IPO Shakeups, NVIDIA’s Big Bets & Must-Know Founder Hacks | E2085

25:08We could do this for milk as well. You could have milk. You could have a camera inside your refrigerator that has where the staples go. You keep them in there. If it doesn't exist, you order it. Now that's complicated, I know, but it is the eventuality that these things will be abstracted away from you. There was another product that Amazon made. It was a little IoT device and it had the logo of Tide on it. So if you were a Tide person. I forgot what they were called, but you would put the Tide logo on your washing machine and you would press the button when you ran out of Tide and would automatically put it in.

25:43Now, AI is going to do this for you. Google's AI is extraordinary. Amazon is extraordinary. You can put things on subscribe and save. That's another interesting product. And you could also order from these things. The idea I've always wanted to do, Alex, I had an idea for something called Mercury Club. I never did it. But if somebody wants to do Mercury Club as a concept, I'm down. It was kind of like GoPro. It was kind of like Cosmo, Urban Fetch during the day, which was I was thinking about, well, what are the most frequently used staples? What if we put them into a truck? What if you park those trucks and the trucks can move?

26:22Yes. So you can have a truck with a dude in it or a gal or a person. There we go. i don't know what i'm supposed to do now there was a there was an executive order what am i supposed to do guys i think guys is still sufficiently gender neutral yeah i think guys is pretty broad okay okay gals guys everybody in between don't cancel me but what if this truck existed it was optimized and it just you open up an app and i you know i gotta tell dara about this idea imagine you open the uber app and the convenience stores were not convenience stores they were convenience trucks. They had a refrigerator section, got everything else.

27:01And instead of being 45 minutes away from you, because you had that option, right? You could have gone on Uber Eats and ordered it in your town or DoorDash. Oh yeah, yeah, I could have. But I have, Uber Eats has gotten to the point when it's so much more expensive than me schlepping myself that I can afford it. But I get angry about - The fees. Yeah, it's a bit much better. So this, I think, would lower fees because the truck would be out there. Now, I don't know if this is feasible. Somebody should definitely do it. if somebody wanted to make this as a business, I would be interested in a crazy logistics and tech person duo who want to do this.

27:32I put you in the incubator, give you 125k, let's model it. Okay. But also, do you recall Spoon Rocket from 2013? This was a great idea. And we invested in a company called Bento. Somebody who had worked for me did Bento version of it, explained this concept. And I think Chamath was in Spoon Rocket. This was so brilliant before GoBreats. So you're at the office and you're hungry and let's say it's 4 p.m. and you're going to be working a little bit late. What you don't want to do is pull up Uber Eats, pick a restaurant, go through the whole song and dance and then have it arrive between, you know, 45 or an hour and a half later because traffic if you're at the office is bad.

28:09SpoonRocket had a set number of meals they made each day and then they put them into cars and then they had the cars out and about. So if you wanted food now, you could pull up SpoonRocket app and you could see the three or four options. You could click, bring me that one. And then they would just pull the car around and there it was. So at the TechCrunch office back in the day, we used this all the time. The economics didn't work out, but it was great for me. The economics were so close. Bento also did this. And I remember the founder was so great. And he figured out, you know, at about nine o 'clock, there would be some left.

28:43And what they did was at 9 p.m., they told everybody in the app, it's two for one at 9 p.m. And in the app, you would have these cars, they would have the salads in a cool bag or in a cooler. Then they figured out that there were warm bags. So they would make the food. They figured out how to make the food in a proper kitchen. You know, like in San Francisco, there's a lot of, we're talking about regulations. A lot of regulations in San Francisco. Worst place you can launch any business. I mean, literally, I think the most regulation ever is San Francisco. But they figured it out. They tested everything.

29:16And then what they did was for the hot meals, you know, because hot meals, I think, could last two hours on a hot tray at a certain temperature was the law. So, you know, the car leaves in the trunk. They would have 20 hot meals in a hot bag. Then they wanted to make sure it was temperature close. They put a temperature IoT device in it and they put a heating pack in it. So there were heating packs that you could connect to a mobile battery and you could keep the thing warm. And then they eventually learned how to plug it into the cigarette lighter, yada, yada, yada. Anyway, this idea was really good, but it turns out getting one of four meals in 10 minutes, not as good as getting one of 400 restaurants in 40 minutes.

30:00Founders, let's be real. Let's keep it a buck. Finding great developers is hard. It's like one of the hardest things you have to do in our industry, especially when you're trying to run and scale your startup but here's the good news i got a tip that's going to save you time money and a ton of headaches you need to check out lemon.io lemon.io has thousands of on-demand developers who can help you they have done all the work to find and vet great developers who are experienced who are results oriented and who charge competitive rate. Great developers we know are hard to find and integrate into your team.

30:39So Lemon.io will handle all of that for you. They're only going to offer you handpicked developers with at least three years of experience. And if something goes wrong, Lemon.io will find you a replacement developer ASAP. A bunch of launch founders have worked with Lemon.io and they have told me they've had a great experience. So here's your call to action. Go to Lemon.io slash twist and find your perfect developer or an entire tech team in 48 hours or less. And Twist listeners get 15 % off the first four weeks. So stop burning money. Hire developers smarter. Visit lemon.io slash twist. So people would rather, and that was the discovery they learned.

31:21There was also a pizza company. The pizza company had a mobile truck. It was called Zuma. And Zuma pizza was the robotic pizza. I think SoftBank dumped a ton of money on them. The thing crashed and burned. And I looked at investing and it was only 60 % effective. But all of these ideas are available to you as an entrepreneur, just by looking at your own life and saying, there's got to be a better way. And then of course you got to figure out economics, et cetera. And there is always new technology that can give you a second chance at making it work. So anyway, people who weren't around for the dash button thing, this is what they looked like.

31:56There was buttons. You could put them on. This is 2019. Okay. So not too long. And well, I mean, that's now five, six years ago, Jason. I hate to break it to you, but time is moving quick. Yeah, time's moving quick. And if you were curious, this is what the Zuma pizza trucks looked like. I think, you know, SoftBank got a lot of stick for the amount of money they put into this company. But I mean, if you think about fast delivery, people love that. And people love pizza, literally the whole world around. I can kind of see the effort there. That pizza truck, by the way, was a Domino's on wheels.

32:32The problem is there's a Domino's every mile or two in America. So they were up against Domino's. If there's a Domino's every, you know, literally every five miles in, you know, any American suburb, which there seemed to be. Yes. And they're in like the dankest office space, you know, commercial space. They didn't have as big of an advantage as they needed to have. No, no. because the overhead for that dank office space is pretty low and maintaining cars is not that cheap, especially if they're commercial. I want to get back to the idea of finding ideas, Jason, in our founder hacks section. But quickly, before we get to that, right before we jumped on the show, there was some breaking news.

33:12I haven't had a chance to get into this much, but the news is that XAI is working on a potential$10 billion funding round that would value the company at about$75 billion. Now, details to come, we'll find out more. We'll figure out if it's Sequoia or Andreessen or Valor that will lead this round. But I think it goes to show that after all the deep sea hype, hysteria, enthusiasm, excitement, people that are building these systems are still looking to raise a lot of capital to put a lot of capital to work on building the infra and the investment and the personnel they need to kind of stay on top of the market.

33:46So I view this as more of an endorsement of the AI market writ large than perhaps any individual company. But it is, of course, a big deal for XAI and its ambitions to take on Anthropic, OpenAI, and other domestic and foreign AI companies. There's a big prize at the end of the AI rainbow. And if you were going to place a bet on who might do well here, Elon's not the person to bet against. He's kind of the person you want to bet with. He gets it. He was early. He funded OpenAI. And he's got a unique data set inside of Twitter that's real-time and powerful in the same way Reddit is, in the same way.

34:27And I saw Reddit actually took a dive. I think they had some search engine problems. I'm not sure if it recovered. And that lawsuit we talked about the other day, those lawsuits are going to keep happening. Content's going to do great. It could mean that these models are going to need more capital, not just for service, but also for content licensing. And I think there's a big prize right now. Search. I've moved half my searches to ChatGPT, Gemini, et cetera. And doing research, you know, here on the show, other things, I've moved 100 % of mine there. I used to ask people to do research for me.

35:04Now I just do it myself and I use deep research, et cetera. 20 bucks, 100 bucks a month, personal or corporate is a no-brainer. Every single person is going to spend between 200 and$1 ,000 a year on their AI language model. That's every human being. Yes. Do you know how I benchmark it? What do I currently pay for my cell phone service? Sure. That's kind of my benchmark for what I think the max for consumers will be. because I am still blown away with the fact that DeepSeek shot to the top of the app stores. And I know we have caveated that with you can game it and blah, blah, blah. But I think it goes to show that what consumers want is the fastest, cheapest, best AI model, period.

35:50And I bet you that in time that will cost something around 200 to 500 bucks a year. Maybe it's a family plan, but it's still going to be a lot of compute, but people just love it. Last note on AI models, Jason, before founder hacks, take a look at this chart. And I was impressed. So this is from the fine folks over at Hugging Face. We've talked about them, talked to them. They host a lot of the AI models out there. Great service. Love it. They have a chart here that shows how DeepSeek's R1 model became, quote, the most liked model ever on Hugging Face just a few weeks after release. It's been downloaded over 10 million times across its different variants.

36:26And if you're watching the YouTube version of this, you can see that the curve. What's the number two one? The blue one that catches up to it. The yellow one goes straight up. I see that. Black Forest Labs Flux 1-Dev. Okay. So that was a image-making model that I think Grok used, if memory serves. Likes and downloads, you know, on things like GitHub, people joining programs, is an indicator of their future success or at least interest in it today. What we're learning, I think, from DeepSeek is open source, I believe is going to win the day. Respectfully to everybody who's doing closed source, I think open source models are going to have such a distinct advantage that we could see a handful of winners, handful being less than five, in the closed source, and they'll have some advantages as well.

37:22And then we're going to see 200 open source projects being contributed to every day in every single niche possible. and it's going to be wild. It's going to be wild how effective these are. And open source is going to win, I believe. 60, 40, yeah. That tracks me. I do think that OpenAI, Anthropic, et cetera, will have value down the road. But I think this is why when we see 157 or$300 billion attached to OpenAI, a little scared. But there's been an interesting shift out there when it comes to startups and AI, because if you go back a year, people were talking about, you know, GPT rappers and that's just a skin around a model and that's not going to have long-term viability, long-term strength.

38:07But now it seems that people are talking about how all the value in the AI game is going to accrue to the application layer, which means building on top of models. So it seems that the thinking has gone an entire 180 here. Not for me. I said from the beginning, applications and brand are defensible. If you, we just talked about Airbnb, Toro, Uber, you know, these verbs, When you have a brand and a network effect, you know, and you're on people's phones, that is defensible. That does, whether it's Instagram with network effects and already being installed and people having their history in it, it's very hard to displace Instagram.

38:44Very hard to displace DoorDash, Uber, and Airbnb in people's lives. I've always felt that AI is going to give every single category of consumer behavior. all those incumbents are going to have to adopt AI and when they do it gives startups the chance to get in there and maybe displace them compete with them or replace them any of those are possibilities when a new technology comes out you used to have there was an Uber before Uber there was a taxi magic app where you could text and get a cab to show up it never worked there was no GPS it was over text it sucked but it was the same concept taxi magic when gps and the gorgeous iphone came out you actually had it work so this is from 2008 observe this iphone 3 3g style uh thing look at that yeah looking successful use refresh button you had to manually refresh but this is in the age of like Ajax happening, isn't it?

39:56Yes. Pick up location, live updates, taxi dispatch, the driver's name, taxi 0.13 model. It was like literally giving you line item updates on where your car was. Now, GPS didn't exist, I don't think, at that time. So you couldn't put it on a map yet. GPS came somewhere or Google Maps and Apple Maps came on the second or third iteration of the iPhone, maybe iPhone 3 or 4. In other words, just because something's been tried and failed multiple times does not mean it's going to succeed. As we just mentioned with SpoonRocket, Mento, and those other... Zoom, et cetera. Yeah, on-demand. Skype was supposed to do what we're doing right here, and now Zoom figured it out.

40:42So an IRC, people use before Slack. So there's a long list of inspirational gravestones for you to look up. You get to that. This is a founder tip, our strategy. You go to the graveyard, you dig up a body, you reanimate it. That is the best segue possible to the founder hack that I'm most excited to share on the show. Okay. So one thing we're going to be doing over the next while is pulling out great tips for founders. Things we find from people we know, we do find from awesome posts from venture capitalists. Pick your poison. We have our ears and eyes open. Jason, what I found today that I wanted to bring to the show is actually from the startups community over on Reddit.

41:21And it's from a thread entitled, we built the wrong startup, broke a startup golden rule and pivoted into success. Essentially, this founder ended up building something for which they were not the ICP or ideal customer profile. Now, people often say build for what you know. So they broke that rule. I thought the post was very interesting. But here's the bit that really stood out to me that I want to pass along to everyone listening from this post. and we'll have a link to it in the show notes and on the YouTube. Quote, don't look for what's cool now. Look for what you would have wanted two years ago in your career.

41:50This will help you find better, less competitive opportunities that will probably be less sexy. But I feel like this is what you're saying. Find the pain point. But in this case, just look back a little bit because there might be fewer eyes looking at it, but probably still a pertinent place to deploy software. I participate once in a while in the Startups Community Physique. I think I'm Jason M. Calacanis on Reddit. And I load Reddit a couple times a week. And so here's the thread in the startups community. Okay. So to summarize, they pivoted, but they didn't work on the ideal customer profile being them.

42:23They found another customer, maybe highlight the section here and let's read it. So essentially they had quote, to ignore the dogma of needing to be your own ICB, because right now we don't have the complex pricing models, global compliance headaches, or enterprise building workflows that our ideal customers do. So they couldn't dog food their product as much, but the idea of building for yourself doesn't always, always apply. What was the gist here? Got it. So there is great startup advice. Build something that you need, dog food it yourself, force yourself to use it. So if you were DoorDash, you'd say, I'm only going to eat what DoorDash can deliver.

42:58And you're like, okay, well, we've got a Tex-Mex place and we've got a Greek place. I really feel like sushi or ramen. Well, now the team's got to go out and find a ramen place to add to DoorDash. Okay, fine. Fair enough. But if you are trying to build something for CFOs who have compliance issues on a global basis, well, you're going to need to go find those folks, interview them, do customer interviews, listen to them, maybe even bear hug them. So this is a great point they're making. And a lot of times, not having experience is an advantage because you can come to a problem and say, I don't work in the hotel business, so I don't feel the need to provide a concierge and a front desk.

43:47And I don't think every room has to have a Bible and needs to be cleaned every day. Now, if you're in the hotel business, you're like, oh my God, how many times do we have people calling for room service and the front desk and the concierge? You have to have. Those are must haves. in the eyes of a hotel executive. Those are must-haves. You know, when Joe Jebbia and, you know, Brian are doing Airbnb, they're like, I need a place to stay that's cheap, that I can afford, and that's in this location where hotels don't exist. So they could build Airbnb and anybody who was in the hotel business would never have done it because they'd say, okay, it's gotta be 500 a night.

44:30They need to have a 24-hour concierge. We need to have room service available, stock the fridge, whatever it is. Fresh eyes can also help. You do not need to dog food your product. You just need to have empathy for a customer base. You need to be in contact with a customer base. And I'll make this even more specific. The way to get this done is to take five or six of your customers and build a customer advisory council. A customer advisory council. And you put them in WhatsApp. And you put three or four of your co-founders in there. And you say, hey, we made a new feature. What do you think? And they debate it.

45:05Now, they might send you on a custom software boondoggle where you're building something for one customer. But if you have five or 10 of them, they might say, this was good for me. And six of them say it's great. Four of them say they don't need it. OK, move that up the priority list. 10 of 10. Let's say you had 10 customers. 10 of 10 want it. OK, you know, you got to build it. But most times, you know, the customers want you to take the top one, two, or three features and just make them better, faster, cheaper, increase the network effects. What are the features everybody uses every day on their iPhone?

45:43Well, they make phone calls. They surf the web and they use messages. when you are watching a keynote and it's you know apple's doing a keynote or google's doing a keynote what do you see them talk about all the time here's the new features of iMessage did for phones we have new airpods that do noise cancellation that automatically pick up the phone when you nod your head we do transcripts we do facetime we do filters in facetime and oh our new browser is faster the battery lasts longer because the browser doesn't do that our browser does X, Y, and Z. They start with the most frequent use cases and making them better with every iteration of the phone.

46:26That's another technique. Take the core feature, make it 10 % better. Why? Because 100 % of your customer base uses that core feature. On the case of Zoom, it would be the quality of the Zoom. Every time Zoom became easier to use and higher quality video, better audio, So that's actually why Skype never worked. Well, Jason, they didn't even have to have high quality. They just had to have video calls that worked. Right. Like that, the bar was on the ground in that case. And there was just, there was billions of dollars sitting there, but everyone thought that WebEx had it locked up. And you know, WebEx was complicated.

47:03It didn't work. The software to install was kludgy. There was just so many blockers. And when you did a video call, it didn't work as smoothly. So taking out friction, really understanding your, I like this as a first foray into, what do we call in this segment? Founder Hacks for now until you come up with a better name. We have a question from Gary with two R's. Jason, why wouldn't you want all customers suggesting and voting on directions instead of five to ten? Okay, great question. I'm talking about startups in year one or two. And you want to have a certain type of customer on your customer advisory council.

47:39you want the one who is spending money is demanding uses the product and you can't possibly take a hundred advice from a hundred people a thousand people it's too many and you won't build a relationship with them what you could do is you could have 10 people in one whatsapp chat getting to know each other getting to your team then you'll launch a second one say hey once you be on our customer advisory council you put them into a second one how do i know this well i started doing dinners because i wanted to understand our lp based a little bit better. And I wanted to understand the members of the syndicate.

48:11What did I do? I had a dinner in New York at Mark's off Madison, great Italian place. My friend, Mark Straussman over by Madison Park. I think a dozen people showed up. I put them into this group and signal and I talked to them regularly. Then I started Founder Fridays and you can go to founderfridays.tech, get a great .tech domain name, tell them who sent you. Founder Fridays is eight founders in Los Angeles. I think to Los Angeles was actually up to like 30. And so they have four tables of six or eight. And then we have like 30 cities, but they're all in groups that are smaller. So you can actually have a conversation.

48:47The proper size of a dinner conversation is six people at a round table. When you get to eight, you have two conversations of four or three conversations of two or three. So you just have to be thoughtful about how many people are in the conversation because it will lead to chaos if you don't, if you go above 10 in my expectation in a WhatsApp. I think once a WhatsApp group goes above 10, too chaotic. I have a personal question to follow up on this point about customer councils. And this is a really ticky tacky one, Jason, but do you give discounts to the customers that are on your customer advisory council?

49:23Do you incent them to do this? You don't have to. It turns out the people who actually really need the product enjoy having access to the founders. And the access and being able to have a say in the direction of the company is what is truly valuable to them. Getting your newsletter for free, actually, if you were to do it for cautiousoptimism.substack.com, it would actually not work because then you'd have freeloaders giving you advice. Taking advice from people who aren't paying for your product is the worst idea ever. Now, you're saying a discount, but even still, users are called users for a reason.

50:08They use you. Customers are sacred. They've given you their credit card. You're in a deep, meaningful relationship with you. You're making a transaction. And they, the beauty of subscriptions is, will cancel if you're not providing value. Yeah. So people over-index on advice from venture capitalists, users, social media. But they should be indexing on a smaller number of people who are customers and who use the product. Because there's really two types of customers you could have, engaged customers and customers. You want to find the most engaged customers and really index off of their feedback.

50:48The person who's bought their third or fourth Tesla, that's a very important customer for Elon to listen to and for the team over there to listen to. My feedback as somebody who has, you know, is on their fourth Tesla and is getting their fifth, super important for them. Why am I going to buy the Model Y, not the Cybertruck? I'd be going back and forth. Why would my wife want the Model X, not the Model S or the Model Y? That's a really interesting piece of information for people to have. That's why being able to segment customers into buckets. So loyal customers who you've had already, who have been on your newsletter for over a year and who open it up all the time, they're going to have one piece of advice for you.

51:31The people who are in month one might have a different piece of advice. So this is where segmenting cohorts. So you'll hear cohorts come up. Very important. The people who lease their Teslas and who have the Model 3 and buy the most basic model and use it for driving Uber or their commuters or the college kids, that's one piece of advice. The people who buy the Model X and spend 100 dimes on it, that's a super user. And that is critically important. And you know what? This is a hit. This segment's a hit. I'm hearing people say this is super useful and commenting on it. Why don't we grab another question from the audience?

52:12How big of a role do you think humanoid robots will play in the next 10 years? Do you really think it's worth all the hype? In the next five years, I don't think you're going to own one if you're listening to my voice right now in all likelihood. I think five of us out of 100 will own one. I will be one. And that would just be like the same person who buys like the Newton personal digital assistant. It's going to be that sort of level or people who bought PCs in the 80s. In other words, there's a bell curve, the technology adoption bell curve. And there's the masses, they're the middle of the bell curve, and then you have the laggers, they're the last to use a technology.

52:53And then there are the people who like to sample new technology, the avant-garde. And people call them different things, but you see the innovators, early adopters, early majority, late majority, and laggers. And this is the adoption curve. And if you haven't seen this before, it's super important for you to know where you are. If we were talking about electric vehicles right now, we are in the early majority and we're going into the late majority. I was an innovator. I bought the Roadster. Early adopters bought the Model S. The early majority bought the Model 3 and Y. The late majority is going to buy, you know, whatever comes after that.

53:31And the laggards are the ones who get dragged along. If you had Uber on your phone in year one or two, you know, you were part of that, you know, Vanguard front. And the Vanguard, they'll buy stuff, not use it. They'll try it. Yeah. You know, they're odd people. They're coveted people, but it's me, you, Robert Scoble, and people who can't shut up about it, who buy the Vision Pro and wear it and annoy people and bring it to Thanksgiving. It's that guy. It's that guy at Thanksgiving. Are you spying on my marriage? Because that sounds a lot like me. I mean, if you brought a VR headset to Thanksgiving in the last 10 years, you're that guy.

54:14You are. By the way, when you look at that chart, we're not even anywhere near the early or late majority. We're nowhere near that for VR. Some things get stuck. They're sticky. In the early adopters era, yeah. It's important to know where you are because early adopters will buy anything. They'll tell you it's incredible. They'll try it. They'll buy it. They'll try it. They'll throw it in a draw. You know, you really need to get to that majority. And it takes time to get there. If the product is a startup, startups should not be going after markets where laggards are using the product. No. Because that means there's competitors.

54:54There's tons. And there's more competitors there are, the lower the margins are, perfect competition, economics 101, etc. Very hard to acquire customers for an Uber or an Airbnb competitor or a DoorDash competitor at this point because they got that locked up. Why would you switch? and to get people to switch requires a technological innovation or it requires being 10 times better, which most often comes from a technological innovation. So the example we gave of Taxi Magic, the iPhone with GPS and maps was 10 times better. And for some reason, Taxi Magic couldn't make that jump and Sidecar, Lyft and Uber did.

55:34Oh, I can tell you why Taxi Magic didn't make that jump is because it was only 15 % better. And you were still talking to our taxi company who didn't control their taxi drivers and you couldn't rate both ways and people were bad and they didn't take credit cards and they blah, blah, blah. It didn't solve the pain points. It just made it slightly easier if you're still at dinner to order a taxi. And this, what you pointed out was somebody who sat there and said, what sucks about getting a taxi? That was Garrett. That was Ryan. And that was Travis. They sat there and they obsessed over every step.

56:04And one framing is what sucks. Can we take that pain out? The other is what's delightful. Can we add that in? And another framing that people will often use in the startup world is friction. Friction is just, what's slowing you down? Another way to describe pain. And one of the things that always slows people down in these systems is payments and tipping. And I had major debates with Travis about tipping. I wanted to add tips early on. He did not. He was absolutely against it. He said, we're just going to make it fast for you to get in and out. if you have to make a decision, the decision should be to get an Uber.

56:43The decision should be to rate the driver if you want to, because that helps the network effects. Let's not add any other decisions. No more decisions. And there were just two decisions, ordering the Uber, putting in your destination, which is the decision to use it or not, which really isn't a decision. So what are the decisions to do after you've decided? Rate the driver. I mean, I think that's the only decision to this day in Uber. Rate the driver. You can pick black X. Oh, okay. You do have to do that, right. What's the, isn't there a middle option? Well, there's XL, there's teen, there's woman driver, there's dog driver.

57:20I mean, actually, because they're all below the fold, the newspaper term, below the fold of the newspaper, less important information. Because they're below the fold, you don't even see them. Yeah. Yeah. So this is a fair point. And adding that friction, there's a good reason to do it. Because if you have a lot of luggage, you need an XL. And if you're an executive and a CEO, you might want the black. And if you're on a budget and you're a kid, you might want the UberX. You know how people joke that all technology is bundling and unbundling and just kind of going back and forth? Sure. OpenAI, when announcing that they're going to, in the future, merge their reasoning and non-reasoning models together after GPT 4.5, they said, we also hate the model drop-down selector because it's kind of anti-consumer.

58:05You just want the best, fastest one. You don't want to pick between 4.0-mini or 0.3-mini flash. And you used to have to pick Dolly to make an image versus ChatGPT to ask a text question. Now, I don't think Dolly's in the menu anymore. So you should ask a question and it should determine if that question requires which model. In pro mode, you can expose those. and that's what it really should be. If you want to hit the dropdown, you hit it. Logging in is throttling. So ChatGPT, their monthly active users went bonkers when they took out logging in. Now they had to have logging in because they had paid customers and they had an SLA service level agreement that they wanted to do.

58:51So there it is. So why don't we talk about your favorite Founder University pitches? The reason we started Founder.University was about half the applications for funding we get at LaunchFun. That's the name of our fund. The applications come to launch.co slash apply. About half of them were people who were thinking about starting a company. They had a couple of co-founders, but they weren't. They didn't have finished products. They hadn't even incorporated. And so they couldn't come to our accelerator. And the way I designed the accelerator was to take seven companies. This last time we did 11.

59:24We did 11 because we had too many great companies coming out of founding university, to be honest. It's a 12-week course. You do it online. It's twice a week, Monday and Thursday nights. If you come for all 12 Monday nights, you submit your progress every week. And then we offer a 25K check as your first check at a million dollar valuation if you want it. Rich people don't take it. That's about 20 % of the people starting companies. The other 80 % ask for it every week. And then we also offer the standard Silicon Valley, Y Combinator, Techstars launch accelerator deal of 125K. When the companies get towards graduation, the best ones who make the most progress, pitch myself and Mike Savino, another general partner at our firm, my partner.

1:00:05And we rank them and we pick them. And we like to invest out of the 250 teams who come to each cohort. We just finished cohort nine. We like to invest in the top 10%, about 25 of them. So of the 100 investments we do a year, probably 50, 60 of them are founding diversity, 25K or 125K checks. And the other 40 come from the launch accelerator or our follow-on investments for the top 5 % of our portfolio. That's the portfolio strategy. That's what we're doing here. So these are very raw, year zero startups. But I just thought I would share some of them to give them some shine and tell you why we're considering investing in them, or we have invested in them, or why they were our top folks.

1:00:45And when folks do get an investment from us and they go through the Accelerator, I eventually have them come on the podcast if they make progress. But this is like, I just thought it was a new thing. we just run through their website show their websites you tell me the name i give you my feedback okay so let's pull up the first one then and it is poker gpt which is a combination of ai and poker jason two of our favorite things talk me through why this one is your jam okay so uh there is a very interesting business out there of training people to play poker these websites typically are subscriptions they cost 50 to 250 a month some of them cost thousands a month and what this product does is you give it your hand and then the hand gets analyzed by ChatGPT.

1:01:35What they're trying to do here is just make you a better poker player. So this business could reasonably get, I would say, hundreds of thousands of subscribers for 20 bucks a month. So it's a small community. But what I've noticed is there is a$100 million revenue opportunity out there for a company like chess.com. So I don't believe anybody's made the definitive poker coach. Search at GPT, you see what's happening with poker. It doesn't take a genius to figure out that there's enough of an audience, maybe not as big as chess, to get to low millions of revenue here, perhaps even eight figures.

1:02:16I don't need, as a year one investor or a year zero investor, to figure out, Alex, if it can get to 100 million, if it can get to low millions, I think as a team, we can figure out what's the next card to turn over. Pull up Tone Bass as an example for a second. Yeah, so to speak. You're not going to take your victory lap there? No. All right, so here we go. I'm just going to drop them. So here's Tone Bass. Here's Tone Bass. Tone Bass went through our accelerator. They're doing phenomenal. They started with, I think, guitar, classical guitar. And then they added violin. Yeah, I've hit the dropdown for violin, right?

1:02:50there. Thank you. Guitar, piano, cello, flute, trumpet, voice, clarinet, double bass, viola. They're now up to their eighth or ninth category. Jason, it's viola. Viola, thank you. Sorry, I've been in too many symphonies to have you defame. I don't know. It's not my jam. The middle stepchild of the stringed instruments. This is masterclass for these instruments. Now, you would think when they came to us with butt guitar or violin that it wouldn't have been a big enough opportunity and you would be right it wouldn't be big enough for venture my job as the world's greatest angel investor is to squint and say yeah let's get that ideal customer profile to our point before and then that wedge is in there open the wedge let's open the wedge get in there how did uber start lincoln town card then uber x uber eats uber pool uber xl you can rent cars on uber now you can take trains in europe you know it just never ends and so i that's what i think is going to happen in the poker space there's other poker games and if they figure out how to coach poker i'm a poker player who also plays chess there are people who are poker players who We'll also play Magic the Gathering.

1:04:08So I love the idea. I love poker coaching. I think there's something here. I want to see what happens on the flop or the turn. What's our next company? I played with it. Super cool. And it's got me excited because I'm going to New Orleans for a week, which means I'm going to be playing a lot of poker. I love it. All right. Next up is JS Mon. I think this is super cool because JavaScript, if you don't know, is pretty much the most important language out there for programming. We're seeing a ton of security startups and there's companies like Palo Alto Networks that acquire great companies, put the CEOs in charge.

1:04:44So there's a clear path to an exit here. And these security experts created JSmon to help people who want to protect websites and run reports on them in security. People are constantly running attacks, what they call white hat attacks. and they do security audits. So automating security audits, if you have the ability to do so, is going to be a huge business. AI is going to play a part in it. So it is what it says it is. And, you know, enterprise 500 a month, you know, security bro, 50 bucks a month. You get the idea. I would get rid of the free and the starter and just go right to the enterprise or security pro.

1:05:26And that's always the thing, you know, our number one piece of advice to founders who have products in the market is almost always, how did you come up with the pricing and double it i was about to double it yep double it absolutely all right another one jason this one surprised me a little bit that it was your jam i'm gonna do my best explain to people what wide worlds is so its tagline is make memes build culture drive growth and this is really interesting essentially there's a meme inside of the crypto world that community is everything and that's predicated on a fundamental truth that humans like to work in groups and we are all kind of community oriented folks okay so what they do is this company wide worlds will help you define your brand language use ai to help you generate a lot more assets for around that brand image and then use agents to help send those assets around to energize your community and i think it's a cool idea but i'm i'm curious what about it grabbed to your eye?

1:06:27Because to me, there's a lot of moving parts. I'm a little curious, but I'm a little skeptical. Okay. So this is a weird idea. And when I see weird ones, I get interested, especially weird ideas with attractions. There is a unique opportunity to create a character using generative AI. I think we can all understand that. And you see characters all the time. mascots is one way to say it uh for different products right so uh if you look at any cereal brand fruit loop she got the toucan sam aunt jemima with the pancakes although they've changed that now i think there's a politically correct they actually use the if you look that up for a second this is interesting my daughter was educating me because she's she's very up on these social issues they actually changed the name of aunt jemima pancakes to the literal name of the woman who inspired it because it's using an archetype, I think for Aunt Jemima, I know we're getting down like a little political wreckness here.

1:07:28But the Aunt Jemima is considered a mammy, quote unquote, offensive trope, racist kind of thing. So they came up with the actual name of the woman who was the inspiration of Pearl and it's now Pearl Milling Company. Yes. Formerly Aunt Jemima. Okay, I don't know if anybody knew this, but now it's Pearl Milling and they have the woman who's, do they still have the woman's photo or no? They got rid of that too. I don't think there's, no, it's just their brand image now, but I'm seeing eggnog pancakes, which means that I'm hungry. I had waffles today, man. I got this great waffle maker. Anyway, I don't know how I feel about this.

1:08:09Pepe the Frog was a funny meme that got co-opted by the Nazi, kind of white supremacist. Anyway, there's a whole range of characters here putting aside races. Tony the Tiger from Frosty Flakes. For getting me back on track. So, in order to make one of these and to make memes, you need to have a tool. And I thought, well, this is interesting. If we made a twist character that went with this for our brand, we could then have the twist tiger. I don't know, it's just an idea. No, no, I love it. And that we could make memes all day with them. It seems silly, but characters in IP are what drive the mouse house, Disney.

1:08:55It's what everybody's aspiring to in advertising. So maybe there is a business here. Again, odd with traction. This is an interesting factor. Something strange you don't understand with traction. Get curious. I'm curious. This reminds me of the fact that most scientific discoveries don't kind of commence with Eureka, but instead, huh, that's weird. Yeah. Thinking about the Twist mascot, we want to have alliteration. So we want T, T. Animals that begin with T include the Taper, the Tarantula, Tasmanian Devil, Termite, Tern, Tetra, Tiger, Tiger, Shark, Toad, Tortoise, etc. If you have an idea for us, AlexWLaunch.co, let me know what you think we should do.

1:09:39Twist Tiger is pretty dope. I mean, it kind of rolls off the tongue. Twist Tiger. twist tiger shark it's interesting twist turtle i'm kidding okay we have one more to do which is oh okay this one i struggle with but you like this we're going to talk about it ladies and gentlemen i give you lumix advertising lumix advertising here what they've done is on a vespa you know when you're delivering food there's a box on the back yeah and in this box which they created. It's got three sides of it, have an LED display. And I thought to myself, low margin outdoor advertising business, probably not venture scale.

1:10:17When I see an idea that's working, I get curious. So it's working. They've got prototypes out there. They're doing it in South Beach. And I thought, you know, Uber and DoorDash have a wonderful ad-based business, right? And so does Instacart, Amazon, right? You're at the shopping cart. So I started thinking about that. And I started thinking about food delivery. And I just thought, hmm, what else is there here? Again, open up the aperture, keep your eyes there. The advice I gave to the founder was to look at the leasing business. Uber had a leasing business. I forgot the name of it. You can look it up, Alex.

1:10:53It doesn't have the name Uber in it, but there was a leasing business associated with Uber at some point. I think Lyft had it too. They might've spun it out. And there are finance companies. Like if you look at car companies today, with the exception of Tesla, you know, Ford's leasing business or whatever. So here it is introducing exchange leasing lease options built for Uber driver partners. So one of the things people don't know is you, if you want to be an Uber driver, you may not have a car or you may have one car for your household that you can use for four hours, but you know, mom or dad have to go pick up the kids or they need it this weekend.

1:11:25And then whoever's driving in the household for Uber needs a car. So you can actually go rent a car. That's how taxis and medallions work in New York. You rent a car and the medallion for 130 bucks. The taxi driver drives like a lunatic until they hit 130 bucks, which on a Sunday might take them five or six hours. You understand why they're under some pressure because they got a 12 hour shift. They get to hour six. Now you know why they're peeing in a cup is because they're like, oh my God, I might lose money or I might make$200 today and I spent 135. So I just thought to myself you know i wonder if they provided and rented the scooter and the advertising box on the back what would happen then i thought to myself man if i was uber or doordash or i was like a local restaurant in partnership with them i don't know if you know like i think doordash and uber um they might have had exclusives with mcdonald's or um starbucks at different times I might very much want to buy that fleet out for the year and not have Dunkin' have it versus Starbucks.

1:12:27Oh, this is just like in search. If you search for Jason Calacanis and you're a brand, you have to buy your first ads. Thank you. Thank you. Everybody's buying Calacanis.com. Yeah. Yeah. If you do that, that'd be a great troll. If you buy Calacanis.com right now, I'm going to send an army of people. I'm going to find one of these Manila army black hats, gray hats to click your ad 10 ,000 times from US IP addresses and cost you a lot of money. Don't do it. Who? Oh, no, you own Calacanis.com. I own Calacanis.com. I'm just saying if you bought the ads on there. So anyway, I usually don't like these hardware things, but I do think I encourage the founder since they have traction to keep pulling the string and see what's happening here.

1:13:06Because maybe, you know, we had invested in this one other company that I felt like I got screwed on pretty badly and they had some traction, but it was constantly having problems. I do think maybe another entrepreneur, another team specifically taking a swing at this part of the puzzle without having to wrap a car could be interesting. So I agree with all of that. And I think actually, if you think about bringing this technology to markets in Asia that have high density and high scooter usage, for example, could be great. My only beef with this is that it's just very visually loud. And I still see it.

1:13:44Ubers go by with the light bar on top that has ads. And I'm always like, ah, okay. I guess everything ends up kind of being what it was before. Kind of like how, you know, YouTube TV was cheap and awesome. And now it's getting more and more like cable. I'll take one more question from the audience. There was somebody here, actually, you know, I hate to delve into politics too much, but somebody asked something about like, why don't I ever call out Trump or Elon? And I wanted to address that actually. If you look at my Twitter feed, I've literally talked about process journalism today and how an investigative journalism and how I felt like I believe the next step for Doge is to hire a team of investigative journalists to take the stuff that's spilling out of Doge, right?

1:14:25And I think the Doge philosophy is radical transparency, just put stuff out there. That was called process journalism. In fact, your alma mater, TechCrunch, believed in process journalism. I did not. Okay. The reason I didn't like process journalism is because that's what Gawker did to a more extreme level, they would release something that came into the tip line, and then they would use that as a cudgel, as a leverage to kind of get the story and get more information. The problem is, you know, you could spread something that's false. Yes. And that's why process journalism, sometimes it's good to slow down.

1:15:00Now, it's the people's money. The mandate is let the information out there. But as we've seen the information come out, hey, Reuters got this much money. Politico got this much money. And then we find out, okay, that wasn't Reuters, the journalist, that was Reuters, the data service, or that was political. The reason it seems like a lot of money is put. Anyway, my advice to them is, yeah, sure. You want to dump everything, do a data dump. That's totally fine. That's what WikiLeaks did. As a journalist, former journalist, I like a little more thoroughness, but I understand many people believe that they should be able to see the CIA documents for WikiLeaks or this or the JFK files.

1:15:43I like a little vetting, but I understand people disagree with me. Why not have five investigative journalists in a room every day talking about these contracts? Yes, the condoms in Gaza might be a different Gaza, whatever. Okay, how much are we spending on condoms? How much are other countries spending on condoms in other countries. Why are we doing that? Is that to prevent disease, to change birth rates? What's the CIA's goal here? You know, the USAID stuff was driven by foreign influence. A lot of times that's like a CIA thing or hearts and minds. Let's have an investigative journalist call people on the ground and say, hey, did this money actually buy 50 million condoms or did we get charged$100 per condom?

1:16:28Like that's the work of investigative journalists. The problem is investigative journalism is expensive and slow and slow so to your point do i ever agree with do i disagree with trump i told everybody i gave him a b and i that the meme coin i don't like i don't like the law fair whatever's going on with the mayor of um new york and doing that like leverage i don't like that either i didn't like the law fair when they did it against trump when they upped those charges i said that very publicly i thought the upscale of the charges was ridiculous like just give them the speeding ticket, whatever.

1:17:01So it's going to be chaotic. I can't even keep up with all this stuff. My job is to talk about startups. Yeah, that's what we're going to do here. When it overlaps with technology. Yes, I'm going to talk about it, but I cannot be responsible for everything David Sachs, Elon, and everybody else is doing in the world. I have countless friends in and around the administration. I'm going to stick to my lane, largely startups and technology. No vicarious guilt. I mean, I'm questioning things all day long. Are people not paying attention? They're not. Okay. So, I mean, just to give you an example of this, we touched on Palantir the other day.

1:17:39And honestly, we talked about how Palantir is a great company with a visionary founder and products that are resonating in the market. And then we said, you know what, though? Maybe this valuation is not entirely moored to fundamentals. And I saw a lot of responses to Thrillette talking about our conversation. And people were just claiming that we're like on the take, shorting the stock. Like one, who would pay us to ramble about Palantir? I don't want to use the R word on the program, but that is Ray. We are not stupid, y 'all. So anyways, it's people need to relax. The thing that I'll say is, and I know we're going to get to this on Monday, but just my little point is right now, there's a lot of people who are posting pro-doge things to engender favor with the people on X that are the most influential.

1:18:28This is not just Elon, this is venture capital community and so forth. And I think that what they're doing is often getting ahead of their facts and their ability to understand the data. And so we're having wasted news cycles of BS that could have been precluded with traditional reporting standards. I love citizen journalism. I love having more voices. I've been on Twitter since 2007. But if you're going to make an explosive claim about criminality uh you know somewhere between a data dump and you know biased journalism on the margin whether it's msnbc npr the daily wire you know opinion journalism there's investigative journalism there's thoughtful center middle of the road journalists just pursuing truth it's not that many anymore everybody seems to have picked a side in journalism there's a lot of sensationalism.

1:19:20I just wish they were like more centered journalistic organizations. I'm really dismayed that that model doesn't work anymore. So what we should do, and I know we're going to say this mostly for Monday, but what if we founded the Center for Boring Journalism? Yes, unprofitable journalism. We are going to talk to you about these things and you are going to be slightly bored, but they are what matters. The Center for Boring and Unprofitable Journalism. as opposed to tribal journalism. That would be pretty great. All right, everybody. It's a lot of show. Next week is President's Day, so we'll be back on Tuesday, I think.

1:19:58Yep. We'll see you all next time. He's x.com slash Alex. I'm x.com slash Jason. We'll have Lon come on on the news on Tuesday, and we'll see you all then.