Do Kwon Guilty, Powell Pivot & OpenAI’s $12B Run Rate | E2168

27:33So we felt okay, having a small percentage of our net worth, extremely low single digits, like, okay, sure, we'll try it and see what happens. But this one to me just was like blaring alarms. Similar to when Niccolo was telling me he was building a Badger, you know, pickup truck to go with it, that would be electric so that he could get retail investors to buy the stock. And that was his motivation as opposed to building a product or service in the world that people needed. So I'm a simple guy. You know, I'm fine putting money into a venture firm to QQQ to buy specific companies that provide a product or service to the world that I can understand and I can understand the customers.

28:15Here, I couldn't understand the customers. I couldn't understand who was regulating the game. and it was too complex of a system for me to hold a mental model of and be able to explain to other folks. And when you were struggling to explain it to me or said just outright, like, yeah, it's hard to explain this to a 12 year old and you add that it's unregulated and it's offshore, just felt to me like, why would I participate in this? It's just like me playing in a poker game where they tell me, yeah we're gonna have like do seven and the suicide queen is wild and you know everybody's gotta play a bomb pod every two orbits and you're like okay this feels like i'm getting angle shooted i just want to play no limit or plo right yeah yeah and everybody gets into a poker game like that where somebody's angle shooting some crazy you know um some some crazy angle against the other players and when that happens i just say you know i'm gonna sit out in orbit you guys have fun.

29:14I'm going to go call my wife and check in on the kids or I'm going to I'm going to go have a steak sandwich and call it a day while you guys play, you know, these crazy games. So I guess what's happened since is that he was arrested. I remember he got picked up in Monte Negro or like some Bond villain area getting on and off a plane in Europe. And then I guess the Koreans picked him up. The key to running a successful startup or maintaining a profitable portfolio? We all know the answer. It's data. If you don't understand your numbers and you're not watching them like a hawk, well, you're flying blind.

29:55But how can you be sure you're getting the best and most accurate information to make decisions and that you're looking in the right place? The answer is AlphaSense. This is the world's most trusted AI platform for market intelligence, currently in use by over 6 ,000 leading companies, including 88 % of the S &P 100. Alphasense's expert domain-specific AI has been trained on over 500 million business documents like earnings reports, equity research, expert interviews, filings, and more. The faster and more efficient your team gets at the research, the more founders you get to meet, the smarter and more informed bets you get to make.

30:35So it's time to accelerate your entire process with deep research from AlphaSense. So here's your CTA. Start your AI search and market intelligence journey today and get deeper insights to power your business. We're even gonna start you out with a free trial. Go to alpha-sense.com backslash twist to get started. He was arrested in Europe and then extradited to the United States. I think this was a multinational effort. And then the latest, just to kind of bring everyone full circle here, is that he pled guilty to one count of, and I'm quoting here, conspiring to commit commodities fraud, securities fraud, and wire fraud, and one count of, quote, committing wire fraud in connection with fraudulent schemes at Terraform.

31:17And it's a$19 million Jason forfeit from him and up to 25 years in prison. Wow. Of course, he probably won't get 25, he'll probably get three. Now the question then became, and I wasn't sure about this, how do you get in trouble for just having a bad idea, right? Okay, they had this idea for a stable coin that was algorithmically backed, it didn't go out well. That doesn't sound illegal, it sounds stupid. Well, it turns out there's a bit more to this just to let people know. They were doing things like saying they had a partnership with a thing called Chai that was running transactions on their blockchain.

31:48That wasn't true. So some of the real world use case was fake. They claimed that the Luna Foundation Guard, which was set up to help defend the Luna token was independent. It wasn't. And so there was actual false representations as well that were part of this. So it was a real crime. So what you get people on and what he pled guilty to isn't necessarily creating a crypto project. It seems like it was under the false claims and then maybe not following exactly securities law. Am I correct there? Was there a securities law portion of this or it's just unclear what he because it seems like he just settled and said, OK, I'm guilty.

32:28Well, the first count did include securities fraud as one of the things that he pled guilty to. So I think it's under that umbrella. One last thing, Jason, you mentioned at the top crypto regulation as why you feel confident in putting a small wager into a crypto fund. I was curious what the Genius Act, the stablecoin-focused legislation that passed recently, had to say about algorithmic stablecoins because their reputation went from high to low in the wake of this catastrophe. So I went back to the text for us today and all it calls for in the Genius Act is a quote study on nonpayment stable coins, a group that it says includes endogenously collateralized payment stable coins.

33:08What's that? It's a stable coin that quote relies solely on the value of another digital asset created or maintained by the same originator to maintain the fixed price. So they're going to investigate this idea, and it may in the future be allowed under the bill, but I think everything else is going to be one-to-one back so we can close the chapter, close the book on this chapter. And this word is endogenously? Endogenously collateralized payment stablecoins. Where the word endogenously, E-N-D-O-G-E-N-O-U-S-L-Y, an adverb meaning in an internal self-generated or self-originating manner and in business economics or science it refers to something that originates from within a system.

33:53So in finance an outcome is endogenously determined if it arises from the economic model not influenced by external forces. So this seems to be the key term of art here that is in the genius act to protect against this kind of shenanigans. If I may use another term of art. Yes. Shenanigans is as serious a term as endogenous. Well, here on Twisted is. Yes. It took me a while. It seems like there were shenanigans. And this is why the stable coin stuff has been brought on shore and tightened up is to get rid of these things. So I guess my thesis This maybe is playing out, which is Gensler told everybody play by the rules or get arrested, which is what happened here.

34:46And now Sachs and this current administration are saying, here's the rule set, the new rule set, follow it or get arrested. And to try to protect Americans, which is the job of our agencies and, you know, the executive branch and the House and everything to pass these laws. So it feels like we're going in the right direction by onshoring the stuff and that Americans need to understand if you're participating in things that are offshore, you're really taking risks. And if you're participating in things you can't understand and explain to a 12-year-old, there's something that I would consider a screaming red flag on fire.

35:32Like if you took the red flag and you waved it for people and then dipped it in gerozine and lit it on fire, that's kind of what you're seeing here. Boom. Like just you can sit this one out, folks. Like don't get greedy. Right now, it's incredible. You can get 5 % or 6 % from a Robinhood, from a Coinbase or from your bank on a cash holding because to the first story today, the interest rates have been high. And now those interest rates, if they come down, you're going to be getting paid 2 % or 3%. maybe people will be like, you know what? I'm not going to stay in cash. I'm going to have other, I'll have some more market participation.

36:07I'll move out of cash because I want to get the 7 % gain on average, not the 2 % or 3%, and people will start making those trades. So this is all important for people to understand if you're building a crypto company or you're raising money. Absolutely. Next up, Jason, let's talk about OpenAI. They recently had their first billion dollar month, putting them on a$12 billion run rate. I wanted to bring this up because we discussed on the show back in June when they announced they reached a 10 billion annual run rate. So for me, I think going from 10 to 12 in, it's a little hard to tell, one to two months is the timeline, is incredibly impressive.

36:42That's 20 % growth in less than a quarter, which is startup style growth, frankly, Jason, that we would love to see from a series A or B company. But in this case, from a company at a 10-figure run rate, which struck me as very impressive. Sorry, 11-figure run rate, which was very impressive to me. What's kind of crazy as well is that in an interview with CNBC, their CFO, Sarah Fryer, said still that, quote, the biggest thing we face is being constantly under compute, which means that they could grow faster if they had more GPUs, which at some point, they're going to have enough, right? I guess, yeah.

37:19And Sarah Fryer is the same. She was the CEO of Nextdoor, I remember. And before that, she was at Block, formerly Square. So she's got a pretty great track record. The CEO track record, Nextdoor did not do well during her tenure, but obviously Square did exceptional during her tenure. And the question that I would have is the profitability of this revenue. So what are they spending in CapEx? What are they spending to deliver this? And then what is the breakdown between hits to their API versus consumer? And I remember it was majority, 60, 70, 80 % at some point of this revenue. We could probably ask producer Claude if they can find us more information on this breakdown.

38:06They also have 700 million weekly active users, wows, I guess, instead of DAOs and mouths. So maybe they have 500 million DAOs. So if you have 500 million people using your product and you're making a billion dollars, they're making two dollars per user but i don't think they're making the two dollars off those users they're probably making if that billion was 50 50 they would be making a dollar off every user which would be in subscription revenue because they don't have advertising yet and their subscription is 10 or 20 bucks a month so you could actually back into how many paid subscribers they have do they release their consumer subscribers i don't think they release consumer subscribers, but I do recall they went from three to five million business customers over, again, a single digit month timeline earlier this year.

38:55I've asked producer Oliver to ask producer Claude for the stat that you pulled out there. But what I want to add though, that they did discuss, Sarah Fry in the same interview, that there was quote, an acceleration in plus and pro subscriptions. So that means that the consumer stuff that you're saying is doing well, is still doing well. But she also highlighted a nice momentum in the enterprise and quote, great momentum with developers. So in theory, this is hitting kind of all the major AI groups right now, the post-GPT5 open AI world. There seems to be a disconnect between GPT5's reviews on Twitter, X, and how they're playing in the market, the models.

39:30So maybe things are a little better than we thought. Well, I think that this could add, I think they could add 100 million in advertising per month out of the gate, a billion. I think they, well, no, I think they'll to 500 million a year so i think they had 40 million a month in year one for advertising we made a bet on when they would have advertising in the previous episode yeah i don't know if we have that documented yeah it's in it's it's in that no okay so we will get that yeah let's get that here because i think this is an important thing to look at is when they release their advertising product because right now the free users are not being monetized.

40:13They're used as a funnel to try to convert people to subscriptions, but you can use a number of free products. And I think 90 % of people are not going to pay for this five to 10 % well, which is probably what probably similar to YouTube. So YouTube has over a hundred million paid subs, I believe. And if YouTube has 2 billion or 3 billion users and 100 billion paid subs, you can get to that 5 % of people might convert to paid. So we could actually back into of the billion people who use the product a month or 800 billion who use it a month based on the 700 million who use it a week, 500 million using it a day.

40:52Let's say it's a billion people use it a month. If you converted 5%, you get to 50 million, 50 million times 10 bucks each, 500 million, which kind of would be half that number. So we can start to approximate and back into a 5 % subscription rate. So of two or three billion eventual users. Yes. There might be a hundred million paid. I think that's going to change a little bit though. They're recently announced a chat GPT go subscription in India. That's about five bucks a month. Once you do currency conversions, clearly targeted at a different price point for a different market. And just to go back to your point about our wager, Jason, we had a bet over under was at 12 months.

41:31for when OpenAI is going to bring advertising to ChatGPT. I took the under on that, and I believe you took the over on 12 months. Okay, so let's make sure we have that in there. And we're making a calendar. So there'll be thisweekinstartups.com slash calendar will be our Google calendar that you'll be able to look at and see all the great things that happened in the history of This Week in Startups. We're going to put our past episodes on there so you can see what happened on this day years ago, who came on the show, you know, the Travis episode, et cetera. So we'll put the notable anniversaries on there of original appearances by famous people like Do Kwan or infamous people like Do Kwan and famous people like Travis.

42:13First episode, all that good stuff. So producer Oliver is going to get on that with our Athena assistants. If you want an Athena assistant, go to Athena.com, Athena.wow.com to get a couple of weeks or months off since I'm an investor and partner with them. Okay. time uh next up anthropics upcoming funding round has doubled in size from 5 billion to 10 billion dollars jason yeah i at some point the numbers get so large that i kind of lose touch with them uh yes but the company was reported to be looking at 5 million at a 170 billion dollar evaluation now it's going to be 10 billion uh the company has seen its revenue scale from about a billion to about a 5 billion dollar run rate this year so not as big as open ai jason but 5x in a year half a year is very, very impressive.

42:57And I pulled some data on this. The number one model on LM Arena today in aggregate is actually Claude Opus 4.1, best in GPT-5 High and Gemini 2.5 Pro. Yeah. And so we're starting to get a revenue, a mental model of revenue for large language models. So you got 12 plus 5, 17. XAI is probably doing$2 billion. So now you're at 19. Maybe Gemini is doing a bit. So we're starting to look at maybe in 2025, there'll be$25 billion in revenue for this category or more. So then you times that by 100, and you get to a couple of trillion dollars. And I have put this market size at a$10 trillion prize. $10 trillion prize means eventually$1 trillion in revenue and times 10 for a fast-growing category, price-to-sales ratio.

43:56So we're going to have a mental model here sometime at the end of next year of what the space is worth. Then we're going to start in 2027 looking at the profitability of the space. and any hand-wringing about, oh, it's a bubble, will be worked out over the next 24 to 36 months as analysts start to say, okay, let's get to brass tacks. Who's going to pay for all this infrastructure? Where's the revenue going to come from? Okay, Meta wants to put$75 billion into infrastructure this year. How do we get that? How do we get double that number back? Or if that's$75 billion you spend this year, does that last three years, five years, seven years, as we talked about previously?

44:43So how are you going to amortize that? Or is it going to be$75 billion a year for 10 years and you're going to be putting$750 billion to work? How do you get a trillion dollars in revenue out of that? Well, if the ad network works better, if people spend more time because the For You pages on Instagram, Facebook, and threads and WhatsApp are that much better. OK, there's multiple ways to generate revenue now. We're going to have advertising, increased engagement, causing more advertising. So you can put that together. Then you have consumer subscriptions and then you're going to have API calls.

45:15I don't believe the space is overhyped in the mid to long term. It's probably underhyped in the long term, appropriately hyped in the midterm and overhyped in the short term, which is just how technology works, folks. We're sitting here over-hyping self-driving today, and we're under-hyping it long-term, and in medium-term, it's probably appropriately hyped. Did you see the MIT report that 95 % of enterprise AI pilots, yeah, this is kind of where I thought we were going to go with this, which is people are saying, oh my gosh, right now, complicated AI pilots in enterprise aren't generating much value, therefore, this is all going to go to zero, which I don't agree with whatsoever.

46:01I think that what that says is it's very hard to get enterprise AI right, but there's so many people working on it. Models are getting better, Jason. I'm not that concerned about it in the medium term. We had a long-term, yeah, we had a long-term, we had a big discussion about this at the top of the All In podcast, which you'll drop at the same time this episode drops, I think. And I think the conclusion is pretty clear. Enterprises are not the best people to execute on new technology. Startups are. And we're seeing, game on the field, our startups are getting massive value out of this. They know how to execute.

46:34And then you have to ask yourself, peeling back the onion, is a large language model, generative AI, just generically going to replace Salesforce or pick a product or service? Today, The answer is obviously no. However, will Notion's AI or Slack's AI provide extra value and get you to upgrade? The answer is yes. We just upgraded both of those products to the AI versions. And when I look at the Notion AI, it's really nice at recapping stuff for me. And the recaps in Slack, I think are underwhelming, but I'm using them. And so those applications, the narrower the application, the greater the output or the greater the potential, I think, today.

47:21And so that same study said like two out of the three verticalized software products were not being kicked out. So if you were to use an investment like ours in tax GPT, a company incubated, that one is getting tremendous value. And people will cancel that product when you drag it from their dead, dying hands. Because when you can make an account 10 % faster during tax season, they can actually sleep or they can do 10 % more tax returns and they can make less errors. So coding co-pilots, tax co-pilot, legal co-pilot, sales co-pilot, sales tools like Grilla Voice, which we're investors in that help you understand who's a better salesperson and why based upon their actions in a Zoom call or whatever it happens to be, or Granola, which we started using internally, which is a similar product just for keeping meeting notes.

48:20And Huddle now in the advanced version, Huddle's on Slack. We'll give you an AI summary. All of those things are doing wonderfully, brilliantly inside the organization. So I think the MIT study confirms that big companies doing wide projects are going to mostly fail at the early start of the technological curve, much like the first volley of apps from corporations sucked. And then United Airlines app is brilliant today. I don't know if you've used it recently, but it is awesome. It allows you to change seats or return a flight or book a flight like almost, I would say 80 % to the elegance of Uber or DoorDash.

49:09Not quite there, but close, close enough that it's highly functional. And you get really great alerts, like which baggage carousel your bags are on. This sounds stupid, but now I trust the app. The app is, the United app is essential to me. And I open it up and I look at what baggage claim are going to be. I don't go to the board at baggage claim anymore. I have that information while I'm walking from my gate to baggage claim. Stupid example, but it took these guys 15 years, bleep that out, to get there. That's how incompetent large corporations are. It's true. And as a United Flyer myself, I can attest to this.

49:52Just to give an example of what you're talking about, Jason, before I came over to Twist, I was at TechCrunch, which is owned by Yahoo, which is owned by Apollo. Anyways, big company, right? Thousands of employees, lots of offices. They installed Glean on our intranet. Glean is a company that this year raised$150 million at a$7.2 billion valuation. It's a big enterprise-facing AI product. And what they did was they turned your home screen of your browser into a Glean page, which didn't do anything. But I know that the idea was to bring together our corporate information to reduce IT desk calls but I didn't need that.

50:28And so what I got was an enormous rollout of a product that did not work for me in any manner of speaking. So I can see why these pilots are struggling, but I agree. I'm still incredibly bullish on a medium to long-term timeframe, even if there will be some chop in the near term, but that's always the case in a gold rush. Yeah. And the reason why small companies beat the big companies, aka enterprise companies, is because they're more nimble, they're more focused and it's not a side hustle. It's the hustle. Whereas all these AI pilots are done by an experimental group. The CEO has some funny money in a budget somewhere.

51:06The CFO says, yeah, give this person 10 million, give that person 5 million, give these three people 2 million each. Let's see what they come back with. And if we can actually impact our business, it's funny money to the organization. They're making billion dollars in revenue. They carved out 20 million to do some pilots, which they always do. And they're just there to learn. So it's great that MIT did this, but people are overreacting and over-indexing on it. So that's fine. All right, Jason, Canva is up next. They are executing a tender offer, a secondary share sale for their employees. And this matters for us because their valuation is going up from$32 billion set last year to$42 billion.

51:46And for people who were around in 2021, The valuation of Canva did reach$40 billion before going down as the ZERP era kind of came to a close. So they are now back above their previous valuation high. And I think this is mostly down to Figma's outstanding IPO we saw quite recently. Yeah, 100 % correct. Yeah, Figma's had an incredible run as a public company. It's obviously come down, I think, correct, from the public market valuation it peaked at? It has. Yeah, and we can pull up that chart in a second. And, you know, they've got over$3 billion. Canvas got over$3 billion in revenue and a quarter billion monthly active users.

52:24It is one of the great success stories of this era. And it is the most successful startup to ever come out of Australia, with the notable exception of Atlassian. And I think Atlassian's public market cap is bigger. but two extraordinary companies to come out of Australia. And the spirit of those founders and entrepreneurs in Australia is extraordinary. You can put them, you know, heads up against American companies, no problem. And they actually trounce the European companies in my mind, most typically. They're dogged and they're awesome. So, OK, so looking at the share price, peaked at 120, is now at 77.

53:10what's the market cap of Figma then, I guess, currently? And then we look at their revenue. So Figma's market cap is$38 billion. And we need to get their revenue. I did the math beforehand, Jason. They have currently a 39X price sales multiple. And so if you applied that multiple to Canva's$3.3 billion annualized revenue as of June, Canva would be valued at about$129 billion. So I'm not sure what to make of the discrepancy between the two companies' valuations. Perhaps public market investors are too enthusiastic and private market too conservative. But I... Yeah, you nailed it. And Figma will do a billion in revenue in 2025,$1.1 billion.

53:57So Canva does three times as much revenue as Figma. Wow. That's... Yeah. I didn't realize that they were three times bigger. This is an important lesson for founders, Alex. Figma is for professionals and Canvas for consumers and prosumers, but I would say not exactly like the professional product designers tool. That would be Adobe and Figma. So really bold move by the team at Figma, which has been able to, the team at Figma has been able to figure this out. Consumers, buying a product can get very big, but people like to go enterprise because enterprise buys early and often, easier to sell into, very easy to target.

54:44So you can play the long game like Figma has, I'm sorry, like Canva has and capture that market or you can capture the enterprise market, but they're both huge markets. design tools are a huge market. Who knew? So producer Claude lets us know that in Q1 of this year, Figma's revenue was$228 million, Jason. And then in the second quarter, just about$250 million on the Figma side. Billion dollar run rate. Billion dollar run rate. Thank you, producer Claude from our friends over at Anthropic. All right, next up, Neuro raises a bit more. So Jason, I think it was back in April, Neuro, the self-driving company that Uber is partnering with along with Lucid to create their own self-driving taxi, raised a Series E of about$100 million.

55:27They've added about another$100 million to that, bringing it to$200 million. And this is when the Uber money has come in. Uber pledged to make a nine-figure investment into Neuro. And so here we are. The company is now worth$6 billion. And in theory, Uber, Neuro, and Lucid, Jason, are going to get about 20 ,000 self-driving cars out in the world starting 2026, which means we'll see 10 ,000 in 2028, which means that in the medium term, to your earlier point, it's a good time to be bullish about self-driving. Yeah, and really the story here is NVIDIA's involvement. NVIDIA has a self-driving platform.

56:05It used to be called NVIDIA Drive, I believe, but I think it's called Thor or Drive Thor or Drive AGX. They have a couple of different platforms here. And what they do is they provide the hardware suite training data, and it's kind of modular. So if you want it to have self-driving level two, level three, eventually level four, you can use some portion of NVIDIA's hardware stack to keep the car in the lane and to do adaptive cruise control, which would be like those two things together is considered level two. autopilot on uh tesla's is considered level two stays in the lane you still got to be active obviously and it keeps your your distance you know and with that adaptive cruise control fsd level four is what waymo has and what tesla is trying to achieve and will have achieved when they take the safety drivers out and when you can sleep in your car what's going to happen here is NVIDIA is building a stack that any automaker can put on their cars and are starting to do.

57:18The Chinese automakers are using this stack. So you'll have a company like Nuro, which will do the software for full self-driving, but they'll use the hardware stack. And what that means is they're going to run the price down and the power up and the ability to buy a car without the hardware for self-driving will become impossible in about five years. Just like buying a car without seatbelts, air conditioning, electric windows will become impossible, right? It's like hard to buy those things unless you're buying like a retro throwback car, like that slate platform. Dot auto. But just to clarify here, Jason, there's two ways that things become baked into cars over time.

57:59One is consumer preference, air conditioning. I'm not going to get into a car that one. And another is rear view cameras, which are mandated by law. You think this is going to be consumer preference versus a federal mandate? Consumer preference for in the short term, and it will be law in three to five years, I'd say. You're not going to be able to buy a car that doesn't have lane control, that doesn't have auto braking when a ball goes in the street, which I don't think is mandatory right now, but the rear camera is mandatory. So the question is, when will the auto industry accept$1 ,000 in additional cost per car to have these advanced safety features?

58:39And I'm guessing you can start to have these advanced safety features in that$1 ,000 range right now, maybe$2 ,000. It's not a full self-driving compliment, but yeah. And that's that tension between the automotive companies trying to be profitable and lower the cost of cars and then the public good of, you know, giving you lane departure data. Like, why isn't that standard? Feels to me like lane departure warnings should be absolutely standard given the fact that young people and older people, too, are on their phones constantly, you know, switching podcasts, checking their text messages, and all this other shenanigans, this stuff will all be a requirement in five years, I'd say, in the Western world.

59:29And yeah, we should look into that over time. But Waymo is also, by the way, going to be selling these packages into Toyota. So if you do a quick search for Waymo Toyota partnership, this one went way under the radar, but Waymo will become an OEM, original equipment manufacturer to the number one or two car maker in the world, which I believe is Toyota and Volkswagen second. These are the car manufacturers who make five, 10 million cars a year. And there it is, yeah. Yeah, this is gonna be enormous. By the way, Jason, we did pull from producer Claude that the automatic braking system that you were discussing is not currently mandatory, but it will be required in all new cars in light trucks sold by September of 2029.

1:00:16Correct. And if you want to see a really interesting tweet, let me give you this one to pull up. Bill Gurley, who was the Series A investor, I did the seed round, Shervin did the B. For Uber, here's a tweet from him in 2020 where Bill Gurley says, the people that predicted full self-driving by 2020 are ones that don't truly understand the complexity of the problem. This is still many, many years away for technology and who knows how long for regulation, but the broader press keeps biting on the bait. Great article. And this is from May 14th, 2020. And, you know, this bill had the inside information on this.

1:01:02We're sitting here in 2025 and the only person actually running a robo-taxi service is Waymo. And then the ones in China, China, obviously regulators, a little more frisky. But the next shoe to drop is going to be public reaction to these technologies, excuse me, through the lens of job loss. If you type in Wuhan self-driving car protests, what you will find is in Wuhan of, you know, Wuhan COVID laboratory fame. I don't know why this city keeps coming up is... Well, it's big. I mean, I think the reason... It's big, yeah, but it's not Shanghai. I don't know. I thought Shanghai or Beijing would be the places we'd be reading about self-driving cars, but here it is.

1:01:46Taxi drives in Wuhan protest against robo-taxi expansion amid fears for jobs. I don't know what the date of this is, but... This was from April 3rd this year and was updated later that day. Jason, the reason why I think you might see more protests in Wuhan versus Shanghai or Beijing is that they're not Shanghai or Beijing. And so there's probably less oversight at the CCP level because they're less politically sensitive. Like if you tried to do a little protest in Tiananmen Square right now, how's that going to go over in Wuhan? You're a little bit out of the radar, if you will. Maybe, yeah. Maybe that's why they did it, yeah.

1:02:18So it's really interesting how this is shaping up. I'm going to stick with my prediction, which has been within 36 months, a dozen players will achieve full self-driving autonomy for under$50 ,000 per car. So that means maybe by 2029, there'll be 12 major players with$50 ,000 cars. And then it's just regulation and protests that determine how quickly these things get on the road. And so if you were placing a bet, and I have placed these bets, I think you could look at Volkswagen, Toyota, Waymo, Tesla, obviously, and Uber, WeRide, Pony. All of these companies are going to get there within the same window, which then means it's deployment and regulations.

1:03:20And we'll start talking about things like the depots. And I sent you a video of some folks who found the depot for Waymo's. Um, this is where the rubber is going to meet the road, keeping these cars clean, charged, uh, safe, uh, the tires, you know, inflated, all this kind of, uh, cleaning the puke out of the backseat. All of these issues are going to be 20, 30 % of the story in each of these cities. It's going to take an army of people to staff. the companies. Here's a shot of this. Just from the video, Jason, you showed people what you're talking about. This is the place where I believe they get all charged up and made pretty.

1:04:09It looks very much like a car dealership, shockingly enough. A supercharger station, yeah. So, anyways, that's what's going on there. I do want to point out, though, that the protest point is material. I want to show you a picture, Jason. So, there's a discussion in Boston about whether to allow Waybo and self-driving cars in. You won't believe who is protesting. Check this out. This is from boston.com, and it shows a picture of people wearing shirts that say, App Drivers Union. Wrong way, Mo. And so what we're seeing here is after people protested against Uber and it's changed to the taxi industry, now drivers are protesting to protect their jobs at Uber and against the new thing.

1:04:50So this, I think, lets younger listeners tell that. DoorDash, yeah. This is a great find. I think this will become a societal issue. Every self-driving car is six gig economy workers and four full-time workers. So just keep that in mind, folks, when you hear about plans for 20 ,000 self-driving cars from Uber or Volkswagen having tens of thousand or Waymo hitting tens of thousands or Tesla potentially hitting 100 ,000 because they can make cars quickly. And EVs aren't selling. So if EVs have kind of hit a ceiling amongst consumers, then all that capacity can go towards robo-taxes, which is what Elon has said he's focused on.

1:05:32Yeah. The market will expand, though. We will go from 1 % of rides being

1:05:41ride-sharing to 20 % or 30%. And the pie will grow because people will take more rides. If rides were half the cost they are today, which is, I think will be the end state. You know, your$40 ride from the airport will go down to 20 to your house. The$20 ride to a$25 ride to dinner will go down to 12. So instead of 50 bucks round trip from dinner, it'll go to 25. It'll be significant. Who can afford that and who chooses to do that? Because maybe it's better to just be in the back of your car. When I drive to my office, you know, in that drive, I like it. I like spending 20, 30, 40 minutes in the car, but when it gets to 40, maybe less so.

1:06:29And if I could be on my phone and have a full self-driving car that I own or be in one, And then what's going to happen is these drivers are going to move out to the burbs. So all the current Uber drivers and Lyft drivers and door dashers will move to the suburbs and the country to get rides because the robo taxis will perfect downtown Austin and San Francisco and some parts of the peninsula. But they're not going to get to Napa for a while. They're not going to drive to Tahoe. So that will be a phase of this. the humans will do the long haul drives or they'll relocate themselves outside the city centers.

1:07:11So brave new world, folks, get ready for it. We have a founder question. I love a founder question on a Friday. All right. So from our friends over on Reddit, we are asking about how to handle equity splits. Which subreddit? I guess we've got to figure out which subreddit it is so people know. This is from JIVI31 on the startups subreddit. Okay. And so after I answer it, producer Oliver, Please go clip it and upload the clip there and just say, hey, I'm a producer. Jason answered this. And let's see if they let us do that. I don't want to do it for marketing or promotion, but I do want to give them the answer.

1:07:44So here we go. So they ask, hi, everyone. I'm starting a startup with a technical co-founder. We didn't talk anything about how we'd split the company. Warning, bells are already going off. He came up with the idea. I worked on ground and developed it. I thought 50-50, but now he isn't ready to give me 50-50. I've been handling business, marketing, while he handles the tech and development. I'm not a tech guy, but we both need each other to win. I was working as a freelancer on a contract basis, but now I work in this full-time and he has a full-time job as well. What should I do now? Jason, I'm sure you have some thoughts about this.

1:08:20Yeah, life's a negotiation. It's good to have this negotiation so that if I was starting a company with a technical co-founder and I've got the gravitas I have at my age and my stature in the industry, I might say, I'll give you 20%, I'll take 80. Or if it was three people, I might say, I'll take 50, you guys take 25 each. Early in my career, if it was just me and Brian Alvey, we split it. Okay, yeah, simple, easy. If the person doesn't want to split it, you have to ask them, okay, what are you going to do that you deserve two thirds and I get one third? Can you explain that to me? And then let them make their best case.

1:08:59And then you can make your best case. And this is a great thing to do because it sees if you can take a complex, dynamic negotiation and get through it. If you can't get through it and you don't accept the value they put on your contribution and you feel it's insulting or lower and you make your best case and the person says, take it or leave it, you can take one or two, you can take a multitude of things from their position. One thing might be, this is a hardcore CEO who knows how to negotiate, and they're right. They have a better track record, more connections, and they are contributing more.

1:09:40Therefore, great, they're not a pushover. And so my one third to their two thirds is going to become worth more because when they negotiate with the venture capitalists or for our acquisition, they're going to fight the same way they just fought with me. Great. I would like to, if Zuckerberg's a lunatic or, you know, somehow you were able to be a co-founder of a company that Travis was doing or Elon was doing, you'd be like, okay, yeah, I deserve 5%. You deserve 95%. So here we go. And, but if they aren't contributing that much and they're delusional and they want to put you under their thumb for whatever reason, or yeah, you just think it's not equitable, you say, great, I have other opportunities.

1:10:20So I'm going to pursue those other ones. And you, as the person here, who's the business person who, you know, doesn't, who believes they should get 50-50, if you're actually that baller, go find another co-founding team that wants to split equally. And then you can tell the person, you know, I appreciate that you think you're worth twice as much as I am, but this other group of founders doesn't. And I'm going to take that other opportunity, unless you feel we can come to a better term. And I like that. I like that tension because there will be other decisions you have to make, like your salaries, the CEO salary or the CTO.

1:10:58The CTO might have 500K offers to work at Google doing AI. And the CEO might say, I'll take two thirds of the equity. You take one third. I'll take 100K draw. You take a 300K salary. And so some people sometimes are salary-driven, some people are equity-driven. So there's another nuance to it. How much cash do you want to take out of the company? And then you should also look at future equity, future cash compensation. After we raise our Series A and we get$10 million, okay, I want to pop up to$300 as the CEO, and I'm okay with you popping up to$500 as the CTO. I mean, those are big salaries for a startup, but let's just say you get some crazy$20 million infusion of cash in your Series A, you could justify it, I guess.

1:11:42There you go. So I would pre-negotiate the salary bumps in the future years if you raise money or you hit profitability. Last thing I'll say, if you have to write down, we didn't talk about anything about how we'd split the company, you made a mistake. Please do not wait to have these conversations until the last possible moment. Talk early, talk often. It's very important. The other problem here is if they've both been working on this and it's unclear, then the person who set up the company owns 100 % of the equity. So if you didn't negotiate this, now that person can just fire you. And if there's no documentation and they set up the company and they've been paying you, they could just fire you at any time.

1:12:21So that's why before you start working on stuff, you got to be clear about this. Before you write one line of code, you have to say, what's our intent here? Because the person who, again, founded the company, named the company, you know, they by default own everything. You own zero until you write, you know, those things. So yeah, they may not, they also may not be formed. And if it's a project, then who, when you organize the project, you're going to have to get an IP assignment from the other person. So that further complicates this. This is where an angel investor, a board member, an advisor who has been through it before can help you work through these issues.

1:13:02So if I was an angel investor in this company, I'd sit them down and I'd say, here's my perception. Here's what I think is fair. And that's where you can get a third party to maybe, you know, explain it. And then also be the grownup in the room and say, okay, even if you're not going to use this person, you still need to give them 5 % of the equity that you can buy back at this valuation or some cash compensation to get them to sign an IP assignment. If not, the company's not fundable because now you've got this overhang of somebody who wrote code that wrote the code base and doesn't have an agreement which do you want to deal with that yeah do you want to deal with that 15 years down the road or whatever i mean no you don't yeah remember the snapchat uh snapchat i'm gonna ask claude third co-founder lawsuit resolution producer claude will get this information to me um Um, so they did settle, uh, the lawsuit involving Snapchat's third co-founder Reggie Brown was resolved through a high profile settlement.

1:14:03Brown, who claimed to have originated the concept for Snapchat and was pushed out early, filed a lawsuit against Evan Spiegel and Bobby Murphy in 2013, alleging he was unfairly excluded from the company's profits. The legal dispute concluded in 2014, the confidential settlement. The details remain private until Snap's IPO filing. Here we go. brown received a total of 157 million in cash from snapchat the payments were divided between 50 million paid in 2014 and 107 million paid in 2016 the current valuation of snapchat is what and what is evan spiegel's stake 12.18 billion dollars as of today snap shares are up two percent to seven dollars and 20 cents i don't know evan spiegel's uh ownership offhand jason but i will We'll get that for you right now.

1:14:51Yeah, I'm happy to get it myself. Evan Spiegel's ownership percentage is approximately 30 to 33 million shares. That's the float. And Snap has 1.28 billion. So he owns roughly 2.6 % of outstanding shares. Not much. That's a very small percentage. That does happen. Um, but the SEC says he owns, he remains a 10 % owner. Oh, because he has voting shares. They have super voting shares. So Evan and Bobby have 96 % of Snap's voting power. Didn't Snap go public with shares that got no votes famously? Yeah, this is why people are having a hard, this is why the stock is depressed, I think, is because of this ownership voting issue.

1:15:41But if they own 6 % of$12 billion, they own collectively, yeah,$600,$750 million. I don't know if this is correct or not, but this is a closely, yeah, this is a tough one for us to determine here on the fly. Yeah, there's a lot of voting rights that come into this. But Snap is a company that has famously had very, very highly skewed voting percentages towards its founders, just as Meta is the same way. Meta is effectively controlled by Zuck still, I believe. So founder friendliness, it still exists in 2025. There it is. All right, everybody. We'll see you on Monday, thisweekinstartups.com slash docket if you want to get involved in the docket.

1:16:22We'll have slash calendar and slash bets up over the weekend, hopefully. We'll get those landing pages going for those. And he is cautiously optimistic. Cautiousoptimism.news. There you go. I got the URL. Cautiousoptimism.news if you want to hear Alex's take on his optimism. optimism and uh he's x.com alex i'm x.com jason we'll see you next time bye bye bye everybody