. And so if you're an investor and you have the capital, you're going to be picky. You're going to pick and choose the companies that have great product market fit, great background of the founders that are actually showing traction. And I don't necessarily have to take the risk of free revenue opportunities because I can actually validate my due diligence, basically, the market does for me. And I love it. Yeah. it's like there's zero to one and then there's zero to one million. These are like two different proof points, like zero to one, you know, you got one customer or whatever, or you launch the product like zero to one million.

46:39You can't really fake. Right. And you are a function of your opportunities. So if there's no companies, you know, in the seat stage, knocking on your door with a million dollars in revenue, well, by definition, you're going to invest in them to help them get to a million. But if there's 20 who've knocked on your door this week, and they all have between a half million and 2 million, well, now you're going to make the best choice out of that group. And I've been trying to explain this to founders who were lollygagging, who were taking their time, who didn't have a sense of urgency. And then they met reality.

47:15And they said, well, I don't understand. I was able to raise money for my last startup, and I'm further along this time. And I'm like, you are assuming that you're competing against your former self. You're competing against another group of founders who are coming, who have a company that went to$60 million in revenue in this amount of time. Sure, the valuation is much higher, but they've eliminated a ton of risk. So it's very much like the public markets today. If you've got a company that has very low margins and very slow growth, and you're up against Uber or Coinbase or Google or Facebook or Google or Amazon.

47:53And people are looking at those companies and go, wow, these companies are growing revenue on big numbers with great products and they have incredible margins and they have momentum and network effects. There's a flight to quality. That same kind of thing has happened even as early as seed now and series A. Yeah. I want to jump in here and maybe say like, there's almost like there's two stories of founders really. And if you look at the data, it becomes super clear, which is the volume of capital deployed in venture is at the highest it's ever been. But the number of deals that that capital is chasing is actually shrinking right now.

48:29I don't know the exact number, but on one hand, you have founders who are like, holy shit, stop knocking down my door trying to fund me. And on the other hand, especially at the earlier stage where the metrics are fewer and far between, where founders are like, no, I think I'm awesome. I think this product is working. Why is nobody taking my call? So yeah, I think the data does support that as well. All right, just backing up Didi really quick here with some data. Here's a chart from the KPMG Q3 Venture Pulse. And as you can see here, I'll just sportscast a little bit. It's a chart showing how much money VCs invest and then the deal count over time.

49:06And deal count peaked in late 21, early 22, and it's been kind of a secular decline since then even though capital has returned to the market so we're seeing a return of dollars dd but not a return of deals which i presume that means your no percentage is going up over at menlo yeah um i mean look menlo specifically is um there's another beautiful graph peter walker put out of on like you know fund returns based on the strategy whether it's a spray or pray or a select fund and a selective fund. And I think Menlo is still very traditional that way. Like we like to pick, we like to do lower volume.

49:43We like to work and spend time with our founders. This is why we do lower volume deals. The reality as like basic eighth graders in math will tell you is if you do a more concentrated strategy, the variance is much higher. If you do a spray and pray strategy, the variance is lower. What is not so obvious is typically, I believe that what that chart said too, is a spray and pray strategy has a lower median return than a concentrated strategy. So obviously at Menlo, we hope that we're not at that median. We hope like we're definitely an outlier when we're doing those select deals. But man, it's not easy, right?

50:18Like everyone is chasing those select founders. So how do you go and actually prove that you can find them and then add value is still very tough. And I think, Jay, you have to make sure you don't overpay. Absolutely. Just to explain to founders who are listening here exactly how hard our job is, then founders want to break the record for their accelerator and get the highest. They want to get$1 million more than the last founder said at the poker game they got, and they're in a competition, which means then we own a smaller amount, which means that on the exit, we get less to return. And so the game becomes harder and harder in terms of eventually getting a return for your LPs.

51:02So it's, I would say, perhaps the hardest job in the ecosystem is being an investor. I think it actually is. Now, you can't cry for them because they get paid huge salaries. You know, and it is a pretty great job when you have the diversification of 10 bets or 30 bets in your fund, whatever it is. but you don't have to sit there grinding your teeth at night, but this is the best time possible to be an entrepreneur who can execute and who's revenue focused. If you can grow your company, you know, I was telling the founders in, you know, when I was out here on my trip, they said, who gets funded?

51:38I said, the person who's growing 20 % a month, person who's growing 5 % a month today, because there's so many deals to look at that people are starting, you have a choice between which high growth company you want to invest in. That's a very different world than it was, you know, 10, 20 years ago. But Jason, explain to me how you said earlier that we're seeing VCs kind of give up and walk away because maybe this isn't the place for them. And then, you know, ever since we've had that point from you, we've talked about how there's so many great companies to invest in. They're growing faster. AI payment tailwinds.

52:10Explain the disconnect between those two points for me. You know, it's just classic supply demand. There were far, far too many VCs. And we saw that break when the private equity folks, as Jay referenced earlier, came down and they started competing and they didn't do due diligence. They just said, oh, if any of these firms do a deal, we'll just do, we'll assume that that company's been successful and we'll do their next round. Just think about how crazy that is. That assumes that, you know, if our firm has a good track record or DD or J's have good track records. There's a person coming in and just blindly betting on that when we all know that we make mistakes, that we are far from perfect in our decision making.

52:53In fact, if you ask most VCs, which company did you think in your portfolio was going to be the breakout and then which one actually broke out, they wouldn't be able to tell you at that time in the first year. So there was just far too much supply of capital. And there's still a large number of companies. That's true as well. But if you're going to put larger amounts into the winners, that means there's less leftover for the rest of the kids in the class, which if you think about it in Darwinian sense, if this was an ecosystem and these are all different species, what we're now seeing is the species of founders who can generate revenue are going to get a lot more of the calories in the ecosystem.

53:41They're the sharks in this. They're the apex predators. So if you can't get revenue, you know, you just may not be able to get the calories. I mean, one way to think about it is remember we had this unprecedented bull market leading up to 2021, right? And if you were, let's say, an entrepreneur turned angel investor that was betting on YC companies that were almost guaranteed to raise the next round of financing and you would have automatic write-ups, it was a good time to play. And those results culminated in launching a micro VC fund, right? And life was great. Life was easy. Just continue betting on those companies, right?

54:20And now the world has changed. We're living in a different world. You know, there was a time where, you know, the tide would raise old boats. It's become a lot more difficult now. And so the ones that actually have experience, have the data pattern recognition, have a reputation with the entrepreneurs are still going to be able to find those opportunities. But if not, and you've been part of that crowd and you've just been kind of riding that wave, it's a lot harder. So those are the folks Jason, I think, refers to that are not going to be able to raise their next fund because they haven't returned capital.

54:53Yeah. And they weren't doing, they were blind betting in a lot of cases they just had a network they bet they thought they were experts at this because the markups happened exactly but then the fundamentals weren't there and then they did all of their bets if their micro fund because i was lps and some of them they might have done all their bets in an 18 month window when the mark the valuations were peaking so there was no time um diversification yeah no time diversification thank you jay and so that was one of the things i learned from Ruloff and Doug Leone and Bill Gurley, the people who kind of mentored me when I was coming up, is like, hey, take your time.

55:29You don't have to put it all to work in 18 months. You take, if you deploy over three or four years, which is typically how funds used to work, four years, not two or 18 months, or some people would deploy in a year, you would have some time diversification as well. And they didn't have that. And so here we are, folks. Yeah. This game is not easy. And we all know one of the primary drivers of, you know, venture capital return is really timing, right? You have to be in the game at the right vintage so that you can exit at the peak and go in low. And it's just the way it is. Yeah. Jared, I actually wanted to talk to you about entry prices because everything that I'm seeing from data sources indicates that we're seeing essentially median pre-money evaluations get back towards where we were in 2021.

56:17But we've also discussed quite a lot about how companies have more revenue. They're growing more quickly. So does it kind of net out okay that we're seeing higher prices today because the companies are of higher value and therefore just going to be more durable to grow into those prices? I think it is. I think, you know, it's become a little bit easier, quite frankly, in terms of discussing valuation, because now we can actually base the valuation on a revenue multiple, right? And as we mentioned earlier, we're more than happy to wait to actually see some early traction and give a fair valuation based upon kind of what the, you know, generated revenue or AR at that particular point in time is, or at least what the expected one would be 12 months from now.

56:59And have a reasonable conversation about what the appropriate multiple is, instead of just trying to draw numbers out of thin air. So while I believe median valuations are kind of normalized somewhat, the underlying revenue growth is actually much better these days. And so thus we're getting more better, I would say, better priced opportunities. Maybe I'll add some color to that. Oh, dude, please. Like, I think the median valuations, I agree with Jay's point. But, you know, what's interesting is if you look at the 95th percentile, that's a whole different story, right? Like the median valuations look like it's kind of reasonable.

57:34oh yeah look a little uptick this year that's nice but things are growing i agree on the extremes i mean it's it's crazy out there right like you have multiple companies with no revenue raising at a billion dollars um you have you know i've had in this past week multiple people say we're ai guys so i need 100 million dollars because here's this cool result no business no nothing and i'm like this is a recurring phenomenon. So, you know, and some of them are going to raise those rounds, whether I say yes or no. So I do think there is also a bit of a story there as well. At the median, I think, yeah, it's fine.

58:12I mean, look, if we're growing so fast, if you're, you look at Anthropic, right? Like, Anthropic is literally the fastest growing software company of all time. Zero, 100 million, a billion, and then this year tracking to 10 billion. These are made up numbers. Like you can't tell people. You got to show this chart. I mean, this is mind blowing. I've seen it. It's crazy. Here is the chart from our dear friends over at The Information, giving them some credit of revenue growth from both OpenAI and Anthropic. This comes from a story, guys, that came out this week indicating that Anthropic has raised its revenue projections.

58:48And as you can see here, it's grown from incredibly little to quite a lot. One billion ARR at the start of the year to seven billion at the last thing that I heard, though, Didi, sounds like 10 billion might be the new number? I mean, around that amount. I mean, I'm following the information leaks. These are not investor comments. This is how much information I get about Anthropic. But look, I mean, the reality is Anthropic has always underestimated their own numbers. So this projection is not as wild as it might seem. if you know the history of how often they have undershot their actual goal, which is every single time.

59:27Every single year, they've undershot their own goals. So I think it's just insane how much growth there actually is for the absolute top companies. One last thing I want to say. Jason said the hardest job is in investors. That's absolutely not true. The hardest job is definitely of founders. I think maybe he was saying like chance of success is lower. Yeah, I'm talking more about like the compensation, but yeah. This is going to be put on VC brags, I can tell. No, no, no. There's two different things. If you really want to make the big money, be a founder for sure. And then joining these great companies is a clearer path to compensation.

1:00:07But VCs don't have to work 996, and they're not. I mean, yeah, that's very true, I think. What are VC hours? If engineers have to work 996 today, what is it like? I mean, I worked, I would say I probably worked 60, 70 hours a week. And people on my team work 50 to 70. Didi, how much do you work? I don't not work. Like, I think I work all the time. I don't think the work is as intense as it was that maybe when I was at a startup. But I mean, you know, look, I'll be just candid. Like, I'm new in this industry. And I think the only thing that I have to add is hours. That's like the only thing I can control is how much information I'm ingesting I'm able to process and then helping the people that I'm trying to help.

1:00:50And so, you know, that means that, frankly, like I will return texts at 1 a.m. I am up working all the time. I don't have a family yet. And so at some point, maybe that will change. But right now, like this is my life. All right. Jay, just because we're doing this apparently, how much do you work? And are you working the 10 to 4 that I hear VCs love to rock? Well, I think as Jason alluded to, I think all of us have to be constantly on. So if you actually count the hours that you're off, you're sleeping, bathroom breaks and whatnot. Maybe you have some private hobbies that you enjoy. I love to watch movies.

1:01:33I'll play some games once in a while. But excluding all of that, like I'm constantly on and, you know, everybody knows this, but I'm have this kind of bad habit. Like I have a zero inbox habit, so I will have to clear my inbox before going to bed. And the first thing I do when I wake up, grab my phone and go through my emails and texts and messages and whatnot. And so I think most functioning VCs are basically trained to work like that because that's the expectation. You know, you kind of have to keep pace with your founders. You have to keep pace with the founders. And then you have to, the deals are getting done so quickly now that, you know, if you need time to do diligence, well, that's on you.

1:02:15You better move quick because a founder can just say like, yeah, you know, we don't have time to wait. We'll talk to you in the next round. And if you are able to get in the next round. So you really do need to. The diligence I can share. Yeah, go ahead. You know, there was an interesting anecdote. I'm obviously not going to name the people involved, but, you know, there was a close friend founder of mine who was raising this round. And, you know, at some point he was like, hey, can you look through my data room? Like, tell me if I'm doing things right. This is kind of my first rodeo doing this kind of round.

1:02:47And I look through it and I give him some feedback at the end of the day. And he goes like, hey, man, we already have term sheets. I'm like, dude, you didn't have a data room. Like, what do you mean? Like you didn't have your financials on there. What do you mean you have a term sheet? He's like, dude, I'm just telling you what happened, man. Like, I don't know how, I don't know what I did, but it is actually true that because of the speed of these deals and speed is a differentiating factor in order for many people to win deals that look, there are some cases where diligence is just not happening.

1:03:20That's the reality. We just we just went through a lesson in 2021 about not doing diligence, investing too quickly and assigning unicorn valuations to revenue-free companies. What the f - are we doing now? FTX. I mean, you're speaking of Sequoia and other firms. A lot of them just drafted on each other's supposed diligence. And it is when the market gets hot, and I've lived through now a couple of bubbles, like three, I think, four. Yeah. And I always see the same thing. People lack discipline, so they don't do diligence, right? so they suspend disbelief. Okay. Number two, they get really greedy.

1:03:59So they don't take the opportunity to liquidate positions when they maybe have the opportunity on, they hold on too long, et cetera. And then I see people actually acting, I see people acting unethically. And that happens in a bubble as well. And you'll see this ZD, like you'll start to see people who retrade deals on either side. It could be founders, it could be entrepreneurs. You'll have a founder sign a term sheet for a convertible note at one valuation, and then another VC pays at a different valuation. They're in the same round. One person has MFN, another person has a side letter. Nobody knows the complete picture of what's going on.

1:04:38And so we've had to unravel some of these things that have occurred. I don't recommend anybody do anything other than standard documents and being upfront with everybody, because when people find out that you gave a special deal to one firm or that somebody had some special side letter or whatever it is, and they might have MFN in their documents, you may have to go correct that in the next round and the valuation's gone up now. Yeah, just always standard documents and don't play any games, you know, and, you know, just best advice. People get very weird. And then the other thing that happens is people, you know, sometimes they forget who helped get them here.

1:05:14You know, we'll have, you know, you see this every time there's a bubble, people will be like, oh, you got an amazing deal. You have to give up your board seat. You have to give up your pro rata. You have to give up your information rights. We invested in you three years ago and nobody else believed in you. Why are you asking us to give up all of these rights? And the founder will say, oh, the VC who was doing this round is saying, if you don't give them up, they're not going to do the round. And they're not, you know, all kinds of shenanigans occur. Just being standard and being upfront and Like you can really see people maybe start playing short-term games instead of long-term games.

1:05:48I'm a long-term game guy. Like really like the long-term game of everybody just being respectful and not trying to get over on people. I saw there was somebody talking, I think it was Jason Lemkin said he saw like a couple of founders just abscond with the money. So when we get these peak moments and then people don't do diligence, you could have interlopers come into the industry. And everybody knows those stories as well. So, D.D., are you seeing, I'm sorry, Jay, are you seeing similar levels of malfeasance and poor behavior as we are in, let's say, a hot market? I mean, fortunately, I haven't directly experienced as much, but I'm sure it's out there.

1:06:27We've heard, you know, stories from our colleagues out there. I do think that the level of FOMO is getting to kind of bubble levels for sure. I've heard stories now of high growth founders being chased at airport security lines, people showing up unannounced at doorsteps with term sheets. I mean, things are getting pretty frothy again. And so, you know, again, understandably, nobody wants to miss out on the next, you know, open AI, anthropic, lovable, whatever that is. But it's getting out of control. All right. So one way that founders create FOMO, and I actually have a tweet here from Didi, is by doing intelligent work on social media.

1:07:11This is a tweet from Didi saying, this is the ultimate TikTok growth marketing guide every Gen Z founder is using. And one thing I'm really bored of is expensively slickly produced launch videos, Didi. But it seems that because everyone knows you can now go from zero to 100 million ARR faster, they're more desperate to grow revenue from kind of day zero versus maybe a little bit later when they have better fundamentals economics and you're kind of tongue-in-cheek playing into this here but is it healthy for the startup land to index so heavily on i would say like in group virality over on x or tick tock um i i actually like don't mind it i don't think it matters i think the problem really is when the product does not follow that marketing right like the the fact that you're doing like viral distribution campaigns is is immaterial to me look x is a pretty terrible place to do viral distribution campaigns in terms of the volume and audience you reach so i think founders doing it on x are maybe more clout chasing than business chasing look the real audience is probably on the other like instagram and tiktok and reddit and things like that and um and i think that's fair look like you're trying to grow your product really fast like absolutely you should be employing all the tricks of the trade to do that.

1:08:24If you do that and your product is absolutely shit, then you're going to have a hard time, right? Like it's just going to, you're going to try to run behind something that you can't keep up with. So that's not great. And on the flip side of this, what's super fascinating to me is what's fascinating to me in the last couple of weeks actually is I've been doing this all the time. I suggest you guys do it if you have like a lot of free time. But do you know how, we know that LLMs and ChatGPT, often when you ask it a question, it will summarize from a bunch of web results, right? And a lot of those web results happen to be Reddit results.

1:08:56And so there are entire companies built on spamming Reddit with high quality content that's completely AI generated, where they sneak your company that they're trying to sell in there in ways that, trust me, like most humans, including myself, would not know that this is fake at face value. And so what's happening is like, it's not just the TikTok and Instagram stuff. It is also like this insidious nature in LinkedIn and in Reddit where there's these not only sponsored content, but just automatically algorithmically generated content from like basically bought farms of anonymous accounts that are influencing people's behavior.

1:09:36And then slop is an interesting terminology. I actually had this interesting conversation the other day where like the real question is, it is only slop when we know it's AI generated. But the reality is, in most cases, people don't know. It could just be a stupid person or just an unsophisticated person reading it and not realizing it's slop. Well, it's not only that, though. The quality of some of these short-form video creation is mind-blowing. I challenge you to notice a difference. Just a year ago, two years ago, you could tell, right? Everything is still in, just the mouth is moving. Nowadays, it's so sophisticated, multiple camera angles.

1:10:12I mean, look, the point that was made earlier about viral growth generation, some of the smartest guys are using these AI tools to do it. Think about the fact that instead of having to record these short form videos multiple times a day, you can actually create a video likeness, virtual likeness of yourself using AI, having scripts that are auto-generated using generative AI, and just having them automatically posted to a specific targeted audience that again you use agentic ai to specifically target we just closed an investment in a company and i've never done this before they are a ai chatbot company and again why would you invest in an ai chatbot company because they are changing the game and the way they fundraise they actually found investors all using their own tools the due diligence the pitch deck and even all the due diligence questions you would come to their website and you would be able to actually utilize their own chatbot to generate the engines.

1:11:10I love that. And so, you know, there are so many AI tools these days that are making every step of the startup business process so much more efficient. You know, if you're not taking advantage of that, I don't think you're doing your job right. What's the company that you just made an investment into? This is a company called Lyzer. Shout out to Lyzer. L-Y-Z? L-Y-Z-R. If you want to look it up, it's lyzr.ai. If you're listening into this live. I have not heard of this company. You'll be hearing about it. Okay. But I just love the ethos that they were basically eating their own dog food in their fundraising process.

1:11:51And they have a little bit of a shout out on a LinkedIn post to announce to other founders to utilize their own service to be able to actually raise your series seed series A. I want to loop back to what Didi said about insidiously, basically SpawnCon being put into human forums. This is kind of the dead internet theory, Didi. Are we at risk of reaching a tipping point in which we've automated so much content generation that we wind up in a world when it's just bots watching bots and humans are just left by the way I said? And I ask because I was expecting a more negative reaction to the Cluelly style hype videos from you all.

1:12:28But instead, very much a different response, which is good to ask questions. But it makes me worry that if you guys are this positive on it, we're going to see so much of it that it could become a preponderance. I think, look, it's just not my position to say, right? Like if I was at Google or at Reddit, I would want to tamp down on this. Absolutely. Right. I don't want my platform to be spammed with AI generated content. But on the founder side, like, look, if that is an avenue that is generating engagement for you, I can't complain if they want to use that to grow their business. So I really do think that it's like I hate Internet theory.

1:13:11I hate the fact that this exists at a philosophical level. I don't think that should happen. You know, there's this there's this great graph called what people are calling the new flippening. I don't know exactly how they got this data, but it shows how much of the Internet's content is human-generated versus AI-generated over time. And what does that mean for the future of how even these great products like ChatGPT work? It's just going to be a loop of eating its own slop and churning out stuff that is not human-generated at all. It means an opportunity, the opportunity for us as humans to in some way, you know, authenticate that the content we're making is in fact us and not AI.

1:13:57And I think what you're going to see is like in Star Wars when they say, hey, we don't want your kind in here on Mos Eisley and they don't let the droids into the bar. That's happening now. No AI. So people are now putting on their movies or, you know, on their stories, no AI. This includes no AI. There's no AI added to this. So it's going to be kind of like acoustic or organic food. People are going to start labeling things human only or 100 % human as a reaction to this. Now, I love the storytelling. So clearly, I'm not sure if their business will be successful or not. We've had the founder on here.

1:14:32He's super spicy. But they know how to storytell. And storytelling is important because the ability to tell a story is something that Elon does exceptionally well when he runs his companies. Steve Jobs does exceptionally well when he does his stories. Travis and Uber did this exceptionally well. They told the story of why this needs to exist. And to do it in a short-form video and make it punchy because today's tools allow you to do that, if that explains your value proposition to users faster than tool tips in the app or word of mouth, great. We have a company, Howie, we're investors in. We seeded the company and they just did like their own launch video about like doing meetings.

1:15:14And a lot of VCs uses Howie software to do meetings and book them. And they've got human in the loop to make sure it does what it does correctly. That ad just added like a lot of brand value to them. It added like a vibe of like, this isn't just a regular service. This is like, we actually care enough about the service to tell the story in a really interesting way. I kind of like it, actually. It's kind of the difference between somebody who makes like a cheap$3 t-shirt merch that nobody's ever going to wear. And then somebody makes some really interesting, great piece of merch that you actually use in the world.

1:15:55It's kind of like that. But I love some of these launch videos. I could watch them all day if they're done well. All right. Well, come on me wrong on that one. Guys, I have to take us to a conclusion here. We could talk for the next three days. I know, but I do want to give each of you a chance to shout out a portfolio company. Tradition that I do these. So Jay, who is your current favorite child and why? Oh, no, no. All of our babies are cute and beautiful. But I do want to give a shout out, since we've talked about Higgs build already, to Abacus. And Jason, I hope they're on your favorite list as well.

1:16:27But I just love the fact that this entrepreneur bootstrapped his way to a million dollar AR. are. And as we all know, while every consumer out there has now tried some form of AI agent, whether it's ChatGPT or whatever, there's certain enterprises and certain industries where you're literally banned from using it, right? And so I think the fact that in the financial industry or the healthcare industry where your data cannot leave the print, you have companies like Abacus that come from the industry, know these pain points, and they've built a model where they basically have created kind of a, I would say, a AI foundry service in a way where they enable the enterprise customers to basically train on your own data.

1:17:14They provide all the tools and the infrastructure using open source models, and they are able to even provide a application generation platform where you can build your own in-house version of ChatTPT to be used for in-house compliance rule lookup policies and all of these things that you know would be great to have on your fingertips they've screened out for hallucinations and whatnot to make sure it's 100 accurate and once it's been proven and tested in an internal environment guess what they're going to try it with partners and customers yeah and so it's just a great great model and just love the greed of the entrepreneur to actually start this from scratch bootstrap it all the way to a million dollar ARR and now they're off to the races.

1:17:59GoAbacus.co is the URL if you want to check it out. Didi, who is your favorite port co and why is it Open Router? Well, I don't think Open Router needs another shout from me, given that Elon tweets about it like five times a day.

1:18:16I mean, it's got the best charts online, man. I love it. I'll talk a little bit about Open Router. I'll maybe do an actual shout that I think deserves it, needs it more maybe. Open router, look, fantastic business. It's part of my thesis that if you are, if you can be a product-led growth company, you must be a product-led growth company. There's various versions of an open router that could have gone and sold B2B to enterprises, but they are absolutely going to be steamrolled by this massive thing that has awareness in the broader world. And they are. So love that company. It's fantastic. We're going to live in a multi-model world people want to know what's happening there's 20 plus great models and hundreds of other models sure with different providers and people want to know which providers are doing what and how can i access the different models so the company you love even more is i love open rudder they don't need my love anymore they're they're they're they're they're kings um good fire is the company i would love to ah yes and um and the reason um a little bit of the backstory on Goodfire is, look, it's a mechanistic interpretability company.

1:19:20That's a fancy way of saying brain surgery for AI models. We don't really understand how AI models work. We don't understand how they think. We rely on them to give us thinking traces, which are poor proxies for what's actually going on inside their head. And these are the top mechanistic interpretability researchers who started teams at Anthropic, DeepMind, OpenAI, who've come together to set out on this mission to really understand what's going on in AI models as it becomes ubiquitous. I think it's absolutely fundamental for society for us to understand this because we cannot trust black boxes with huge decisions.

1:19:56And I think they have tremendously cruel scientific breakthroughs that they will announce at some point that are going to be extremely valuable. I was excited about Goodfire if we got into deep tech because I know that Jay's made some investments into robotics companies. And when I saw that they were an AI interpretability research company, I was like, yes, that. But we got to move on. Jason, last question is for you. Who is your favorite portico or at least the one you want to shout out on the show? Well, I mean, I could have done Abacus, but I'll do a different one and spread the love. Next visit AI.

1:20:27This is an AI powered, you know, clinical note taker, except it works in a very specific vertical, which is psychiatry. And that has very specific note taking details, as you could imagine. And this is a great co-founding team with a psychiatrist and an AI specialist. And they very quickly got this to work. And you can imagine psychiatrists who are doing counseling or just sessions and prescribing drugs and a lot of other modalities. They spend like half their time, a third of their time just doing paperwork. And this stuff has just nailed it. The fidelity, when you go after something narrow as a startup and you get it right, it can really pay off.

1:21:13And so this is one of the things we try to teach people in the accelerator. They're concerned, oh, VCs are not going to be impressed because it's such a small idea. Like trading a stock for free is a very small idea called Robinhood. Like watching your taxi come to you on a Google map and paying for it is a very small idea, but it's a profound idea that then things were built on top of like Robinhood is not just trading a stock for free. Now you can get margin loans, you can do crypto, you can do prediction. Mortgage, Jason, you mortgages now. Okay. I wasn't even aware of that one, but you know, it's, um, I always like this laser focus.

1:21:52I tell people like you can turn on a flashlight and just like kind of point in the forest and like, yeah, we're going to do something out there. When you take the laser pointer and you're like, we're going to do that. See that thing right there? Right there. That's what we're doing. That laser. We're going to do it really, really well and make it perfect. So the customer has no choice but to use this product and they love it more than anything they've ever had. That's kind of the magic in that year zero or year one. Just do one thing better than everybody in the world that makes that customer just so delighted.

1:22:24They won't shut up about it. And that's when your PS score gets to like nine or 10. And so everybody wants to do, you know, the co-pilot that transcribes what you're saying. But man, just getting in there and knowing that this person is taking this ADH medication on this level and it's this code and it's this type of ADH medicine. It's not the generic. It's a specific one with a specific one. Like you got to get that right. You can't get that wrong, right? Or else it's just not worth using. It's not like a hallucination when you're trying to figure out what movie to watch this weekend. Favorite Paul Thomas Anderson.

1:22:57Jay, your favorite Paul Thomas Anderson? Speaking of movies, you say you like movies? You have a Paul Thomas Anderson guy? Yeah, the most recent one. What is it? One battle after another. It's crazy. I was on the edge of my seat the whole freaking time. That was crazy. It is like a three-hour... Oh, yeah. It's an explosion. It's a marathon and a sprint. It's crazy. but it's you know my Paul Thomas Anderson favorite still remains The Master oh that movie was The Masters, that's next level Didi's looking at us like how do you guys have time for this we're from a different generation, we cared about cinema alright everybody see you next time alright bye everybody, this has been Twist, see you all soon this was so much fun guys, thanks thanks for doing it Jay, thanks Didi