In short
Odd Lots Podcast Episode Summary
Episode Title
Ethan Kurzweil on Venture Investing in the Post-ZIRP, AI Era
Hosts
- Joe Weisenthal
- Tracy Alloway
Guest
- Ethan Kurzweil: Founder and Managing Partner of Chemistry VC, formerly with Bessemer Venture Partners.
---
Episode Overview This episode discusses the evolving landscape of venture capital in a post-ZIRP (Zero Interest Rate Policy) environment, particularly in the context of artificial intelligence (AI) and changing economic conditions. Recorded live in San Francisco, the conversation dives into the implications of rising interest rates, the scarcity of computing power, and Kurzweil's approach to venture investing with his new firm, Chemistry.
---
Key Topics Discussed
- The Shift from the ZIRP Era
- Historical Reflection: In the 2010s, venture capital flourished due to cheap capital, leading to an oversupply of funds and overcapitalization in many startups.
- Current Landscape: Venture capital today is different due to rising interest rates and a more cautious investment environment.
- Introduction of Chemistry VC
- Reason for Establishing Chemistry: Kurzweil believes there is a need for a venture fund that prioritizes personal relationships and hands-on service, which has been diluted in larger firms.
- Fundraising Experience: Chemistry raised $350 million for its first fund amid skepticism from limited partners (LPs) regarding the performance of the venture capital asset class.
- Evaluating Companies in the AI Era
- Impact of AI: AI technologies have transformed how startups can operate, making it easier for entrepreneurs to deliver sophisticated solutions without extensive technical backgrounds.
- Investment Criteria: The evaluation process still centers on market potential and business viability, but there is a broader reference frame for what is achievable with AI.
- The "NVIDIA Tax"
- Cost of Technology: Startups today often have to allocate significant portions of their budgets to access advanced AI technologies, which can affect their financial strategies.
- Balancing Act: Founders must balance investing in high-quality technology while ensuring business sustainability.
- Future of AI and Technology
- Predicted Trends: Kurzweil indicates that while AI will lead to new applications and efficiencies, there will also be cycles of hype and disappointment, reminiscent of past technology trends.
- Human-AI Collaboration: The future will still require human oversight and creative direction, with AI serving as a tool rather than a complete replacement.
- Macro Economic Considerations
- Interest Rates and Valuations: Kurzweil discusses how rising interest rates should theoretically dampen investments, yet excitement around AI has led to high valuations.
- Market Dynamics: The interplay between private and public valuations remains crucial as the tech landscape evolves.
---
Key Takeaways
- New Era of Venture Capital: The transition from zero interest rates to a more complex financial environment requires a reevaluation of venture strategies.
- Chemistry VC's Unique Approach: Emphasis on personal relationships and early-stage investments aims to fill a gap that large firms have neglected.
- AI as an Enabler: While AI democratizes technology access, evaluating its impact and potential remains a critical skill for investors.
- Market Cycles: Awareness of potential cycles of disappointment in AI's promises is essential as the technology matures.
---
Conclusion Ethan Kurzweil's insights provide a valuable perspective on the current and future landscape of venture capital, especially concerning AI's role in shaping investments. The conversation highlights the importance of adaptability and strategic foresight in navigating the complexities of today's economic environment.
---
Further Listening For more episodes of Odd Lots, visit [Bloomberg's Odd Lots](https://www.bloomberg.com/oddlots) for transcripts and additional content.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Transcript
Automatic transcript. May contain errors.0:00Your best bottling plant employs 3 ,300 people. How do you get 3 ,300 people working at peak efficiency? Your best store has reduced waste, water, and energy usage. How do you make every store like your best store? Your best property has every guest raving. How do you make every property like your best property? The answer is Ecolab. Better performance, better outcomes, better impact. Ecolab. Now every location is your best location. For enterprise organizations, managing all your food needs is a tall order. But with EasyCater, you get a single workplace food vendor with the tools and resources to make it easy.
0:40Giving teams across your organization an easy way to order from a huge variety of restaurants, all on one platform. All while consolidating your corporate food spend so you can control costs, streamline billing and payment, and simplify reporting. EasyCater, your business tool for food. To learn more, visit easycater.com slash podcast.
1:05Bloomberg Audio Studios. Podcasts, radio, news.
1:21Hello and welcome to another episode of the Odd Lots podcast. I'm Joe Weisenthal. And I'm Traci Alloway. Odd Lots listeners, you are going to be listening to a special recording of the podcast, one that we recorded live in San Francisco. Yep, that's right. This was a conversation that we had at the San Francisco MoMA on November 20th. It was an event sponsored by Principal Asset Management. And our guest is Ethan Kurzweil, the founder and managing partner of Chemistry VC. Yep. We talked about all things tech, software investing, how investing today is different than it was, say, in 2014 when rates are at zero.
2:04We obviously talked about AI and how that changes the game of software investing. Take a listen. Thrilled to be here with the perfect guest, Ethan Kurzweil, at his new fund, Chemistry. It basically launched like three weeks ago or something like that. And prior to that, 16 years at Bessemer. So literally the perfect guest to talk about, you know, VCs, the landscape changing over time or something like that. Well, thanks for having me. This is actually the first episode of anything we've done since we launched Chemistry. So that's very exciting. We're thrilled. There is obviously so much. There's so much we could talk about.
2:44We talk about the macro environment. talk about AI, we could talk about the political environment. Maybe we'll touch a little bit on all of it. So I'm just going to ask like a really simple question to kick it off, which is in the 2010s, you know, people talked about the Zerp era and some people even look on that period quite fondly right now with nostalgia, even though at the time Zerp was sort of seen as like this negative thing. Didn't seem that bad out here though. Strictly from a macro standpoint, you've been in this game, so to speak, for a long time. What's the difference right now versus, say, if we were having this conversation in 2014?
3:21Oh, the good old days of 2014. I miss those days, too. I wish we could go back. So right now, there's lots of things happening in sort of the tech landscape broadly, as well as like venture. And so maybe just taking a few around venture, you had this era of explosion of different things, lots of different funds, new products, money being kind invested in the asset class beyond what it could take, beyond the capacity of those companies to absorb the capital and do good things with it. I'm an optimist about tech and venture. I think more is generally better, but there's a limit to that. I think everyone would now agree in hindsight we went a little bit beyond that limit.
3:59Now we're in this new era where that's happened. We're digesting the impact of that, of all this capital coming into the space. and you have this kind of new technology phenomenon. And by the way, it's not really new. It's maybe new as it applied to startups. You've heard about it for a while. I've been hearing about AI for, I don't know, a few decades or something like that. We'll talk about that. We'll get there. But that's now kind of the building blocks are now there. The technology startups without a lot of capital can take advantage of it. And so that's getting people kind of very, very excited again.
4:31And even as everyone knows, there's still this fresh memory in everyone's head of how we kind of overcapitalized everything. And so those two forces are sort of countervailing. And it's having some interesting impacts that I think we'll probably get into. We definitely will. The new fund, why does the world need a new venture capital fund? Or if I was going to phrase it more diplomatically, like what is it that you can do at chemistry that you couldn't do at Bessemer? The world does not need a new venture capital fund. That's the last thing the world needs. After chemistry launched. That was the last one we needed.
5:04Even before chemistry launched. OK, OK. We were oversupplied on venture capital funds. But the world does need the right venture capital fund. And I'll get to why we launched chemistry in a second. But I do think the effect of the capital that came into the asset class over the past kind of five to 10 years has been to create a little bit of a misalignment. A misalignment between LPs, that's who invests in venture firms, and the venture managers, and then a misalignment with founders, ultimately. And that's what got us this sort of passionate idea to bring venture back to its roots. Chemistry is sort of a simple idea.
5:36It's a boutique venture firm. It's small. It's designed to scale very slowly. We are not a hyper-growth startup, even though we try to find those to invest in. We want to bring some sort of personal service back to venture capital. We don't have big teams of people. We are the portfolio services team that works kind of hands-on with our startups. And that ethos we felt like was missing from a lot of the way that kind of as the asset cuts got institutionalized, you lost a little bit of the personality and the personal relationships. And we felt like it didn't have to be that way. There's nothing bad about the way venture used to be practiced.
6:09And that really it just became so missing that we felt like, OK, we'll go do this. Let's say I have a lot of money. I'm an endowment or whatever. And I'm thinking about allocating money to a VC fund or firm. That sounds really nice, personal relationships, all that. But mostly I just care about getting returns. And let's say my assumption is, okay, yeah, again, it all sounds very nice, but there are advantages to scale. There's deal flow that large firms see that maybe they're the first call in some round or something like that. Why would that be wrong? It's not wrong, but it's not the only way to practice venture.
6:51There's definitely advantages to scale, but I think it comes at a cost of being able to focus uniquely on companies that are at this inflection point moment, this pre-inflection point moment where they're about to take off. Because when you have a lot of capital to manage, you're going to make decisions that aren't necessarily about how do I find that company that needs a$3 to$7 million check at that moment. You're thinking about how do I move the merchandise, move the money that I have in the system. And so it may be appropriate to make that investment, but it may be appropriate to make a whole host of others that are at cross purposes with finding the one defining company of that era.
7:30And I think for us that have been experienced at working at venture firms and identifying the patterns that lead to that, we felt like we could pick those out pretty well. and that we would have a good sense of where to spend our time without the resources and without the brand magnets of other firms and that it's a small community. We could get our brand out there pretty quickly to folks that are used to identifying those patterns and referring those deals on to us. So I think you just finished your first fundraising. Was it$350 million? $350 million was the first fund, that's right. What was the fundraising experience like now versus, say, going back to the good old days of 2014?
8:10Oh, yeah. Oh, yeah. Good question. So it's different in two respects for us because we were a new entity too. And so we had a whole bunch of vetting around, hey, what's our track record and experience? Do founders want to work with us? Because this whole premise was on we're going to bring the individual personal service back. We need to be able to say our personal service is good. Like you want to work with the chemistry team because we're known for that. So there was a lot of vetting around that. There were LPs that felt like the asset class had under-delivered. Those were generally, you know, came into conversations very skeptical.
8:45And our argument to them, and some of them invested, some of them didn't, but our argument to them was, look, that's true, writ large, but by having exposure to just the earliest stages of the asset class, going back 40 years, that phase of the market has always performed. If you took a slice of the venture market and looked at just early stage investing, just the phase of your typical kind of series A and series B investment. Maybe the median fund hasn't performed, but there's always been outlier funds throughout that period. If you looked at all of the asset class writ large, including all the growth checks, the leader stage investments, the sort of pre-IPO rounds that came on, that asset class has really underperformed over the last five years.
9:25So we were sort of orienting around, we give you exposure to just the early stages, and we don't want to do anything else. We formed the firm just to do that. So there's no guarantees ever in venture, and we know there's outliers. But the basic idea here is that venture may be cyclical or maybe structural, but early stage is not cyclical in the same way that these potential returns have been stable at this level. If we make the right number of investments, we have to make the right investments, and we have to get a few right. Maybe there's a little luck involved. But if we do that right, that will outperform.
10:01We won't water it down with bad investments later. Let's talk about how to make good investments then, because that's really what matters. Obviously, the 2010s, the sort of cheap cloud computing and all the SaaS trends that made people a fortune. How does evaluating a company today, and this is where like, I guess the AI part comes in, whether the company is AI specific or in some level is going to be plugged into an AI model somewhere. How does that make the process of evaluating a company different? It makes it radically different and exactly the same all at the same time. All right, so what do I mean by that?
10:42Radically different in that the entrepreneur can now promise pretty incredible things. You can talk to a system and get it to code for you. It's something that three or four years ago you would have said, sure, good luck with that. You need some engineers on your team. You can now make promises like that. But ultimately, the way we're evaluating companies is thinking about the end market that they serve, the business user or the consumer, and how are their lives made better? How has this process improved? If you're making a consumer video editing app, how is that awesome for consumers to use? And so you start with the question of can the technology deliver what the entrepreneur says?
11:22That's radically different. And then you step back to, hey, is this a good business opportunity or not? Is this something that people will pay a lot of money for or that will be able to monetize itself in some other way? That's very similar to how we've always done the job. Is there a difference in the sort of due diligence process for AI versus old school SaaS? Not terribly. Honestly, old school SaaS often had a data element to it that's somewhat similar to AI, how they harness data in the application. What's different now is you can apply a frame of reference of what's possible that's just much broader, that's just much more interesting, that's just much more potentially transformative to business users or consumers.
12:06That's a little different. You might not be as skeptical about a founder's ability to deliver. There's this whole democratizing element to AI, just riff on that for a second, in that you maybe don't have to have the most ultra-specialized skill set of engineer to be able to deliver something pretty transformative. And so if you back up from that and think about a due diligence process, you don't necessarily need to spend as much time questioning the entrepreneur's ability to deliver. And there is this maxim that most entrepreneurs will build what they want to build. Just, is that the right thing?
12:42And is the timing right? With the AI era, you get that on steroids. Most products can be built the way the entrepreneur says them. Now, will they have the impact that the entrepreneur thinks they'll have? That's still a question that we have to answer when we make our judgments. What's the differentiator in that case? If it's not necessarily about the skill set of the engineer, what is it that makes you think an AI project is better than another AI project? Ultimately, it comes back to what impact will it have in the market. I don't think about AI as a category so much. I think about AI as an enabling tech, just like cloud computing or mobile or mainframes back in the day, or data center technology.
13:20It's just a way of building tech that can potentially allow an entrepreneur more weapons to be able to deploy. But there's nothing inherently like the end user that's using a finance application or that's using a communication app. At the end of the day, they're using that app because they want to do something with it. They want to communicate. They want to run their expense reconciliation process. They want to do X, Y, or Z. There's nothing different about AI that makes that any bit of a different analysis than we had before around what is the impact that that particular product is going to have in the world.
14:10Your best bottling plant employs 3 ,300 people. How do you get 3 ,300 people working at peak efficiency? Your best store has reduced waste, water, and energy usage. How do you make every store like your best store? Your best property has every guest raving. How do you make every property like your best property? The answer is Ecolab. Better performance. Better outcomes. Better impact. Ecolab. Now every location is your best location. For enterprise organizations, managing all your food needs is a tall order. But with EasyCater, you get a single workplace food vendor with the tools and resources to make it easy.
14:51Giving teams across your organization an easy way to order from a huge variety of restaurants, all on one platform. All while consolidating your corporate food spend so you can control costs, streamline billing and payment, and simplify reporting. EasyCater, your business tool for food. To learn more, visit easycater.com slash podcast. Today, we got NVIDIA Earnings, a company that people may have heard of, and they were really strong. I think the stock slipped a little bit. Is that an important company? Yeah, and Jensen Wong saying, you know, AI is full steam ahead, and they have all these scarcity.
15:28When you're writing a check to a company today, you know, one of the things that characterized the 2010s was just persistently falling cost of computing power. When you're writing a check, how much of that today is going to pay some sort of NVIDIA tax to have access to that? And how does that make the sort of capital decisions of a company or the types of companies that you're invested in going to look different than they were? Well, there's this interesting kind of two countervailing forces, because the more of your check that goes to pay a tax like an NVIDIA tax or an open AI tax or something, are being built on some – generally for us, it's being built on a model.
16:09They're not using the sort of bare metal of the GPU. There's puts and takes there. But the more you invest in that, the more you've got your own technology that's more defensible. So a lot of times we're seeing open source. So a lot of times we're seeing not a lot of money go towards that, but they're deploying open source tools or they're built on models that are freely available to anybody. And so it's a question of, can the founder or the entrepreneur make a process improvement or productize commercially available technology to everyone in a radically different, unique, 10x better way, so much better of a user experience.
16:46The deep tech founders who are doing what you're saying, where a lot of the check goes to building the core technology, you have to believe then in the business outcome being so great, that it's worth it. It's worth this huge R &D investment or this huge investment in training specialized models. But just to be clear, you say, OK, the non-deep tech ones that are using some existing models, is the amount of money that's going to, say, an open AI or some entity that already built the model. Is that fundamentally look different than, say, the expense sheet of another software company in 2014 when they think about how much outside tech they're paying for?
17:26It's not radically different for the ones that are the thin layer. Yeah, the thin layer. The thin layer ones are not that radically different. Think of it as a small incremental tax on top of their Amazon Web Services bill that they might already be paying. And the technology is so good. It's so performant. It's so available to everyone that most of the companies we look at, because we believe in the lean startup. And most startups that can be built on that kind of technology will build on that kind of technology. It's not a huge tax and cost. The huge tax and cost comes when you try to sell it.
17:56And you scale up the go-to-market operation, the sales and marketing. But the tech itself, there's only a handful of companies where that's a real barrier to entry. And there are some. Just on this sort of big versus small point, I think one of the weirdest things about the sudden rise of AI over the past couple of years has been the fact that Microsoft has been really good at it, which I think three years or so, no one would have expected. Going forward, do you think, who's going to be the best at this? Is it going to be the incumbents who now have a head start, who have the deep pockets, the access to data?
18:33Or is it going to be the leaner startups who are maybe experimenting with new things and building on top of existing models? Our view would be that the foundational layer, like the model layers that a lot of people build on top of, or that we as consumers use for sort of our basic kind of chatbot style applications, is going to go to the big players plus maybe one or two new entrants. And that looks like that game is sort of established. I mean, I don't know if you count OpenAI as a separate company from Microsoft, but they're clearly around to stay. And maybe there'll be one or two others. But that's not, in our view, a humongous startup opportunity because there's such amazing capital investments that need to be made there.
19:10On top of that, how do we take that tech, take those abilities that the amazing researchers at OpenAI, aided by Microsoft and others, have built and make it useful to the end consumer and to the end business user? I think that's where we're going to see kind of this new era, kind of like we saw of cloud computing, where there were a few early entrants in HR, tech and financial applications and things like that. And then this explosion of cloud computing applications that disrupted the status quo. The platforms that emerged to dominance in 2010 just exerted to varying degrees, but just tremendous lock-in for their clients.
19:47I'm not talking in the formal legal sense, although maybe we'll get to this, but monopolies. But de facto, just some truly without competition. And there's like this debate about like, when it comes to these foundational models, there seem to be, you know, there's a lot of entities. There's not thousands, but there's quite a few that can make incredibly impressive performant models, some open source, some closed source. Do you see any of them emerging with the same sort of like true lock-in kind of dominance? Or when you look at the companies that you're funding, do they seem like issues like, yeah, we could use OpenAI, but also without too much trouble, we could switch to another provider fairly trivially?
20:32I think it's a really good question. I think the lock-in is not too dissimilar from the cloud era, where you could switch off of one cloud computing vendor from one to the other. But there weren't that many of them. Now, in this era, there might be a few more. And I think there's no open source cloud computing provider. Someone's got to plug in the hardware, air condition the data center, make the networking work. With large language models, there are open source models that are going to get to be pretty good. And so I think that's another element here that's a little different from the cloud era, where it probably allows a little more fluidity than even you have among clouds.
21:09But there's not going to be dozens and dozens of models. Because to be performant, to be human-like, be able to provide people with responses that make sense, that have emotion, that really fulfill on the promise of what AI can do, there's so much money needed to that that there's not that many companies that can capitalize on it. But setting aside the big foundational models themselves, what's the coolest application of AI that you've seen so far that's sort of built on the big guys? Well, the consumer applications, the companion apps are probably the coolest right now. They're not very realistic yet, although they're sort of getting up there in that they start to emulate real people in the world.
21:52Like podcasters? I think there should be a Tracy character app that's out there in the public. In fact, with a Google system, you can actually create a whole podcast from a notebook that you submit to the application. Some people have done some odd thoughts versions. It's lacking a little color. No, but this is really important because I've listened to some of those Google creators. I mentioned this on another episode. I've listened to some of those, and they're not as good as me and Tracy are. Nothing could be. Completely unbiased. But they're not terrible. It sort of disturbed me because I listened to the AI-generated podcast about some document that the Department of Energy made.
22:36And I was like, oh, shoot, this isn't that bad. It's not totally boring. It's not a terrible way of consuming that content. So someone I know well had to read an entire book and generate a 12-minute podcast on that book. And it was not – I totally agree. It was sort of like the personality was lacking. That's right. And they tried to make it personable and it just fell flat. And I think that's a little bit what's missing today. But AI will get better. Yeah. Not as good as you guys. That's true. We're due. No odd loss. Well, okay. I'm just going to ask this question. You know, most people in this room have probably been talking a lot about AI for two years.
23:21Now, probably in this room, it's three years. In the rest of the country, it's probably about two years. You, as you alluded to, have been probably thinking about AI in some respect for 30, probably 40 years. Tell us a little bit about your background, having thought about this for at least three or four decades longer than the rest of us. And how does that inform when you make predictions now, when you try to pick winners, how does 40 years worth of experience inform your choices today? All right, so a confession. The book that the podcast that I just mentioned is written about was my dad's book, which you're alluding to.
23:58That's perfect. My father, an AI technology futurist, has been thinking about AI for about 65 years, is what he would say, and large language models for 40, because that's his field, too, is pattern recognition, is being able to recognize patterns and apply them to language. That was one of his first companies was that. So did the podcast do a good job of talking about your dad? So I was debating with him. He thought it was great, and I said, I think the podcasters aren't as good as Tracy and Joe. Thank you. Without your names, that's sort of what I said, and that was the debate we had. But it got the substance right.
24:29But yes, to your question, AI, and it's maybe made me both more excited about AI and slightly more cynical about this moment because AI has been around for a long time, and now it's becoming – And hype cycles, there have been many – yeah. And we will be in a disappointment about what AI brings hype cycle in about, by my calculations, 2.6 months from now. And then we will come back out. That's a very precise estimate. And then we will come back out from that, and then it will start to happen. How many months after that? 4.3. Okay. Wait, what's the catalyst for disappointment? Yeah, yeah, what's coming?
Read the full transcript
25:04That some of the companies that have been hyped to fulfill on this promise of completely human-like, lifelike understanding, reasoning abilities, and emotion won't quite fulfill on that promise right away. And we'll have to wait another 4.9 months for that. Or longer. A couple years. I can't tell how much of a joke they said. Anyway. You know what else is coming in about two months is a new administration. I've heard about that. Yeah. Yeah, it's kind of been in the news. You talk to lots of people in the VC space and in the founder space. What are people saying about the incoming administration?
25:44What are the hopes, dreams, fears that people are talking about? There's people that are prominent out there that advocated for this or for or against it, that have their passionate points of view about why the new administration is good or bad. At the kind of surface level, day to day, this felt like, okay, there's a change coming. And there's just not a lot of translation between that and the day-to-day of venture capital. Because technology is this force that sort of plows through market cycles, technology administrations. Unless you're in a very highly regulated industry, crypto, for instance, or something like that.
26:23I feel like in my circles around how is AI going to be commercialized for business and consumer, it's not an event that people are as focused on as perhaps the most prominent personalities out there.
26:50How many vendors does it take to meet all your organization's food needs? Just one. EasyCater, the workplace food platform that lets teams order from a huge variety of restaurants, over 100 ,000 nationwide, all through a single vendor. In addition to all that variety, EasyCater also gives you full visibility of your organization's food spend with invoicing, centralized reporting, and seamless integration with expense management systems, all on one platform. Easy Cater, your business tool for food. To learn more, visit easycater.com slash podcast. You're thoughtful about where your money goes. You've got your core holdings, some recurring crypto buys, maybe even a few strategic options plays on the side.
27:36The point is, you're engaged with your investments, and Public gets that. That's why they built an investing platform for those who take it seriously. On Public, you can put together a multi-asset portfolio for the long haul. Stocks, bonds, options, crypto, it's all there. Plus an industry-leading 3.8 % APY high-yield cash account. Switch to the platform built for those who take investing seriously. Go to Public.com and earn an uncapped 1 % bonus when you transfer your portfolio. That's Public.com. Paid for by Public Investing. All investing involves the risk of loss, including loss of principal.
28:15Brokerage services for U.S.-listed registered securities, options and bonds, and a self-directed account are offered by Public Investing, Inc., member FINRA and SIPC. Crypto trading provided by Backed Crypto Solutions, LLC. Complete disclosures available at public.com slash disclosures. I'm very curious about, like, the sort of tech-inflected side of this administration, the influence of Elon Musk, J.D. Vance having been a VC. But the really low-hanging fruit policy question is on the merger side. And we don't know who's going to run the FTC. We don't know who's going to run the DOJ. It certainly seems plausible, however, that the new administration will have a much more liberal attitude towards letting mergers go through.
28:54How much is just from a nuts and bolts standpoint, when you're thinking about returns, when you're thinking about investments, exits of various flavors, does a big sort of 90 degree turn or maybe 180 degree turn on merger policy change your thinking? Not a ton, but there's no question that the current administration, not just here, but in the EU and the UK, where any global merger now has to get through basically three antitrust bodies, has been really a lot tougher than any administration we've seen, Democrat or Republican, in the past. So when you're sitting with an entrepreneur and getting excited about some big dreams for tech, are you thinking about, well, I wonder what Lena Kahn's thinking about, you know, the consolidation of the market for design tools?
29:36No, not at all. But is it probably a good thing for entrepreneurs' options to be able to exit their business and our business, which relies on that? Yeah, probably it is. The reality of the prospect of an exit into machine design tools or the prospects of who would be a buyer, those kind of conversations weren't coming up in the early stages of a conversation with an entrepreneur? Not with an early stage founder. Now, in a growth context, it's tremendously important. Because if you don't have the option to exit for billions and billions of dollars, you can only go public. There's a pretty narrow set of criteria you have to meet to be able to go public.
30:10So then it's a pretty material thing. We're at chemistry focused on the earliest stages. Is this technology going to get out and have an impact? There, you just kind of take a flyer. you kind of assume that if it does, it'll be valuable in any kind of context, whether it's in an M &A one or some other exit. So another aspect of the incoming administration is they seem to be crypto friendly. And I'm going to co-opt a question that's been submitted by the audience. But what do you think about crypto in general as an investment? Is it something you're interested in? That's one where the administration probably matters a lot.
30:47I have been pro crypto for certain use cases in the past, thinking about what's the infrastructure layer needed to make crypto a part of the financial system. And so that's the security, the protocols, the permissioning, the privacy, all that kind of stuff, I think is really necessary because crypto is still such a wild west to where you have to be pretty deep in it to benefit from it. And so I still think there's that like bridge technology to make it useful for kind of everyone in their everyday lives, like, you know, the Coinbase is kind of wallet type software for everything else. And it's probably true to the extent you can kind of read the tea leaves on these things that the current administration is a lot more friendly there.
31:25At least that's messaging. How will that manifest itself in policy? No idea. But right now, a lot of people are scared of this space because there's a lot of uncertainty around it. Then the other element is just sort of the people in the orbit, you know, tech, accelerationism, exciting things, getting to Mars. Seneca said the Mars question specifically. a lot of it seems very vibes-based, and I don't know what policy levers any of that means. In your view, are there other policy levers that could be pulled that would be good for the American tech infrastructure, or sorry, not the industry? Probably, yes.
32:07It's really hard to start a company and have it be successful. There's so many things stacked against you. So what are the things you can do to remove all the unknown obstacles that might come up beyond the really hard ones of like, will you deliver the product on time? Will the product, can you deliver it for a reasonable cost and will it have an impact on the market? The regulation that kind of is another curveball that you might have to answer to, that's an impediment. That's a blockage that serves at the cost of innovation for sure. And so I think what probably has an impact, just maybe using crypto as an example, is clear regulation and lack of uncertainty, where you have a sense of what are the rules going to be.
32:51Not that we'll apply these arcane tests and we don't know exactly how a court will interpret it, but exactly do these six steps and you'll be fine. That's taking out any uncertainty that's beyond the sort of normal startup risks is a good thing for innovation. Another question from the audience. They mentioned that obviously we've been talking about AI a lot, and you talked about the disappointment and redemption cycle. Are there any other nascent tech areas or growth areas that you are excited about? Beyond AI. Yeah. Oh, we have to think beyond the AI Rubicon. Let me think about that. I mean, I think the democratization of tech broadly, this is aided by AI but not principally.
33:33the fact that a normal business user or even a consumer can now create an intelligent system. It doesn't have to be a coder necessarily to be able to write a complex kind of logic flow and be able to build an application or a messaging tool or something for a business context. I think that's pretty powerful. I mean, I've always been interested in the democratization of tech. Even cloud computing had a democratizing impact. Because you could give lots of people logins to a system and let them have impact, even if they weren't technical. You could let people kind of edit the flow on a website, personalize a page, be able to engage with their customers directly on the website without having to code anything.
34:12And so I think there's this democratizing aspect of the internet, of cloud computing. AI is a part of this that's going to give more people the ability to be more creative. And that's going to have a kind of second order impact on just the kinds of things we're going to be able to do, even for like little niche audiences. Do you see yourself writing checks to companies that are making apps for virtual reality goggles? For virtual reality what? Goggles. Possibly. I've written one before. Virtual reality. The goggles. Goggles. Yeah. Is that exciting to you? Yeah, virtual reality gaming could be a thing.
34:51Virtual reality messaging, communication, working in virtual reality. Before I was a venture capitalist, I worked at a company called Linden Lab, which is the company behind Second Life, if you remember. Oh, Second Life, that's a blast. It's not a core theme for us at chemistry, so odds are we won't, but I'm open to it. Sorry, I had a flashback to the time when you thought electric scooters were the future of transportation. They are. They're so great. They are. I love the electric scooters. It's part of the future. So I'm just going to, this will be now. I forget, when was that that we came out here and I was like, oh my God, Lime scooters are going to change the world.
35:28But for me, like I've taken, it's Waymo this time and like I'm just so completely Waymo pilled. I'm just going to, I'm just, it's just so amazing. I don't never, I never want to take an Uber again. The wow of the Waymo experience is greater than the wow of the Lime scooter experience. Yeah, it is. Because there's a lot less risk of death. Well, I mean, maybe if you think the Waymo might crash, but they don't. No, it feels so safe. I felt so comfortable. And actually, then I took an Uber today and it felt worse. And like it was, it was, it was a worse experience. And now when I go back to New York and take an Uber, it's like going back to the land of flip phones.
36:02You're going to have to move out to the West Coast. Yeah. Okay, so when it comes to AI, one of the debates that's been going on is like, well, do you invest in the actual AI companies or maybe you invest in sort of picks and shovels and data centers and things like that? Is that like on your radar at all? Or do you, this is a question from the audience, at a minimum, do you look at, for instance, investing in new technology that could help AI manage energy usage or something like that? I think there's a lot of second order effects of AI that we can make investments to make better. Energy usage being one of those or helping create the primitives to allow developers easier access to some of the more advanced functionalities of AI.
36:45There's a whole side theme that's maybe orthogonal to your question around the provenant. You don't really know what data has been inputted into an AI system is relevant to your answer. So it could have stolen some content or it could have, like, who knows what trained it to provide you with that particular thing that it said. And so there's a whole side theme of like, okay, how do you make that okay for the people that made the actual IP that that AI was trained on? So that's another kind of side theme of AI. Oh, interesting. So managing the IP. Managing the IP, the rights of that, the privacy.
37:20You might, for a base level, want an AI trained on a corpus of data that's pretty basic, but then the Taylor Swift of data, the really advanced IP holders, you want to pay more, but then you want to get some of the money to the people that created that IP that made that AI even better. That's hard to do right now, but maybe not yet solved. I want to go back to what you were saying about how having thought about in your life AI for four decades, that in some ways it makes you more optimistic because you see like this grand sweep, but also at least a temporary sort of cynicism because you know that AI winters exist.
38:02And it's very plausible. And, you know, there's all kinds of stories about, you know, running up against current limits of scaling and all of that. Can you tell us a story about what was the past AI winter that happened? What was something that at some point people were like, oh, we got this. This is moving. And then they ran into a wall. And what's a lesson that can be drawn from a past experience? Well, speech recognition was probably a wall where people thought you'd be able to talk to AI. Yeah. What was the years? In the 90s, that was one of the eras of AI where my father was involved with.
38:32In fact, he named one of his companies Kurzweil AI. And AI stood for applied intelligence, not artificial intelligence, because it was a bad word to say AI and have it mean artificial intelligence. Because it felt like it was under-delivering on an artificial intelligence. It was more applied than the AI that we think of today. And so that was an era where you could do discrete things. What was the moment, oh, this is not growing or scaling or improving the way we expected? People forget. And then something that counters that, a counterfactual comes out. and then everyone kind of hones in on that.
39:06And that period where people are forgetting about it, that's the trough of disillusionment, where the beginning of that is the trough of disillusionment period where there's a lot of prognostication about how this technology didn't deliver. Then people forget something does deliver and then move on to a new cycle and the expectations get high again that probably can't be met. All right, another question from the audience, also sort of a Trump-related question, polymarket and other prediction markets. Do you think those are like in for, well, where are they going? That's a really good question.
39:39I don't know. I mean, I think that probably, I believe that system had the best way of taking stock of kind of all the known universe of information that was out there and distilling it down into a, okay, what does it mean for a particular event like the election? So that's kind of cool. Now, is it legal? I don't know. It sounds like maybe not because somebody's going to jail, but. No one's going to jail yet. Somebody might face criminal prosecution about it. But it did seem very easy for an American to use it, which does not seem like it's supposed to. Right. People having real money on the line does create a more purist system that it's sort of hard to replicate with any other approach, similar to how the stock market sort of works.
40:22And in theory, it gives you kind of the right price of every particular asset that's listed. So I think for prediction markets, that incentive is hard to replicate any other way. Should there be prediction systems? And there's a policy question that I don't know. Speaking of prediction markets, and it's sort of a broader philosophical question I was wondering, like we live in an era of people betting on everything. And one aspect, I think, of people sort of betting and speculating on all kinds of things is that it seems to me that the sort of VC mindset of you want to just have a couple of gigantic winners and get that big score is spread to the non-VC world, right?
41:06And people really look for those right tail opportunities, both in investing in stocks, their careers. Do you perceive that the sort of VC worldview has seeped out of the VC realm and sort of, I don't want to say infected because that's like a bad word, but as a transfer - Are we all VC people now? That's exactly the question. Does it feel that way? 100 % what you're pointing out is true. And it'd be interesting to do root cause analysis. Like, are we to blame for that? Yeah. But first of all, is it a bad thing or not? Yeah. It's certainly happening. And I think the cause of it, I would say, is, well, there's probably a number of kind of psychological causes.
41:51But there's been this democratization of access to private assets that's happened over the last 10 years, too. We haven't talked about either where rather than trading, it used to be you traded stocks. Then it was sort of you could trade IPOs. And now there's some access for Main Street consumer to private company assets as well, more typically venture vetted. And there's investor protection laws, but the move has been towards more and more and more democratization. And so the VC way of thinking is just seeping into more things. Is that good? Is that bad? There's probably pros and cons. A lot of people here probably want to know how to spot winners in the market.
42:31But part of this is about avoiding losers as well. What's your best tip for spotting, I guess, or finding, identifying froth in tech? Well, I'm the wrong person to ask. Because we back as a good, even as good venture capitalists, we back so many losers. Like it's just an occupational hazard of the job. Now, you asked about froth though. And so that's the sort of like perception disconnect of like what's the reality of a particular tech? And I think you have to go to the source. You have to see like what impact is this having? Not what are other people saying about it? What are these kind of second order effects?
43:07What is, does the entrepreneur, you know, look the part in some way? Or are they kind of playing a role that makes their impact seem great? But like what's the impact of the technology? and kind of like have blinders on for the noise that's out there in the ecosystem. Because that's another impact of everyone, you know, more and more VC-like thinking is there's more and more hype around really exciting tech. Some of it's real, some of it's froth. We talked about this earlier, the fact that there are not terrible podcasts that are produced by AI. And it does cause me as a professional podcaster, like, yeah, it causes me anxiety.
43:47But like, this is the other sort of big question, the sort of future of labor question in a world where AI gets better and better. And like, what are we as humans good at? I mean, I'll start with that question. What are we, and I'm talking not just like in the next year or five years, but like in 20 years or 50 years. And I know your dad made predictions that were 50 and 60 years out. So you probably think – I imagine you also have in your mind predictions that are 50 and 60 years out. And so when you think about like what are humans good at, what are we going to be good at? People say the same thing about VC, by the way.
44:26It happened for many years. Like a – you'll be able to put this data into a system and it will be better. And you can't even really do it for stocks yet. Right. Probably at some level there will be systems that aid people. But I still think what tends to happen, I'm not the 50, 60 year out banker. I'm the sort of five year out kind of thing. But let's go with that time horizon. We tend to, as humans, kind of move up the stack. We still have to do the creative work. We still have to guide the AI systems. We still have to sort of harness the tech that's coming out of them, figure out how to apply it.
44:58We can't just plug in more energy into our systems. Are we just going to fall further? I think we'll still stay on top of the systems. We're going to tell the systems what to do. What challenges do we want them to solve? What's the problem space that we want them interested in? What's success look like for these systems? There's real work as we move higher and higher up, and we do less of the grunt work. That's the sort of pattern that has been existing today. So we've been focused on software for obvious reasons, but when do we get the good robots that can do the terrible jobs? I don't like the future where AI can do a podcast and write songs and poetry and all the fun stuff, but I still have to vacuum and fold my laundry.
45:42Fold clothes. There are robot vacuums. Yeah, okay. They're not that great. But there are – They can't fold the laundry. Yeah, talk about that. Yeah, I don't know. I mean, I have seen some systems now that have sort of like – or some robots on it that have kind of human-like characteristics. They can walk upstairs and carry things and pick things in a warehouse. And so because it's hardware, there's less of an exponential to that. So it's more a blocking and tackling around the sensors. And what's the cost of the particular parts that go into that? But cars now drive themselves. So that's a big step change versus what we used to have.
46:16Drive themselves in difficult environments with rain. So it's coming. Now, how soon and which industries is it going to hit? That's a hard one. The first question I asked you is like, how is a venture different today in 2024 versus 2014? And you gave a good answer to the question I asked. But actually, I meant to just ask you about what the impact of 5 % interest rates were specifically versus zero. And then I asked a very vague question. But I am curious about the sort of the strictly macro element. We're also in a weird moment because rates have gone up, but the NASDAQ is at all-time highs. And it seems intuitive that private company valuations have some tethered to public company valuations, either by dint of being acquired or by IPOs.
47:05Is there a difference, though, just from the sort of macro environment rate side that affects your thinking today versus 0 % rates in 2014? Well, in theory, it should be a lot harder. Yeah. Because there's so many more alternatives for where to put your capital these days that are appealing. There's macro big tech that's taking advantage of a lot of the trends that we're talking about, not just startups. And there's, you know, risk-free treasury bonds. Now, countervail that with all this excitement around AI and you kind of have the current moment where it should just feel like a sort of post-2000 era bubble.
47:42That's what it should feel like. If you take the financial flows, that's where we should be. And we're not. because there's just excitement about what technology can do, and the cycles are getting faster and faster. It's like people don't feel like it's 10 years away now. It's like almost here. Ethan Kurzweil, thank you so much. That was fantastic, and I really appreciate you doing this live oddlots with us. Thanks for having me. This was fun.
48:16And that was our conversation with Ethan Kurzweil, founder and managing partner of Chemistry VC. And a big thank you to everyone who came to this live recording. It was a very rainy evening in San Francisco, so appreciate so many people coming out. And a big thank you as well to our sponsor, Principal Asset Management, for making this possible. Joe, should I leave it there? Let's leave it there. All right. This has been another episode of the All Thoughts Podcast. I'm Tracy Alloway. You can follow me at Tracy Alloway. And I'm Joe Weisenthal. You can follow me at The Stalwart. Follow our guest, Ethan Kurzweil, at Ethan Kurz.
48:53Follow our producers, Carmen Rodriguez, at Carmen Erman, Dashiell Bennett, at Dashbot, and Kale Brooks, at Kale Brooks. Thank you to our producer, Moses Andam. For more OddLots content, go to bloomberg.com slash OddLots, where you have transcripts, a blog, and a daily newsletter. And you can chat about all of these topics 24-7 with fellow listeners in our Discord, discord.gg slash OddLots. And if you enjoy OddLots, if you like it when we do these live recordings, then please leave us a positive review on your favorite podcast platform. And remember, if you are a Bloomberg subscriber, in addition to getting the first heads up about these types of events, you can also listen to all of the AuthLots episodes absolutely ad-free.
49:36All you need to do is connect your Bloomberg account with Apple Podcasts. In order to do that, just find the Bloomberg channel on Apple Podcasts and follow the instructions there. Thanks for listening.
49:55Thank you.
50:23EasyCater, you get a single workplace food vendor with the tools and resources to make it easy, giving teams across your organization an easy way to order from a huge variety of restaurants, all on one platform, all while consolidating your corporate food spend so you can control costs, streamline billing and payment, and simplify reporting. EasyCater, your business tool for food. To learn more, visit easycater.com slash podcast. Residents at Brightview Senior Living communities enjoy enhanced possibilities, independence, and choice. Brightview Dulles Corner and Great Falls offer vibrant senior independent living, assisted living, and memory care services through various daily programs, chef-prepared meals, safety and security, transportation, resort-style amenities, and high-quality care.
51:11Take financial possession of your apartment by December 31, 2025 and save. Discover more at brightviewseniorliving.com. Equal housing opportunity.
From the publisher
In the 2010s, we saw an incredible boom in the venture capital space, fueled in part by cheap capital as well as cheap compute. Fast forward to today, and many things look very different. We're not in the ZIRP era anymore. And computing power has become a scarce resource, particularly when it comes to AI. So how do things look different today from the perspective of a veteran venture capitalist? In this episode, recorded live in San Francisco in November, we speak to Ethan Kurzweil, a founder and managing partner at the new VC firm Chemistry. Ethan spent years at Bessemer Venture Partners, where he was involved in numerous software deals. He talks to us about his strategy for the new fund, the case for starting a small firm, what technologies excite him most right now, and the general landscape for seed-stage investing.
Become a Bloomberg.com subscriber using our special intro offer at bloomberg.com/podcastoffer. You’ll get episodes of this podcast ad-free and exclusive access to our daily Odd Lots newsletter. Already a subscriber? Connect your account on the Bloomberg channel page in Apple Podcasts to listen ad-free.
See omnystudio.com/listener for privacy information.