415. 2024 Predictions P1: A Perfect Storm for Crisis, Feast & Famine in VC, Big Tech AI vs. Startup AI, and How LLMs Become the Ultimate Accelerant & Tax on Venture (Ramy Adeeb)

415. 2024 Predictions P1: A Perfect Storm for Crisis, Feast & Famine in VC, Big Tech AI vs. Startup AI, and How LLMs Become the Ultimate Accelerant & Tax on Venture (Ramy Adeeb)


Ramy Adeeb of 1984 Ventures joins Nick to discuss 2024 Predictions P1: A Perfect Storm for Crisis, Feast & Famine in VC, Big Tech AI vs. Startup AI, and How LLMs Become the Ultimate Accelerant & Tax on Venture. In this episode we cover:

  • Venture Capital Industry Challenges and Potential “Great Resignation” of VCs
  • The Current State of VC Investments, Startup Challenges and Revenue Metrics
  • Investing in Startups, Crypto, and AI
  • AI Adoption and Its Challenges for Startups and Incumbents
  • AI’s Potential Impact on Various Industries and Applications
  • Interest Rates, and Startup Valuations
  • AI and Tech Industry Trends

Guest Links:

The hosts of The Full Ratchet are Nick Moran and Nate Pierotti of New Stack Ventures, a venture capital firm committed to investing in founders outside of the Bay Area.

Want to keep up to date with The Full Ratchet? Follow us on social.

You can learn more about New Stack Ventures by visiting our LinkedIn and Twitter.

Are you a founder looking for your next investor? Visit our free tool VC-Rank and weโ€™ll send a list of potential investors right to your inbox!

Transcribed with AI:

0:18
Welcome back to TFR. Our good friend Ramy Adeeb is back to recap 2023 and talk predictions on 2024. You know, Ramy, I was listening to last year’s episode. And unfortunately, you had not predict the Ukraine war from the prior year. And then this year, I noticed you failed to predict the banking crisis. So we may need to revisit you as our predictions expert going forward considering you missed these enormous events?
0:48
Absolutely. I mean, my whole spiel is to predict the global macro economic financial trends, but the fact that my own financial institution is about to go run out of money, because it learned more than it had available on its banks was totally lost by myself and my entire community. But really, the the biggest crisis was what a shitshow.
1:08
Should we have predicted that? I mean,
1:11
look, there are many. I think it was a glimpse for what a banking failure will look like 30 years from now, because what happened is particularly fascinating, you don’t want to Washington Mutual went out of business in 2008 10 billion of deposits were withdrawn in 10 days now Washington Mutual was like the biggest banking failure in like the past 15 years. With Silicon Valley Bank 40 to a billion of withdrawals were taken out in four hours. It’s insane. How because we all had mobile banking. That’s right, two factor authentication, and Twitter. And we all felt the same people. And they told us our wire so we sold we think it’s a fascinating, you know, Hemanth, a lot of VC said, Well, that was the so we did not want it doesn’t matter what we did or what other people did, but a lot of a lot of folks in our community were like, look, we have to wire the money because the future you’re a responsible thing to do. And I thought her mind, Taneja JC had a great quote, he said, causing a bank run at the most important institution in your sector can hardly be considered an act of fiduciary duty. And I really appreciated what he said there. But it was inevitable. Here’s a great driver. So you know, fortunately, that was gone. First Republic followed. But it does raise a question. In a world where every depositor has the capacity to wire all their deposits instantly with a cell phone, which by the way, is not the case today. Like if I told my dad, Hey, Dad, you know, how would you wire all your money out of Toronto Dominion Bank into like Royal Bank of Canada is he’s not sure. But we are in the business of sending wires
2:52
to his VCs, what we’re good at. So we’re good. So we’re like, Oh, your should I wire the wire? That’s value add right there and leave the money at risk? And just wait it out? Or should I wire it out?
3:05
And I think it raises a question like,
3:07
do we have
3:08
actually the right controls against the bank? Or, you know, historically, there was friction and bank runs? Because you don’t see the line? And then you saw the line? And then like the bank closed at five sorry, guys, like whoever didn’t come to the to the teller by five o’clock. I bet you can
3:23
guess what movie was on last night. It’s a Wonderful Life. Amazing. The classic George Bailey bank
3:29
run. Absolutely. And you know what I love in that movie, he convinces people to not take their money out. Right leaves a role he gets the opportunity
3:37
to see your money is with giant exactly in his house. And
3:42
he convinces people to take $20 or something, right? He can just like, Hey, who wants like 20 cents on the dollar right now? Let’s do this. Yeah, let’s live for another day. And he survives. It’s such a beautiful
3:51
movie of the CEO of silicon back
3:53
trying to do something very similar on a zoom and to disastrous consequences. But I do think the biggest takeaway, which nobody has acted on yet is we need circuit breakers, similar to what you have in the public market. stock dropped 7% while trading, right, that’s otherwise things crash in a world where the posits are fully liquid fully electronic, right? 20% of people just wire them out of the bank, circuit breaker, get people to take a breather, but yes, it was
4:21
addiction. We need some stop loss prevention in our banking environment. That’s right. That’s right. And it’s interesting around me because I think most of the general public, most of my friends, they’re like, oh, yeah, those venture banks, like they were loaning all this money to startups and they created their own demise. And that wasn’t the problem taking
4:39
this topic. They weren’t able to lend actually. So they ended up taking all the deposits and buying Treasury bills, which is the world’s safest, safest asset until interest rates start going up in price and if you have to mark to market, then all of a sudden you fail to meet your deposit ratios. You know, startups are actually very hard group to bank, precisely because they’re a very hard group to lend into it as a consumer grade, you deposit $5,000 a month in salary, and then you take out 300k in mortgage, the perfect customer for a bank. But startups wanted to deposit 50 million 100 million. They don’t borrow anything. So you have to do something with the cash. And they did what they thought was the safest thing. But in the process, they were taking an insane interest rate risk, just like many in our industry, we’re taking an interest rate risk, the only catastrophic
5:37
service provider crisis that we suffer here. I don’t know if we can handle another one. Let’s hope so as well, Nick. Let’s hope so. Alright, so the banking crisis is behind us. It looks like Silicon Valley Bank has survived as a part of citizens. And FRB of course is with JP Morgan. Absolutely. We’re
5:56
now back to jpm. The number one global whatever, systemically GSI GIS globally,
6:04
I did get two notes from LPS during the last capital call. Are you sure you want us to continue to wire the First Republic? And I said, after the JPMorgan Chase acquisition? I think it’s okay. All right, very good. Let’s move on. The seed market is on fire the series a market is on fire. There’s no time like now. Well, there
6:25
really is no time right now. It is 2024 We’ll gear up to be the best vintage V banner vintage for our industry. I couldn’t agree more.
6:35
I was in Boulder. Recently, I was talking to Linda Lachman. And the discussion was about how this could be the best vintage of our career, it is
6:45
the next 12 months are going to be the 2009 and 2010. Uber Airbnb squares, you want to be investing right now, you want to be investing in early stage right now, sadly, most of our listeners, though, aren’t necessarily starting companies, many of them have companies that are running, and those are going to, you know, there are gonna be a lot of challenges ahead. And for the VCs, there’s also going to be a lot of challenges in the year ahead, going to be probably the great VC resignation, in peril of being the best investing vintage year.
7:19
So tell us more. What do you mean by that the great VC resignation?
7:23
Well, you know our industry, we take money from LPs, we invested in startups, we get the exits, we give it back to the LPs and then the LPs, give it back to us. And the LPS haven’t been getting anything back. And so they don’t have anything to give back to us. And so the money coming into venture capitals is drying out. And it’s drying out, not necessarily, you know, part of it for lack of excited sentiments about the about the industry, but part of it for just lack of liquidity for many of the folks that are overexposed to venture. And so there is going to be despite the quality. If you’re not a top quartile, or really differentiated fund, you’re going to be forced to dramatically scale back are not able to raise and most funds did raise in 2021, or 2022. And we follow two three year cycles. So you know, unless you’ve raised in 2023, you’re probably due to raise this coming year due to raise this coming year. And that’s going to be tough.
8:15
I think the great VC resignation implies, these VCs are opting into that in some way. Because we’ve seen that in the past, for the old VCs that had so much success, they decided to kind of hand the torch on to the next generation. But I think what you’re talking about here is there’s a lot of VCs that would love to be in business that will have no option. It’s like almost like more of a riff. As as sad as it may sound like many of my conversations with VCs these days, they’re, they’re talking about being in the bottom of the J curve. And you know, the translation for that is like IRR is or negative. You know, they over invested probably in 2122 at inflated valuations. And now TVP eyes and IRR is and portfolio is looking shaky.
9:00
Okay, I do think there’s a lot of tourists coming into the industry. UVC is a very desirable career. And one of the best quotes I’ve heard is it’s really easy. You know, it’s the easiest job on the planet. If you’re a kid being an average VC, and it’s the hardest job in the planet, if you’re trying to be a top quartile, VC, the amount of extra work you have to do, to find companies to diligence them to help them is immense. And this will be a year where those who worked that hard will probably remain in business and many others will, will be forced to leave. And part of it also will be luck. I mean, you’re right. It is a bit of a riff. But
9:35
it’s a bit of a tale of two cities, right? Because there’s I don’t know that there’s ever been a better time to be a VC, like getting money in the ground. Now. The pricing is I mean, it’s come way down. There was a lag effect. Right when we did this last year. See pricing hadn’t fully corrected. It’s taken a while to work its way through the system. But I’m seeing pricing like we’ve never seen extensions being done, you know, companies that are trending towards multiple millions of ARR that are extending their seed rounds
10:04
are pristine. Yeah. And we’re getting the so there’s also fight. So there is a quality on the valuation front, there is also quality on the team front, right? In 2021 2222, everybody would start a company today, to really have a burning need a desire to solve, like a hair on fire problem for someone else, or be an AI engineer to want to start a company. Right. And so you know, we’ll get back to the AI engineer in a moment. The quality of the founders is is insane. It’s absolutely insane.
10:33
So does this mean a higher percentage of the founders that you meet with results in an offer?
10:38
Yeah, they’re serious. You know, you’re no longer starting a company for the sake of starting company because they know how hard it’s going to be to fundraise. But I also think, you know, to your point, are we investing in a higher percentage, somewhat a little bit, but we’re a lot more impressed with the caliber of the founders. We’re meeting every day.
10:56
So tough time to be an average VC great time to be a good VC, a tough time to be a startup founder, regardless of what decile or quartile you’re in, I think.
11:08
Absolutely, absolutely. Very tough time. There are some silver lining, though it’s easier to recruit, it’s much easier to build a team right now. Yeah, there’s also clear market signals, you know, 2021, when corporate spending is up to us, you just kind of at the end of the day, as a founder, you’re trying to find signals of product market fit. You’re like digging a weather deserts and hoping to see a little bit of water so you know, where to dig deeper. And in a frock that time, there was a lot of fake signs of product market fit. You know, if you wanted to do a startup that sells kombucha to HR teams and tech companies, and you did that in 2021, you probably get to like 10 million in revenue. In a try that today, you’re not going to get a single sale. But when you do get a sale today, from an enterprise, you are solving a real problem for them. You’re either improving top line or reducing bottom line. And when you do that, and in the current funding environment, you’re building the business in a way more cost conscious way. And a way where the unit economics makes sense, in a way where you’re not selling dollar bills for 75 cents, and that’s just built, when you build the business from the ground up was the right unit economics. You know, the path is the passes. Incredible. You know, I would start convoy today. But I started today. unfortunate demise this year of cowboy. And that was Yeah, and that was a tough one. Right? That Was that too much capital raise
12:29
what was absolutely huge.
12:31
I mean, we’ve had a good article death from overfunding and obituary for convoy. Perfect. And think about it. Yeah, literally. The Congo is a marketplace. When you have a marketplace, that GMV is a vanity metric. GMV is a vanity metric, the only metric that counts is your take rate. That is a revenue, right? Because some marketplaces take 1%. So marketplaces, take 10%, some marketplace take 40% confluent was trading from day one based on GNP and GDP growth. And then the tea crate was, like, publicized to be 8%. But sometimes it was negative, fully negative. As I’m like, There’s no revenue, the company actually has no revenue effectively, right, but raising billions of dollars valuation. And I think that topic of vanity metrics is an important one for 2024. There are many vanity metrics that we I think, as an industry, subscribe to in 2021. And we’re slowly reading away from use of GMV instead of net revenue in businesses as one. The second one is waitlists. Remember, when you see company, they were like 1400 people in your weakness? I’m like, what does that mean? Just like you have 1400 friends on Facebook all agreed to click on that link to build product and you can’t monetize. That means nothing. Whereas the product, we don’t have the users waitlist is not a metric that’s gone. We don’t see weightless startups anymore. And then obviously, when cash is dear buddy of just dumping, dumping cash to acquire customers ratio enterprises are doing that. And so seeing top line growth with ridiculous burn rate is not something we’re seeing as much as we did back in 2021, or 2022. Yeah,
14:02
I had this good interview recently with Mitchell green at lead edge capital. And he was talking about something he calls the hierarchy of bullshit. And it’s like, if you don’t have high quality revenue, you talk about low quality revenue. If you don’t have low quality revenue, you talk about GMV. If you don’t have GMV, you talk about number of users. If you don’t have users, you talk about number of followers. It’s like it keeps going further and further back to like, waitlist. In your example. You talk about the vanity metrics, that that seem positive, instead of the true metrics, what’s the revenue? What are the margins? What’s the growth rate? What’s the retention?
14:40
I do want to caveat, you know, the question you asked about, you know, it’s gonna be a good year for good a great year for a great VC but not familiar with the challenges. It’s, it takes decades or at least a decade. It takes years to figure out the VCs are performing or not. And the great companies and generate all the midterms typically start appearing on the portfolio by year six or seven. But think of figma. And so I think we will lose some amazing folks from the industry who just haven’t had enough time to prove what they would be capable on,
15:14
there’s more to running a venture firm than picking. There’s a lot that goes into it. You got to do a lot of work communicating with LPS working with LPS develop,
15:24
you communicate pretty ropey.
15:28
Okay, so we’ve talked a bit about the seed market, what are you seeing, beyond that? Talk a bit about like, Series A, Series B, what do you see in these days,
15:37
and let’s start with the very late stage, right, the company’s every B’s and C’s, we apply a lot of band aids as an industry in 2023, trying to keep these companies alive, that don’t need to be alive, there shouldn’t be alive. And I think those monies will run out, right. And so you’ve seen a lot of great rounds, continuation rounds, sometimes down rounds structure, that what we saw the most were flat to the last round, or slight uptick for the last round. But then if the company is not doing well, there’s also some options, basically, make it a down round. But on paper, it’s like flat to the last round, can lead to too much structure. At the end of the day, you know, if your valuation is ahead of your product market fit is and you don’t see a path to get there. And another year, before you run out of cash, you are in a tough, tough, tough, tough spot. I think you know, another way to think of it is we suffered was trading that color 20x Multiple now turning net 6x Multiple, just taking gross numbers for a moment. So you need to grow 3x To get to your last valuation. And if you don’t have that, if you’re growing 40% year over year, can take a long time to get to 3x You’re not going to get there.
16:46
But the reality is, if you’re a VC, and you’ve been in the business long enough, you all have some of those, because any startup that raised during the hype phase, probably raised at a multiple that, you know was a little disconnected from reality. Yeah, it depends.
17:03
I mean, it also depends on the startup and the phase. But yeah, we all have some of them in our portfolio, and you stay hopeful. And obviously the late stage funds are the ones who are dumping more and more cash, some of which condors have managed to exit just returning, you know, maybe fractions of $1. This year, properly in our portfolio and brief boosts managed to orchestrate ASICs for themselves. They’re both in the Prop tech sector founders reasoned, pretty pretty astutely, and rightfully the year that there’s just no hope, between fun between the fundraising environment and the high real estate, high interest rate environment, which affects proptech affects real street transactions, which in turn affects profit companies who just you know, hope for their business model for the next few years. And so they decided to sell and orchestrated a quick sale, but also not everybody gets an opportunity to sell, right? And once you start selling, selling, you can’t really wait. I mean, it’s really hard to raise. But once you’re selling mindset, that’s the end you’re either selling or running out of cash.
18:03
Which of the later stage investors do you think navigated the past few years the best?
18:09
I honestly think it’s benchmark I hate to always refer to them as the market.
18:13
This is not they’re not gross age. Good
18:15
point. Yeah. But they resisted the urge to become a growth stage. And they looked very different to crypto but only invested in like geneticists and like one other, even though they were looking Dee Dee Dee Dee Dee Dee Buck stated their conviction instead of follow the hype. I really respect them for doing it.
18:30
So where do you land on crypto now? You’re very bearish the past two years, took my
18:35
401 K in order to invest in from the font all the above? I think anytime people are excited about crypto because they think they can make money. It’s a bubble. It’s a speculative bubble.
18:46
The coin is back Rami. I know, like 44k ish. But
18:51
I think it’s fine to play these bubbles, you can make a lot of money from these bubbles. It just doesn’t at all, coincide in any way with our thesis of like software, solving your hair on fire problem. I mean, actually, I do think that back getting rich is a heroin problem for most of most Americans to cryptid us solve our problem just in a somewhat and a somewhat convoluted way. On a more serious note, there is a lot more in from our part in founders who are actually using crypto to solve interesting problems. And we met with more crypto companies. We’ve made one investment in a company that uses the blockchain. They don’t use the word crypto in their name. They used to have the word chain in their in their company, but they rebranded so nobody knows that they’re using the blockchain. But the blockchain actually has a scholarship that allows them to do some of the stuff they do. Perfect. And so I think that’s more a sign of the time. These are the kinds of companies that want to invest in using technology to solve real problem, not because it’s the height sector.
19:55
Well, it takes her flying Rami, a diva is invested in a crypto really Did blockchain startup, but I’m glad it was a real problem. And that was the right technology solution, as we’ve discussed ad nauseam in previous episodes. So the hype of the day Rami, the hype of the day is AI, how are you approaching AI investing? What’s hype? And what’s reality?
20:19
So I do, I don’t put AI in crypto in the same
20:24
in the same bucket. From a hype standpoint, though, that’s from a hype standpoint, the two of them. Yeah.
20:32
And but the reason is, because the applications of AI, there was many applications of AI immediately, with immediate benefits to consumers and enterprises that we’re seeing, as opposed to sort of build to them, they will come and one day, we’re going to move the banks, you know, the banking system, from Fiat into non Fiat and whatnot. Right, like, if you go back to so the, with AI, the biggest challenge is, by the time a sector is hot, it is too late for seed funds. This is a very painful, very painful recognition. When you wanted to invest in AI, we should have invested right, you know, when it you know, did the check and opening Hi, that those are pretty controversial check, like most early bets, but right now, I mean, a lot of founders who are, you know, starting a foundational model, because they don’t know what else to do with their lives. And the number of startups doing the same thing has just ballooned? Yeah, you want to writer, here you go. You want an AI for customer support is like 100 companies trying to do that. AI for sale, automating sales, every company is doing that. And the truth is, most of the tech companies are able to incorporate the AI incredibly effectively into their workflow. Unlike mobile, mobile really created an innovators dilemma for nurses at Yahoo, for example, we knew that our users wanted mobile, but every time we send a user from the desktop to the mobile to check Yahoo Mail, we knew that we’re getting pennies on the dollar. It wasn’t mountain cube. And so we wanted to bring them back to the desktop. So the case was AI, enterprises love to integrate AI and love to provide the AI solutions to their users. You know, the author of The Innovators Dilemma, Clay Christensen has a great, another great thing he talked about is segregating new innovations into two types, what he calls disruptive innovations, which disrupt the incumbents and sustaining innovations, which help the incumbents and he has a fascinating litmus test for what constitutes a disruptive innovation. Guess what it is,
22:29
it has to cannibalize existing cash cows. So that
22:32
is the outcome, that outcome, but actually, it’s actually very interesting. It has to underperform the existing technology. And it’s early days, it shouldn’t be as good and feature exactly, it is not as good. So a great example is to vehicles. Mercedes Benz owned a whole bunch of Tesla and sold it in I believe, 2011, and 12. Why? Because the range was not there, and the battery sucked. And when they got an accident, they exploded. And nobody wanted to drive an electric vehicle. And you and I got to just focus on internal combustion engines for now. Just focus on it. Fast forward 10 years, very prudent management decision in 2011. That was this haunting them like crazy. We need to pivot the entire business to electrify you know, the electrification, or we die. But with AI, what CEO is going to say we don’t want to commend Dr. Someone who hasn’t read the Wall Street Journal in a year. It’s just so obvious. And it’s so much better. And so everybody’s incorporating it. So what is the UiPath? For AI? It’s UiPath. What is it notion of AI? It’s notion. And so, for a startup to compete with existing tech players based on AI, it’s very tough. What we are seeing though, I think what we’re excited about are areas where you’re using AI to compete in sectors where the incumbents might just have no tech, right? And what I tweet makes a difference, then it gets interesting, like a private equity company that was acquired company caused by private equity that kind of go, they slashed r&d Team increased ACV did all the good financial engineering things and operational improvement teams are private equity does. And now they’re like, you know, making 800 million in revenue a year $6 billion publicly traded company, or private equity health company. Yeah, let’s just go after these with AI. And we have a couple of examples. I can talk about more data. And other AI that we’re really excited about actually are areas of AI, outside of large language models. Large language models just sucked the air in the room. But there’s so much innovation in machine learning was like improvement in predictive analytics, which is amazing, right? Things have gotten so much better now because of neural networks than they were in the past where they were simply running a regression and time series data and we’re seeing a lot of applications that are business so we think there’s a lot for it than just the eyes writing in language and language models are sucking too much of the air out of the room. And it helps that you know, opening I think has done a great job hiding that almost creating what I would call a fanciful fear. or have, you know, artificial general intelligence? And really, you know, I think all it is, is computers have gotten really good at language when they were previously very good at numbers, but not at language.
25:11
So your Fear Index is not at a 10. coming
25:15
in. Okay, cool, right. I mean, there are bad actors today in countries and trying to use technology to penetrate our utility grids and sabotage our infrastructures. That intent already exists in North Korea, or in some hackers groups in Russia. And we’re fighting against that. Yeah, adding some AI on top might not be helpful. But the the threats humans plus computers today are already capable of a lot. So
25:42
I want to follow up on something you said before, so you think most of the AI benefit is going to accrue to the incumbents, and you think it’s gonna be hard for the startups? Can you? Yeah,
25:55
to compete a loan based on to compete a loan based on AI capability? Like, you know, we, I mean, let me take an example. My
26:03
first startup came in, yeah, like infrastructure, but not application. Oh,
26:08
even for an application. I mean, like, let me I don’t, I don’t like to pick a specific examples. Because, you know, like, let’s stick document writing, for example. I really I don’t write English is my is my second language. And my paragraph structures often needs a bit of refining at the end. And so I just love what notion has done. There’s just ask AI improve writing, and it fixes my paragraphs, that partners at the firm used to do that for me a lot. Well, it’s my turn to write our newsletter. And now I’m like, Oh, I’m good. Thanks, guys. I got I got AI. But then there’s a really good startup I met really awesome founder who was building from the ground up AI first document writers. It’s better than notions to little bit better. But all my documents are in notion. And so I’m just using notion for now, I’m not using notion more than I’m using Google Doc. Because for I don’t know, blog, Google somehow has not added some like basic AI capabilities a Google Doc, like I just have no idea what Google’s doing. And how slow and dysfunctional is that?
27:13
I am less worried about income as but notion
27:16
that it’s an excellent point. It’s an excellent point, Nick. And I think maybe in my world, as a seed investor, a serious B Company is an incumbent, right, and are a serious C Company or whatever notions that now that’s fair, is an incumbent and they’ve done a great job with it. And ditto for so many other you know, superhuman has done an incredible, you know, I used to prepare me for my email. Now, that email arrives with a summary on top. And I had heard pitches about like email was AI, superhuman isn’t, has added it. And Google has done a pretty good job adding the complete AI functionality, part of
27:53
what you’re talking about is a general purpose thing that everyone needs, which is like writing help. And even even with writing help, you’re still depending on notions, use of, you know, some LLM to help you, you’re still within that environment, you’re within that use case. And, you know, it’s almost like these MLMs. And the infrastructure layer is almost like fuel. And all the startups and all the applications, regardless of sector, they’re going to have to pay the freight, you know, just like they pay for GPUs, or CPUs, and compute, but it’s all going to end up being pretty application specific, in my mind. And
28:33
so you’re envisioning an application, some applications to emerge that are so AI dependent that the incumbents can’t, can’t provide, I just
28:42
think every startup is going to have to incorporate AI to some degree, whether it’s machine learning, predictive analytics, or LLM. And the nuance of the use cases and how they deploy it, how they train the models, like is going to be very sector and application specific. And so it’s easy to say, oh, all the big incumbents are going to do everything in AI, Google and everything. Apple can’t even figure out how to do auto techs. Correct? Correct. I mean, like, they’re going to do the big things. They’re going to build the infrastructure layer, and then they’re going to do big things like email correction and integrating MLMs into that. But I have very little fear that the big incumbent tech companies are going to take on every application where AI can be a huge accelerant, just like software, it’s like the next iteration of software. You know, the Internet came out, oh, it’s gonna take over everything. Well, it ends up spawning many, many types of companies, ecommerce, software, logistics, et cetera. And it’s still reliant on domain experts in those spaces, building companies that can leverage the technology the right way. Technology isn’t a solution in and of itself. It needs to be purpose built. and used specifically to address problems.
30:02
Very well said, very well. So cool. So
30:06
what else in AI? Should we be focused on and thinking about for 2024? How about open AI Sam Altman
30:13
feud? I
30:14
know I used to work with the node in company at COSLA, early investor and open AI?
30:19
No, I think I forget about the drama, the drama is a just, you know, sometimes or Silicon Valley episode of why like, traditional corporate governance, it’s probably a good idea. I think there’s a bigger question, let’s assume that AI is going to be so foundational to so many applications, will the value accrue to the application or to the generic model, because open AI presents itself to the world, like large language models are the future GPT for is, and then Deputy Fire, whatever is going to be the future. And over which we’re going to, we’re going to be so big at a company that we have this 100x, you know, return limit for investors for fear of capturing all the value in the universe. And I’m like, you know, sounds like something Netscape would have said, in 1994, right, and look how far they’ve gone. They’ve invented the internet browser, you know, yeah, the only get to the internet with the Netscape and how much value they capture. So I do think the definitely has the best technology in the market today. There’s many very interesting innovations that are behind probably not as good. And opening, I will definitely maintain the loop for a while, but not for every application. We have companies that are not reliant on GPT, for an all for anything, or that rely on other open source models or other platforms, and are doing well, like pure element companies doing NLP extraction and applications. You know, I started my career many, many, many years ago, I just got a call tell me and we use voice recognition to automate call centers. So back in 2000, if you call fidelity or Wells Fargo, or southern five, or 18, t or e trade, we enter all these phone numbers and we use, we did not have our own voice recognition technology, we had the application, the company that built the voice vision technology was new ones, and they had the best on the planet at the time, they captured like pennies, and we made millions. And, you know, we could have swapped with AT and T which wasn’t as good. But at the risk of prices, we could have done that. So I’m just again, I think, I think open AI has led the way with AI be interested in opening a billion dollar valuation, because I expect it to be the largest company in the universe. Something
32:31
we’ve thought a lot about lately in in startups that we’re investing in, especially those that are leveraging open AI and LLM is access to proprietary datasets, the more that dataset is proprietary, the more of the moat, we think that they can build and defend over time.
32:49
You know, in tech, we like to take a lot of concepts and reinvent them, right for tech. And I do think one of my favorite concepts they are you know, the Porter’s Five Forces, competition stars, you know, supplier power, customer power substitutes, and whatnot, switching costs. But one of the specific specific most interesting one there is he talks about the power of suppliers, right? Because if the supplier is opening a supplier, and there are many factors that go into suppliers capacity to control price, but the simplest and most obvious one is number of suppliers, the number of buyers, number of suppliers, the number of buyers, sure, because if only one supplier is gonna be tough. But there’s three suppliers, just three suppliers. In fact, all it takes is two suppliers to get the prices down. It’s a whole notion of antitrust. It’s why since I moved to New York City, I’ve done the Lyft app, because the Uber prices are so insane. Like what are you guys thinking? $50 for six blocks, no way. You know, in San Francisco, I’ve never felt that insane, insane surge pricing. Now just use left, when Uber’s really annoying me if Lyft did not exist, holy shit would have been insane. accomplishment for a recent investments in lifters not to create lift but to reduce overall market cap.
34:12
Will the LLM ‘s end up looking like the compute players? You know, you’ve got AWS, you’ve got Google and Microsoft. And then you got a bunch of like small Oracle and whatnot that are trying to make a run at it. Is that what we’re gonna see with LLM? I
34:26
think it’s I think it’s very possible. I think there are three layers, right. There’s the LLM. And then beneath that there is actually the clause where they live in and then he said there’s the hardware that they’re trained on and we can see some crazy vertical integration happening here. But yes, that I think that analogy is very analogies very, very sound. Gonna have to be a handful of players. Hard to imagine just one.
34:49
All right, let’s let’s let’s talk about zip a little bit. So about one zero interest rate policy from the past, so it’s probably the best and the worst thing that could have happened to our space, it artificially caused it to grow very fast. And now we’re paying the consequences. And the Fed just came out and said that they’re holding rates with plans to reduce next year. So talk to us a bit about predictions on how interest rates may affect 2020 For
35:20
venture. Well, I mean, I do try to avoid predictions, right? To the extent I can, I’d love to, I just want to like analyze, which is why we always invite you, I just like to present and like talk about the past a little bit and use it as a guide. Look, the low interest rate environment enabled certain business models that obviously do not work in slightly harder business requirements, like lending. One of the next startups, you know, assumed they would borrow borrow at six and lend at eight but now they’re boring at 10 and lending at eight Well, it’s a problem and then they want to lend that 12 But the customers don’t borrow at 12 the low interest rate environment created interesting models in real estate where people were holding the assets but in a low interest rates environment. Home prices only went up now home prices are actually going down so holding homes in your books is a catastrophe which is really one when various could quote that was said about Open Doors early days as you’re collecting pennies in front of steamroller you are collecting pennies and superlatives, Kuma, so the low rate interest rate environment created a lot of it, that’s gone. And when the Fed talks about reducing the interest rate from 6.5 to 5.5 to five, that’s still got the woods of zero and we’re very far from zero interest rates. Right. Now, districts also create what I like to call zero interest rate behaviors, zero interest rate habits, I think, you know, this year, convoy died. That was rough zoom pizza died, right, you know, pizza making company with a robot. I call that a low interest rate phenomenon. Right? Okay, so the pizza was gonna come to you, but actually, he was going to be, you know, cooked by the robot. And you can buy a Domino’s Pizza for 10 bucks. 799 sometimes, why are we talking about here? It’s two
37:01
degrees warmer with Zoom. Because, yeah, but then when you
37:05
get it, it’s always so hot that it burns the top of your top the top of your mouth. Actually, it’s a feature not a bug.
37:11
I mean, that’s classic over design for a problem that doesn’t really exist. Exactly.
37:16
And what are the over design come from a moment of early stage capital? What is the abundance of research habit comes from the low interest rate created a lot of over design, I mean, really articulated connection much better than I than I have in the past. So that was like crazy humanoid robotics ideas. So that was HR benefits. We’re gonna get, you know, we’re gonna sell kombucha on HR. And you saw that was coming in, we see what’s coming on the buses that’s coming from the decrepit massagers or the carpet this, you saw that was like online mental health. Obviously, we all need, at some point in our life, some form of mental health, not quite, you know, not 10 sessions a month. And, you know, the world should not be a bunch of people therapy, giving therapies each other and not doing any work. And that roughly was the world being portrayed by some of the more interesting mental health companies and just not not happening anymore. No, that’s seven cardinal sins,
38:12
gluttony and sloth
38:15
and gluttony startups that are gone. And but we now have a pill
38:20
for it. We have a zentech. Yeah, I am. I
38:23
would love to, I would love to just my wife and I were talking about this, I have no idea. I’m not a I have no idea. I’m dying to know, if in 20 years, it will be like, in the same bag as like cigarettes and asbestos. And like just all the lawsuits going after automakers, or whether it’s going to be like, this is the pill that saved humanity. I have no idea. It’s too good to be true. And so I’m slightly on the first camp to be honest, it just just can’t be. You know, it can be that simple. Take a pill Slayer with a Moloch. But that’s
38:59
America. Rami, we need a pill, a pill for everything. Everyone wants a pill. Everyone wants a quick fix. Nobody wants maintenance. People don’t want to put in the hard work. You know, they don’t want to go in concern. They want a quick fix. And honestly, we are in the ozempic Venture environment right now, as I think about it, like, we’re kind of forced in this. We were in abundance, and now we’re in scarcity and everyone has to tighten their belts, or give up or maybe they won’t have the choice. Maybe it’s too late.
39:29
Yes, that’s my point. The key to lose weight is to stop eating not to keep eating and take a pill. Because then it’s like venture capital and the low interest rate environment. Like keep burning and I’m gonna keep giving to you and you’re gonna be fine. It just just doesn’t make sense. Right?
39:45
Right. It’s an it’s an Yeah, it’s an external instead of an internal Alright, give me your bearish and bullish for 2020 for maybe three segments, categories sectors, you know, what do you what are you most bullish on give Meet your top three, you’re very excited about and you’re most bearish. I’m
40:04
very excited about areas of AI outside of large language models, improvement in machine learning, predictive analytics voice is fascinating. I’ve worked in voice for many years, and it was always an input medium or an output medium was never a substitute for human connection. Now, for the first time, it is the text to speech engines. I still don’t know what applications will be, but the technology is like I could, I could have a podcast with a computer right now. He’s not
40:31
nuts next year, it won’t be the two of us. As you mentioned that just yesterday, my son who’s six years old, I asked Alexa something yesterday morning, and she said, Good morning, Nick. And my son said, Oh, she’s so nice. And so it’s like, anthropomorphize the technology. Like, he’s only six, but he sees it as a thinking feeling. thing. You know,
40:57
I am sometimes less excited about a specific technology, and more excited about trends happening around this, because I think most of the amazing the biggest outcomes in tech happened less from the cabin based on a very small technological improvement that was resulted in a product at the perfect time, similar to what you said at the start of this show, where you said that technology is just by itself is not sufficient. I mean, think of what made Uber work so well, or what made Ubers possible. GPS chips and phones. Yep, all it is, right, you know, what many phones
41:29
and then GPS chips. And what made Netflix possible,
41:33
slightly, you know, enough DSL in every home. And what made figma possible, you know, WebSockets, which basically allowed a very small improvement, WebSockets just allowed the web, the server to push content back to the browser. So the website became alive, and could get updated without the user’s manual, refreshing the page, web sockets ever made it into like the top of TechCrunch, you know, 20, great innovations of 2013, that that was Google Glass and a bunch of other things. So there’s a lot of small ecological improvements, interesting things happens in the open source world, there’s interesting things happen, that we find fascinating.
42:12
Part of the reason why I’m skeptical when I have VCs on this program, when they’re overly prescriptive about their thesis, and micro about all the opportunities is because it’s just what you’re saying. Like a lot of these subtleties are completely missed. If you go back in history and look at thesis construction. And the ones that get it right, often, they might have some macro awareness and thesis on big trends. But they let the entrepreneurs guide the insights around WebSockets or it’s actually the GPS chip in these phones that’s going to unlock the ability to do the startup at this time.
42:48
100% I mean, there’s a beautiful, beautiful talk by an if you have if you know the founder of by the there is a beautiful, beautiful talk by Bill Gross, you know, Bill Gross started.
43:01
Yeah, beta works or not better works. Better. Idea labs. Idealab. That’s right. I like created
43:08
so many companies, 100 companies, so many, like, all the early tech companies came out of IDEA labs. And Bill Gross has a great YouTube video, everybody should watch, called the single biggest reason why startups succeed. And it’s time. What is it that timing is a combination of technology, enabling a product that latches into the zeitgeist of the moment, right was Uber, GPS chips on phones plus a Great Depression. It’s a great recession, where everybody’s looking for a job. And now Hey, you can go make money driving a car. Magical. I want to think of the best three performing companies in our portfolio in 2023. And the companies that have grown teknicks, basically, are 5x. These are early stage companies aren’t large couple. One of them actually is leveraging large language models in a very esoteric industry, in the automotive repair industry. One of them actually is leveraging the rise of pickleball. It’s a platform for that, again, I don’t want to use this podcast to like promote our our companies. I’m just trying to give examples of, of what it means to be in the right, zeitgeist, Americans love forks, that Americans are just not willing to pay for sports training except for their kids or their experts. Right. And so for the past 25 years, only way you could make money for training was children training or, but it was pickleball what we’re seeing is so many people want to take classes. And so this company is just doing an online marketplace, connecting consumers with coaches to go meet in physical life and play pickleball and then they started with pickleball and exploded in your mouth at five, five new supports and the company is called Teach Me and the you know, we’ll see when this company goes but it continues with the growth it’s at right now. You know, it could be an interesting range. portfolio company admitted for years when he asked me what made this company successful, I’d say pickleball, literally, one VC is going to talk about pickleball. Well, we know because there were articles written in The New York Times this year about how Pickleball is this fascinating phenomenon. We as VCs, can’t go to our LPs and be like, we are investing in the rise of, we’re gonna see the pickleball rise, you got to talk about technology and the future and whatnot. Third company in our portfolio that has also exploded this year, had to do also was we lost, it’s a company called Fe that’s a marketplace for nutritionists, for patients to talk to nutritionists online. And apparently, this is like the biggest trend in online healthcare today, where patients talk regularly to nutritionists, humans, about their needs. And this was not possible four years ago, because pre COVID insurance will not pay for it. But during COVID Insurance out of the payment code. So now it’s possible. Food is important. As you as you mentioned, this was epic. So, you know, my advice to everyone is look around you look for signs of changing consumer behavior. Yes. Look for signs of evolution of patterns, look for startups that are matching on these and just invest? Yes,
46:19
consumer behavior, market behavior, technology shifts, regulatory shifts, all these big macro shifts create opportunities. If you’re looking closely enough, you can you can find them. And then finally, you know, on a positive note here from me, what are you bearish on? Give us your top three most bearish areas.
46:39
Look, I’m very bearish on any capital intensive startup. It’s obviously not particularly profound. But one area that is going to particularly struggle, our services enabled software startups or software enabled services startups, depending on how the founders pitching it to the VC, the software enabled services are going to be tough, not an area that investors want to back even though I think you might still be able to build massive businesses they are so just right now are too skittish from the rising capital costs are too scared or too scared that downstream investors won’t finance it. I think if you just have the balls to go invest in the next time you bank in Africa, that will be impressive. International Investing and ideas. Oh my god, if you think the US is crossed, there, the market is so much for innovation, there’s very few incumbents, but the VC financing totally dried. Totally brought. But you have secret financing is a big risk, big risk.
47:34
big risk. Love it. Love it. So I want to circle back quickly. So with regards to data, which of the big tech companies do you think is going to become the dominant force leveraging AI to its advantage? I think there’s there’s some obvious choices. But I’ll tell you, we were talking about it as a team yesterday, and my my colleague, Nate, he made a really good point that Apple could have the most to gain the most to lose from leveraging AI correctly, because they may have the most data on people, their behaviors, their biology, everything. You know, Apple has probably the best data hooks in into consumers. And so I’d love to get your take on what’s best for
48:18
the consumer. But Apple is also has the most stringent security requirements, privacy requirements, and it’s part of their brand. I’m shocked every time I go to the Apple run different my credit card, and she’s like, please enter your email. I’m like, I’ve done that before. She’s like, Yeah, but we don’t store this information. I’m like, You’re the only ones on the planet anywhere else. Last shot more than once I give them my credit card. Tell me, would you like me to send you the receipt by email, right? Except at Apple. And I really respect him for that. Apple has so much consumer data and so much loyalty. And you’re ready, you know, and they have the photo. But there’s also very, I don’t know how comfortable they are pushing the envelope on privacy or training on data, that there’s gonna be a cultural battle they’ll have to deal with. So Google, on the other hand, I think actually has the capability doesn’t suffer from concerns about privacy, that Apple hold tie. They still have good data, they have good data, but they don’t seem to have the will. Just don’t seem to care. It’s mind boggling.
49:19
Low hanging fruit in their existing.
49:21
I mean, again, Google Doc is the only like, online thing that doesn’t have any AI in it. So natto like my superhuman is summarizing my emails my notion is writing for me like guys, what are you gonna ship it?
49:34
The fact that superhuman is so far ahead at Gmail is mind boggling.
49:39
It’s just, you know, just too much money too many millionaires too much. It’s it’s yeah, I’m very I just don’t even have a lot to execute to be honest. And it saddens me because I think there’s tremendously intelligent people who work there. Just to reminder that the culture doesn’t I think Amazon is a very interesting one because no Once you’ve kind of had, like I said, connections with us is a very complex one. We’re all customers of Amazon. Then we also some people shopping Whole Foods, which is another touch point, within everything we use running on AWS servers. So almost like they’re coming at us from many different points that don’t seem all connected. At one point, they actually connect it all up very, very, almost scary place, and they have very strong AI capabilities. You know, a few years ago, I went to, I went to this like artsy, sort of like a one day event where people like do funky project. And one project was called previous day delivery. And the idea is that Amazon as technology has gotten to the point where they can figure out what you’re going to order the day before, which makes sense, because they just delivered it that day, instead of waiting for you to order it. And then we actually were assigned a task of going to talk to people and be like, Hey, this is your Amazon order. What is that you’re anything like? Yeah, no, no, this is the, you’re gonna order it tomorrow. But we just wanted to make that today into the cplg actions. It was mind blowing.
51:05
So if it’s accurate, then they’re gonna be like, holy shit. It’s like the project with Google years ago, right?
51:11
Precisely. So I think Amazon has a fascinating hole in that space. I think open AI is going to need to figure out whether it’s a consumer company or enterprise company, whether it’s a platform company application company, has a long, they’ve done a good job with, with revenue with making the product accessible. But we’re already seeing more specialized versions gain a lot of traction, like for health care, and so forth. It would be interesting to see what open AI becomes, but I doubt it would be the winner here.
51:38
So your vote is on Amazon. Yeah, you know, a couple of years ago, it was kind of funny, because like one in every two venture dollars went to advertising, like Google and Facebook. And I feel like one in every two venture dollars in the next era here is gonna go to the MLMs. And the compute, and they’re gonna win, they’re gonna win peg. We’re very good. Just to wrap up here. Rami, what is the biggest thing you’ll do differently in 2024, than you did in 2023,
52:04
I am going to embrace being in New York DC, I moved from San Francisco to New York, the secret is out. And in January, and I had been flying to LA to SF, and I’m sorry, realizing that there is an insane, insane number of founders that actually are moving to New York. And I’m actually long SFI to maintain a place there. It’s kind of my home, I feel a bit like a foreigner in this town. Yeah. And we go back to SF, probably a week a month anyway. But I am going to embrace being in New York, DC, trying to find, you know, meet those founders or roaming the streets, who would love to meet somebody in person. And I will try to just, you know, bring the San Francisco VC to, to New York, whatever that means. I think there’s a certain culture is a little different here, which is has been one part of my hesitation to like, define myself as a New York VC. But I just want to embrace being here more and more people in person.
53:00
Well, the firm is 1984. The man is Rami, a deep and he can now be found in New York City. The big secrets out thank you, my friend for joining us today. And looking forward to the next one.
53:13
Thank you very much, Nick.
53:20
All right, that’ll wrap up today’s interview. If you enjoyed the episode or a previous one, let the guests know about it. Share your thoughts on social or shoot him an email. Let them know what particularly resonated with you. I can’t tell you how much I appreciate that some of the smartest folks in venture are willing to take the time and share their insights with us. If you feel the same, a compliment goes a long way. Okay, that’s a wrap for today. Until next time, remember to over prepare, choose carefully and invest confidently thanks so much for listening