SPEAKER_00: I want to make one more little point about the labor discussion because Andy Jassy, the CEO of Amazon, dropped a public memo yesterday and the key quote, which I just pulled off. I was thinking like, Jason, it's almost like you're getting the inside scoop. This is what he said about Amazon and their deployment of generative AI products. He says, as we roll out more gen AI and agents, it should change the way our work is done. We will need fewer people doing some of the jobs that are being done today and more people doing other types of jobs. It's hard to know exactly where this nets out over time, but in the next few years, we expect that this will reduce our total corporate

SPEAKER_04: workforce as we get efficiency gains from using AI extensively across the company. That to me is so,

SPEAKER_00: yes, true. I mean, we talk a lot about productivity improvements and when you have more productive individual units, you need fewer of those units. So that to me is the non-doomer take though,

SPEAKER_07: because the doomer take is white-collar work is going to evaporate overnight. That's not going to

SPEAKER_08: happen. This Week in Startups is brought to you by Squarespace. Turn your idea into a new website. Go to squarespace.com slash twist for a free trial. When you're ready to launch, use offer code twist to save 10% off your first purchase of a website or domain. Inbound. Connect with visionary leaders like Dario Amade and Amy Poehler at Inbound 2025, September 3rd through 5th in San Francisco, the epicenter of tech innovation, and transform your business strategy for the AI era. Use code twist 10 for 10% off your general admission ticket at inbound.com slash register. Valid through July 31st. And Northwest Registered Agent. Starting your business should be simple. With Northwest Registered Agent, you can form your entire business identity in just 10 clicks in 10 minutes. From LLCs to trademarks, domains to custom websites, they've got you covered. Get more privacy, more options, and more done. Visit northwestregisteredagent.com slash twist today.

SPEAKER_11: All right, everybody. Welcome back to this week in startups. It's your boy, J Cal, Alex Wilhelm here

SPEAKER_13: as well. How are you doing, Alex? I'm doing fantastically well. I'm caffeinated. I had a shirt

SPEAKER_17: to wear. I'm feeling great, man, but you're still, you're still on the road somehow. I'm on the road.

SPEAKER_16: You can tell when I'm running out of, you can tell when I'm on the road, I'm running out of shirts. You know, I've got to do a second laundry run. Oh yeah. Yeah. On this thing. And I got a little

SPEAKER_18: bit of a thing. I can't, I get really upset about paying for the hotel to do my laundry.

SPEAKER_19: It's a per piece item. It's $20 for a t-shirt. It's $12 for a pair of underwear. It's $6 for a pair of socks. I'm like, you could buy a pair, a t-shirt for $20. It makes me a little mental.

SPEAKER_20: And so I like staying at Airbnb houses. So I'm going to be at an Airbnb in LA for the, we're having a little party for the launch of the all in tequila on Saturday, but I'm not supposed to, I don't think I'm supposed to actually like broadcast that too, too widely. So I didn't tweet it or anything, but yeah, still on the road. And I'm at this conference, which I think I could say is a co-to, you know, co-to that big firm. It's their conference, East meets West, 150 people.

SPEAKER_19: A lot of interesting discussions going on. I did a little all out panels, me, Sachs, Brad Gerstner, and Philippe, you know, the co-founder of Kotube. And yeah, we had a big debate

SPEAKER_20: about immigration and AI and job disruption. And I kind of gave my little, I don't want to say

SPEAKER_19: stump speech, but I gave my position of why a lifelong optimist is now doomerist curious. Okay. I'm kind of find myself like a, like a hanging out with the, the doomerists and, and, and

SPEAKER_20: chilling with them because I was talking to the guy from Zipline, which has that incredible, um, drones. They started with fixed wing, but you remember Zipline and I think they're in the, uh, twist 500 are, are tracking up 500 most important private companies, twist 500.com. Long story short, you know, he's really doing a lot. I don't want to give any numbers, but I had a meeting with him yesterday. He's doing a lot of deliveries and, uh, you know, they have

SPEAKER_19: Walmart. He's got a bunch of other similar customers and DoorDash is doing their own drone delivery, uh, internally. And so, you know, Zipline is, um, no, I don't think that's it.

SPEAKER_30: Is that Zipline? Oh yeah. So the, the video loops around and you can see the, the little

SPEAKER_19: box. Yeah. So for those of you watching, the way it works is it stays 300 meters in the air or a hundred meters in there, 300 feet in the air. And there's a little cargo that zips down on a tether, releases your stuff in your backyard and then goes away. So the drone

SPEAKER_20: is so high and quiet, which makes it quiet. So if you have a backyard, it drops it, but the pickup is really the interesting thing. So those little boxes are, um, you know, easily

SPEAKER_19: filled by somebody at a cafe or a sandwich shop, you know, pick your poison, uh, and they're doing a lot of, you know, eight pound deliveries and these things come in minutes, you know, sub 15 minute time delivery. So it's going to really change it back to AI and how, and how this is making you doomer adjacent. Cause I'm really curious. So automatic delivery is a big one. And Andy Jassy spoke here and I got to ask him two questions and I asked him about

SPEAKER_20: return to work and the kerfluffle member was all this kerfluffle. And he said, you know, it's actually turned out to be not that dramatic because the people who we, we knew going into it, we would lose people, um, who we didn't mind losing that weren't essential. And we lose

SPEAKER_37: people who we loved working with, but we decided we wanted to have a certain culture and it's turned out the culture's back at the company. And so we love it. So that was my second question. My first one was, Hey, job description, job destruction, disruption, and retirement. You know, you've seen these, the figure robot sorting things, you have tons of robots in your factory, self-driving, you're doing souks. It seems to me that when a lot of those jobs go away, and obviously you're not hiring a lot of people, you have less people at Amazon than you did a couple of years ago. So it's happening in white collar. And, um, he was basically very upfront and candid that he did think there was going to be, um, a lot less people working at companies and that it was something society would have to navigate. Yeah. And that is the tension. And, you know, when you say that to a room full of investors, and I just said, you know, I think we have some responsibility. Uh, I don't buy into like technology is just going to technology and let the chips fall where they may. I think we actually should think about it. I know it's a bummer, but we should think that, Hey, some jobs will be disrupted. Now, is it our responsibility to create all the new jobs in the world? No, no, of course not. But I think we should just be thinking it as human beings on planet earth, because there could be some disruptions and we have safety nets and the safety net used to be being a door dash or an Uber

SPEAKER_38: driver or working at Starbucks. And, you know, now the safety net that we created, like, I think gig economy work is that is the great safety net. We might be looking back, remember all that hand wringing and the press and all the union leaders. Oh my God. You know, the skig work is terrible. You got to work a shift. You got to come in at 6 AM, work till, you know, 4 PM, whatever it is. It turns out like, we're going to look back and be like, wow, that, that gig work was like key. You could pick who you wanted to work for. You could work for three different people. They were all competing with each other for your work. If you didn't like working for one of the customers, someone you can go to the next. So long story short, we're going to have job

SPEAKER_20: destruction. Everybody agrees on that. The only thing people don't agree on is the timeline. And if we can create new jobs in that time period, and then Philippe had something, you know, I think, I think it was Philippe who just said something really interesting about like, people have to also be motivated to start jobs, uh, to create jobs. And that's going to be very interesting. Very interesting moment in time where, you know, group of five, 10 people, like when you graduated from college or I did, when I graduated, there weren't a lot of jobs in 94, 92, 93, we're coming out of like, um, a recession and everything. Some people didn't get jobs. They just stayed at home. I think that might be a trend that we're going to see young people without

SPEAKER_00: jobs. Yeah. So that means we're looking at job destruction now from two different perspectives, because we've talked a lot on the show, Jason, about our concern that AI is going to automate early career positions in law and programming and other kind of like knowledge, heavy industries. But now if we take away the gig economy safety in that as well, I mean, that does really give people right out of college or just right out of high school, frankly, a pretty limited set of gigs.

SPEAKER_19: That's, that's worrying. Well, and it's happening to white collar too. So, you know, we we've seen in big tech that they're like, yeah, we're not going to hire young people because we don't have to,

SPEAKER_20: we don't want to have to train them. And the time it takes to train a young person versus the time it

SPEAKER_11: takes to automate some amount of work at your job, which might be easy to automate it. So I think this is the tension that's going to become very real in the next, you know, five years over the next five

SPEAKER_49: years. And so anyway, but that means there's a job that means there's an opportunity. If there

SPEAKER_20: is surplus, cognitive surplus, white collar, and there's physical surplus, people doing things in the real world, moving atoms, moving bits. If you had a surplus of young people, what could you do with

SPEAKER_52: them? What could you do with them? That would be a creative. I don't know. All right, founders,

SPEAKER_53: let's talk about your website. I know, disgraziad. You're ashamed of your website. I know. Well, it's time to clean it up. Give your brand a quick refresh with Squarespace. That's the all-in-one platform that makes building a stunning, professional, gorgeous website ridiculously easy. Doesn't matter if you're selling products, doesn't matter if you're offering services, or, you know, if you're just showcasing your portfolio, Squarespace gives you everything you need to grow. They've got this great new AI product. It's called Blueprint. You got to try it. You just answer a few questions and you get a fully customized website in minutes, personalized layouts, on-brand visuals, and voila, you're done. I've been using this product for over a decade. Check out squarespace.com slash twist for a free trial. And when you're ready to launch, go to squarespace.com slash twist to get 10% off your first website or domain purchase. Once again, squarespace.com slash twist.

SPEAKER_56: I mean, I have a lot of ideas, but they're not very good capitalism, but like, you know,

SPEAKER_00: we could use a lot more cool trails through the national parks and forests. We could use a lot more painters. I don't know, but nothing that's like plus GDP immediately comes to mind.

SPEAKER_20: So if you have a lot of data to do professional development or recover from teaching and you I'm sorry to sound like a socialist, but you just increase the number of teachers, 20%. So maybe like if every teacher took off Friday and just back of the envelope map to make it easy for people to understand my, my point here. If every teacher took off Friday and had a professional development day, right. Uh, to research stuff, whatever, uh, or maybe do one-on-one tutoring, then you need that theoretically 20% more tutoring. So you hire, you know, whatever, a science person who rotates and takes the kids, you know, uh, 20% of the kids each day, or maybe you just take class sizes from 30 on average and make them 15. Yep. There are 3.8 million public school teachers

SPEAKER_00: in the K through 12 system today. So if we can make that 5 million and get class sizes done, that'd be tremendous. Um, I want to make one more little point about the labor discussion, because Andy Jassy, the CEO of Amazon dropped a public memo, uh, yesterday and the key quote, which I just pulled off. Yeah. I was thinking like, Jason, it's almost like you're getting the inside scoop. Yeah. Um, this is what he said about Amazon and their deployment of generative AI products. He says, as we roll out more gen AI, gen AI and agents, it should change the way our work is done. We will need fewer people doing some of the jobs that are being done today and more people doing other types of jobs. It's hard to know exactly where this nets out over time, but in the next few years, we expect that this will reduce our total corporate workforce as we get

SPEAKER_04: efficiency gains from using AI extensively across the company. That to me is so yes, true. I mean,

SPEAKER_00: we talk a lot about productivity improvements and when you have more productive individual units, you need fewer of those units. So that to me is the non-doomer take though, because the

SPEAKER_07: doomer take is white collar work's going to evaporate overnight. That's not going to happen.

SPEAKER_68: No, it's, you know, there, you're going to need less people because people will do more. That's the upside. That's what could save the entire economy. You know, the, the, the idea that we've had so much disruption geopolitically tariffs, you know, the Israel bombing, Iran, that conflict, um, you know, all this type of chaos we're seeing immigration, whatever, uh, changing in the United States here, everybody knows my feelings on that and the stock market continues to rise. So, you know, putting aside the, the, the freedom day or whatever that, you know, shock and bore wound up being, you know,

SPEAKER_11: shocking 150% tariffs, then bore. We just want reciprocal, you know, that's the standard playbook for negotiations with Trump shock and bore. It's like crazy shocking. And then it's like, okay, this is kind of boring. We can just not pay attention to it. Uh, it's not going to be impactful

SPEAKER_26: in a major way. Like, it's not going to be disruptive. It could be impactful, but not

SPEAKER_73: disruptive. And it would be a fair way to say it yet. Here we are. Stock market keeps climbing

SPEAKER_74: and the feds not cutting rates. So they kind of have a, they have a perspective too. Yeah. So

SPEAKER_00: their perspective is economy's good. Economy's good. Uh, I checked the data this morning, according to fed watch, no one's pricing in rate cuts until September, at which point investors are a little split between a 25 or a 50 BIP cut, which is 0.25 or 0.5%. But yeah,

SPEAKER_79: it's going to be static for a while. I just find it very surreal to your point that my family's net worth is probably at an all time high. And yet when I read the news, it's like the world burns down

SPEAKER_82: again. And I'm like, oh, okay. I guess. So weird, you know, split screen. Here's a poly market,

SPEAKER_43: fed rate decisions. Yeah. So June, July 30th, September 17th. Am I on the right thing here? Let's

SPEAKER_68: see. Fed decisions, fed rates. Yeah. So let me, uh, I, I have a bigger, better screen for that. Sure. I got it. Yeah. So 55% chance, no change in September and everything else is, uh, 41%,

SPEAKER_89: 25 BIP. So, you know, we're basically now down to 50, 50 on screen sharing. I want to show you

SPEAKER_00: something funny. Yeah. Okay. So, uh, I had prepared a different poly market for today, but we're doing this early. So this is totally fine. Um, so I just pulled up the same page and look at what they've put. They actually literally have the live stream from the fed open markets committee

SPEAKER_35: on the page for, I love that. Yeah. You know, it's, it's kind of working back from,

SPEAKER_20: you know, you and I were talking about how journalism is changing, um, you know, and I, I jumped up the journalism boat a little earlier. Uh, you're still like half journalist, you know, um, and, uh, half, uh, entrepreneur now, I think like, this is part of the future of journalism, which is starting with every placing a bet and saying, here's my skin in the game. Yeah. And then, uh, you know, placing, you know, literally putting skin in the game and then working

SPEAKER_19: backwards to the discussion. And so if you think about that versus traditional journalism, and again, isn't not here to beat up journalism, the world evolves, you know, when you see people talking about sports, now they're talking about their wager on the sports on ESPN and giving their position, having a bet, and then they watch the game. And then during the game, they're talking about maybe changing their wagers or how the odds have changed. So if you just think about what are consumers wanting from media, I think they want, despite all the fake news and, you know, all the, you know, everybody's terrible at this job or whatever, there's a group of people who are saying, Hey,

SPEAKER_37: we're going to make some hard decisions, put skin in the game. And we have a more interesting way of getting you the news podcast is one of that groups. That's why I'm always trying to make this podcast better. When we cover companies, Alex emails them, emails the CEO and says, Hey,

SPEAKER_20: we're covering you today. Here's two questions. And that's why we get inside information because we

SPEAKER_37: literally just do this old school thing of doing that. But we also talk about bets and we make bets on this program, although we have to do better at following up on those bets. I need all the sunny

SPEAKER_105: bets looked up, uh, uh, editorial director lawn. And so we can maybe settle some of those.

SPEAKER_106: Okay. Got some notes, got some notes. No, no, I'm just, I'm enjoying, uh,

SPEAKER_00: the discussion of how things are changing in regards to journalism, because one thing I'm still wrapping my heads around my head around is simply the power of industry insiders who are conflicted to have an enormous media position discussing, not just their industry, but the industry at large. And I, that is so antithetical to the way things were done in the eighties, just to roll the clock back far enough to be uncontroversial that I don't think we fully digested what that means for truth telling, but I do agree that it is incredibly popular, impactful, and people do find it useful. So I think we are going to see more hybrid stuff. Um, speaking of which, did you see that?

SPEAKER_111: Just to build on that position, good position. Um, used to be, if you wanted to tell your story,

SPEAKER_20: you were Andy Jassy, you had to, or you were Jeff Bezos, you had to go to the New York Times or Wall Street Journal, give them the story. They interpreted what you said, and then they put it out there. And in the best case, you felt like, okay, yeah, that was, that was fair and balanced. In the worst case, you felt like they spun it and had a position they were going for. And it depended on the publication, depended on the journalist. And over time, people felt like maybe it wasn't balls and strikes. So they went direct. So now you have direct podcasts with the principles on it. And, you know, like if you watch all in, uh, and the four of us are talking, we're talking our book by default, the audience understands that if you look at like the back channel, they're like, okay, David Sachs works for the, you know, for the administration. So he's going to be in solidarity with them. That's how it works. You know, Chamath has these investments. Jason has this investments. Freeberg's a man of science has this, uh, you know, uh, a hollow company. Uh, and so I think the audience is saying, I understand the conflicts. I'm willing net net to take that. I think ultimately where it winds up is a hybrid. I think it's a hybrid,

SPEAKER_14: you know, I agree. But if you think back to when blogs were new, like literally back when

SPEAKER_79: web blogs were still called that, I remember when a company would open a blog, people were like,

SPEAKER_00: what they're just going to like post on it and write things. Well, today I, for my newsletter and for the show, pulled up the Amazon, sorry, the about Amazon webpage to read the memo from Andy Jassy directly. I did not need an intermediary. And that's kind of the, the gist of what we're

SPEAKER_114: talking about. Well, and that's probably because he doesn't want to be interpreted. He wants to

SPEAKER_37: be taken at his word because he feels this is super important. So that's actually a really important insight. And I asked Steve jobs, why don't they blog at Apple? It's like a famous

SPEAKER_11: clip from the D conference. The wall street journal had a conference called D back in the day with Walt Mossberg and Karis Fisher. And I was, but a, you know, 30 year old, uh, you know,

SPEAKER_37: running web log sync, I think at the time. And I asked him like, why don't you blog? And so it takes too much time. And then people are going to say stuff and it might not be in sync. So yeah, it's not for us, but now you see people going direct. And now you start to see pockets. I saw the Carlson brothers, the stripe, speaking of startups, stripe launched their own conversations. I think Connor Perkins just launched a blog. I'm like, my God, how many blogs are there? Like, does anybody

SPEAKER_57: gonna get any work done in this industry? Pretty crazy. You know, here at this week in startups, we try to keep founders just up to speed on all these trends in AI and marketing, but we can't do it alone. But you're also going to need to connect in person with your colleagues and partners and innovators. There's no substitute for in person. So if you want to connect with visionary leaders and personalities like Dario Amodi, Victor Ripperbelli, Dharmesh Shah, my friend, and more, your chance is to do that at Inbound 2025. That's right. The epicenter of tech innovation that will help you transform your strategy as we approach the AI era. It's happening this year from September 3rd to 5th in my old hometown of San Francisco, California. It's a rare chance to really immerse yourself in San Francisco's unique venture ecosystem where new companies can quickly evolve into billion dollar unicorns and beyond. Use the code TWIST10 at checkout to get 10% off

SPEAKER_119: your GA ticket at inbound.com slash register, inbound.com slash register. That's inbound.com slash

SPEAKER_122: register. Okay. Well, if we're going to do that, here's a, here's a website for you, Jason. Here's this from open AI introducing the open AI podcast from three hours ago. So yes,

SPEAKER_00: everyone is now doing this. I think probably because Sam Altman just wants to be able to kind of like

SPEAKER_79: not answer the same three questions 10,000 times, but I think it's going to become the norm. The

SPEAKER_126: question is what's the first company with its own podcast media publication, stable coin. Yeah.

SPEAKER_19: Sovereign everything. Yeah. Yeah. Uh, well, you know, we'll see who keeps publishing. You know, that was always the interesting thing. People underestimate the amount of work that it takes.

SPEAKER_20: And here we are 2000 episodes into this week and startups three days a week. And I just,

SPEAKER_11: I'm a machine about it. You know that, uh, you work with me. So I think a lot of people will just give up because it is arduous to put on a performance, you know, every day or every other day. I do four days a week. You do three. Like it's, it's not easy. It's not easy to have that kind of

SPEAKER_26: consistency as a player. So, all right, what's in the news. Let's get back to the news now. Here we

SPEAKER_122: go. All right. I want to talk about a couple of twist 500 companies that are in the news. First up,

SPEAKER_79: Pano AI is a company that I added because I thought the bet that we could use AI to detect wildfires was

SPEAKER_00: a pretty good one. And what they do, Jason, is they put the cameras on top of like cell phone towers. And then during the day, Pano AI looks for essentially smoke and at night it looks for heat. And the idea is what if we could just better detect and crack wildfires much more quickly and easily. I put them on the twist 500 a little while ago. The latest news is that they have raised a 44 million dollar series B led by giant ventures with participation from both initialized capital and Salesforce ventures. They're selling Jason to people that you love, including private, uh, ski resorts and landlords. So your two favorite things owning ranches and going skiing. And also of course, governments and utilities. The key thing here that I'm really excited about is not that they raise money. That's cool, but not really proof of progress. Uh, the journal wrote that they have now

SPEAKER_134: reached contracts exceeding a hundred million dollars, which I think shows strong early commercial

SPEAKER_136: momentum. Yeah. Um, and you can see it there on the screen. If you're watching, it's, you know, very simple. It's, um, you know, a camera high up and a bunch of software on the cloud. You'd think

SPEAKER_68: about how much it costs to deploy this, you know, these cameras are probably, you know, commercial

SPEAKER_20: grade couple of thousand dollar cameras. Um, I'm sure they make them pretty rugged because of where

SPEAKER_68: they're located. Um, and then a pretty good connection, but you know, 5g exists everywhere. And, um, this is, I think, you know, probably

SPEAKER_20: what people are underappreciating about it because I'm sure people were putting up cameras, uh, and looking for smoke previously, you know, but here, if you look, it's showing you like where you have transmission lines and main lines and, um, the wind speed, wind gusts, temperature, humidity, and then what assets you have, right? So when it shows a fire here, it's showing on the screen, like, oh, there's a Valley Ford fire. It also has two nearby assets and it's 1.2 million to the closest asset asset. I'm assuming here is something to put out the fire. So what we, and we're thinking about jobs, right? As to our early part here, this will be a complete closed loop system shortly. I've brought up zip line before the power of drones. There is literally going to be helicopters or drones that have large water payloads that are going to be able to go out without waking somebody up in the middle of light. Firefighters not going to be able to need to get their gear on and run out to this fire. They're going to have a drone go first drone is going to go and get above the fire and just monitor it, right? So now you got a quad second drone quad captor is going to get out of the way or film from a distance is going to drop a payload of water or fire retardant on these things. The other thing that's going to happen is this is going to be doing an analysis and the AI for an algorithm, instead of, you know, showing you your next video and tick tock as you're melting your brain and frying your dopamine sensors is going to say, these places are where the fires keep happening. And then you say, Oh, this place waters and has,

SPEAKER_38: you know, better vegetation. There's never a fire there. So over time, they're going to know exactly or this idiot keeps burning their garbage and sending embers up. The second, third order impact of this kind of a system is going to be incredibly underrated, incredibly underrated. You're going to catch the arsonists when they know there's fire. Oh, you're just thinking it? Yeah. Well,

SPEAKER_00: I mean, if you have a combination of, uh, a time series, a bit of information about where fires began and you cross that with GPS data from cell phone locations, you're going to be like, well, that started and there were seven people nearby. And one of those phones has been near the last three hotspots. And then we're going to take that person and we're going to three point them right into jail. The downside though, Jason, is that my backup career, which has always been going to sit in a firewatch tower and not talk to humans for 10 years as apparently now sufficiently indicated that I will not be able to do that. And that's a real, there's a, there's a really good, uh,

SPEAKER_147: there's a really good sitcom or short run series here, uh, of the last fire warden, you know, the

SPEAKER_20: what do they call those fire watchers? What do they do? What are they, what's their technical name? I, I don't, there's gotta be some great name for them for this new series coming from Calacanis studios. Uh, the story of the last fire watcher, uh, and they have to shut the, uh, thing down and

SPEAKER_153: they're fighting to keep it going. But there's a robot in there with them. There's like an optimist or a figure robot living with them that they're training to do the job. And their reward for

SPEAKER_20: training them is that they lose their job, but then they decide that the robots would be bored and, uh, sad being alone. So he decides to stay anyway, and he makes a business doing some other business, you know, running a campground around the fire watch tower. And they tell stories about

SPEAKER_157: the old days lawn's gonna quit and go write that, uh, no, no, this is part of lawn being here.

SPEAKER_20: Lawn is here so that we can do our creative ideas and, you know, we have a secret project we're working

SPEAKER_56: I've fallen, I volunteer to be the, to do this. Uh, by the way, they're called fire watchers.

SPEAKER_159: So that's a pretty on the, on the nose. Fire watcher. My question is what about selling to

SPEAKER_00: governments? Because back in the day, people said, don't sell the school districts. Don't sell to counties. It's too slow. It's too hard. It's too painful. But here we have panel AI, which does sell the government's doing quite well. So I'm kind of curious, Jason, when is it a good choice for a startup to sell into hard markets like that? Yeah. So we're talking specifically about our

SPEAKER_20: government, uh, you know, being slow and incompetent and filled with grip and fraud and stupidity and waste. We're not talking about the UAE or Singapore. Uh, and you know, if you go around the world and if you show this to people in Singapore, Singapore pays people working in government three, four, 500,000 a year, they are the premier employer. If you are a smart person in Singapore, you would first look for a job, not at Goldman Sachs or at Google, just picking them because it's G's you pick government, government over Goldman and Google. That's what you do in Singapore. How do they accomplish that? They have created a sense of pride with working in government and a sense of like, that's like, um, one of the most proud things you do because of Lee Kai, um, Lee Kwan Yu, uh, this incredible leader of Singapore. And when you watch his videos of how he built that country, man, he's like the iron

SPEAKER_37: fist, he built that country. But one of the key things is he said the best and brightest work for government. And so, you know, there's people you can sell into there's, you can go for those governments and then you can go for people who just landowners. I, I, I heard you mention they were going after ranch owners, et cetera, who have to deal with fires too. And by the way, I was spending time with Rick Caruso who ran for mayor, lost to Karen Bass in LA, Karen Bass, complete incompetence, disgraziad, like the worst leader you could ever have during a crisis. Uh, as we saw her like, freeze up when they were just asking her questions about what you're going to do when she was on some boondoggle out of the country while the city burned. And, uh, I hope he runs again and I hope

SPEAKER_165: they, they put him in there, but. Investors like me are not going to invest in your business unless

Chamath Palihapitiya: you're structured properly. So founders, if you're serious about raising money, you need to set up your business the right way. And that starts with a registered agent before a venture capitals can wire you a single dollar. They're going to check if your company is incorporated and it's in good standing and it's compliant. That's where Northwest registered agent comes in for just $39 plus state fees. They're going to handle your paperwork, keep it compliant and make sure investors see you as a serious business concern. That's worth putting money into in just 10 clicks and 10 minutes. Your business is officially investor ready. Thousands of founders trust Northwest. So don't let paperwork cost you your next funding round. Go to Northwest registered agent.com slash twist and get your business investor ready today for just $39 plus state fees. You can set up your company the right way, fast, private and compliant. Go to Northwest registered agent.com slash twist today. You know,

SPEAKER_20: he figured out this incredible technique and he created the Grove. He created this place in Pacific

SPEAKER_37: Palisades and he came up with this incredible technique, combating fire. It's almost as perhaps arguably more, um, innovative than Panos. It's called water firefighters.

SPEAKER_170: Incredible combination. So when, so you have people who use water, right. To put out the fires.

SPEAKER_37: Yeah. So when, uh, you know, this place I'm at this Rosewood here in the Santa Barbara area,

SPEAKER_20: Montecito was gorgeous. When they had fires here, you just like, they knew it was fire season. So

SPEAKER_142: they just stationed a couple of water trucks and then there are private firefighters you can hire.

SPEAKER_37: Yeah. So they were like, okay, we'll hire some private firefighters. If sparks come, we can put things out, yada, yada. And then they build things to be a fire resistant. This is all just planning and will. And, um, so if, if one government group can't do it, I like that you keep your costs low and you sell into people who are impacted by this. The insurance companies are going to love this as well because they are losing their ability to provide insurance in certain regions in America.

SPEAKER_20: And this would bring it back. So yeah, ski resorts would have an easy time paying for this. And I think this is a solution that's probably, I don't know, a hundred thousand dollars to install five of these cameras around your ski resort, you know, and then, uh, maybe they charge 50 or a hundred thousand a year, maybe they charge 10 K a month for something like that. And so you're looking at going, okay, 120 K a year, a million dollars over the next 10 years. What do I save

SPEAKER_00: on fire insurance? Maybe it pays for itself. Oh, I could easily see that. Yeah. But, uh, just to underscore what Jason's saying about insurance companies having a hard time insuring, this is a data set from NOAA, uh, the government agency tracking billion dollar disaster events, uh, that are essentially driven by, you know, droughts, wildfires, flooding, et cetera. And Jason, there's a pretty clear upward trajectory to this chart. So as the climate gets a little more unpredictable, a little bit more difficult, we're going to have a lot more problems. Wildfires are in this list, by the way. So I do expect a lot of companies to try to make insurance possible for people, because if we don't have insurance, Jason, who is going to subsidize sports broadcasts on cable television, if it's not Geico and fricking farmers, right? Like someone has to pay for sports. So we've got to keep the insurance companies in business.

SPEAKER_14: Yeah. Insurance is critically, I mean, insurance will change. So, you know, this is an interesting

SPEAKER_20: thing for us to discuss as we move into AI boom, AI doom, AI boom, AI doom. That's going to be our new segment lawn. Make us a jingle. AI doom or AI boom should be like me with a devil on his side. Like

SPEAKER_183: one person on the side, like it's the end of the world. The other person saying it's going to be like

SPEAKER_20: utopia. Um, the AI boom side of this equation is costs go down. So if you were to eliminate wildfires as a concept, well, then insurance goes down. If you have a car that can't crash or you reduce the ability to crash 99%, well, why do you need insurance? Or if your insurance is going to go down 99%. So with technology, you know, we should see massive amounts of money in the economy free up and insurance is a big giant waste of money that should go away in many cases or be reduced at

SPEAKER_42: least 80, 90%. Well, you're saying that if you can reduce the underlying risks, then make insurance a requirement, then we don't have to have that essentially risk rent on the economy. I mean,

SPEAKER_68: do you have insurance for, um, you know, I don't have ski insurance as an example. I mean, I don't even have health insurance, but I don't have like insurance on my skis, you know, braking

SPEAKER_20: or something. I don't have insurance on like this laptop braking. Now they upsell me on like, get the insurance and laptop brake. I'm like, I replace it every three years. The chances of it braking is very low. They make these things like, you know, remember everybody used to get for their phones insurance for the screen braking and then gorilla glass. You can literally drop your phone 10 times and it will not break. I mean, don't do that with an old phone. It will break on the first try. And literally for young folks with your iPhone, you paid a hundred dollars insurance when you bought it and you got one break. And then they would, you go to the store and they would replace

SPEAKER_142: it. Now it was worth it. It was worth it because you were, it was going to slip out of your hands and you break it. Now. I mean, I, I kicked my phone across concrete all the time. I'm distracted. It's

SPEAKER_105: breaking and, and you know, I get little chips in it, but I've never broken the glass. When's the last time the glass broke on an iPhone? It hasn't happened to me. Three phones ago for me. Yeah. Yeah. It was three or four phones ago. Right. So it's a perfect example of technology,

SPEAKER_142: reducing expenses, making things more rugged. And we love that. What's the Reddit I always talk

SPEAKER_00: about buy it once. Uh, I believe it's buy it once. Yes. The idea of purchasing a high quality thing. Although I will say you have given panel AI really great second act, because if they do figure out this wildfire detection system and it works as well as they want, and they get the coverage they hope to have, all they have to do next is just team up with a drone company and launch their own

SPEAKER_79: firefighting service. Like what, why not? They already have the data. Yeah. I'm sure there is

SPEAKER_20: somebody making drones for firefighting already. So that could be like an API call or they could just

SPEAKER_11: white label it, or they can make it themselves. Uh, it's clearly what's going to happen. Yeah.

SPEAKER_193: Buy it for life is the Reddit, by the way, not, not buy it once. We were slightly off.

SPEAKER_194: Buy it for life. Yeah. Great, great, great, great subreddit.

SPEAKER_00: One more little news bit. And then I want to talk about some founder stuff, but really quickly, we've talked about scale AI and it's partial sale to meta. We also discussed a little bit about how some of its customers might be backing away from it. Well, open AI said in the immediate aftermath of that deal, that they weren't going to back off with CFO, Sarah Fryer saying, you know, we don't want to ice the ecosystem because acquisitions are going to happen. If we ice each other out, I think we're actually going to slow the pace of innovation. That didn't last. According to reporting today, open AI is going to move entirely away from scale AI, which means I believe that this aqua

SPEAKER_79: hire of scales founder and some of its technology is going to pretty much kneecap the company.

SPEAKER_196: So I guess there's going to be a lot of talent from scale AI that's available for other companies to hire, but it does seem a little shocking. I don't know about that. Yeah. Okay. Okay.

David Friedberg: It might be that Zuckerberg is so ambitious that he will keep everybody employed and just make his AI

SPEAKER_20: make his AI the best one. So I think that's what's going on here. It's almost as if like you bought Nvidia and then you take all the chips, not making less chips, going to make more chips, but nobody else gets to buy them. So obviously it's impossible to do that with Nvidia. There's so many of them, but you could imagine with grok or an inference chip, you know, a smaller, a sub $10 billion chip. Somebody might say, yeah, I'll buy grok for 10 billion and every single chip I'm going to take for myself. And, uh, yeah, if we have any leftover, maybe we'll sell them. So that, that could be the situation here. And I suspect it is, is that Zuckerberg is now taking a meta glasses, you know, metaverse approach to AI, which is I've got the most money and chips outside of Apple, maybe in Google. Therefore I'm playing for legacy. I'm pushing all my chips in. He is, you can say whatever you want about, um, you know, what he's done to society and you know, how he treats customers and the negative impacts of social media, which are vast. And we've talked about here countless times on a business basis. He is a complete berserker, like in the category of fighters, he will copy everybody's innovation and laugh while doing it and do it better than you, because he doesn't have to come up with the original idea. He doesn't have to come up with the three ideas that don't work to get to the one original idea. He'll just copy whatever your innovations are so fast that your head will spin. And then he has all that free time. Since you spend all the time doing the product discovery and product market fit work, you can just keep

SPEAKER_19: polishing your innovation. Then if there is a possibility that there is a disruptive technology

SPEAKER_20: and he doesn't, he made a mistake meta and VR did not disrupt anything about social media. Zero

David Friedberg: had zero impact, none. So he made a colossal mistake and burned $50 billion. It doesn't matter.

SPEAKER_20: That would be like me, you know, betting $50,000 or you betting 5,000. It's like, okay. Yeah, it stings, but who cares onto the next, right? You can make another $5,000 bet. I can make another $50,000 bet. He can make another $50 billion bet. He's playing with a different chip stack.

SPEAKER_11: The end, you know, and I was playing low stakes poker the last two nights. I just, my VP was a hundred

SPEAKER_20: percent. I played every hand and I told everybody I'm playing every hand. I won three pops last night, three, six off, eight, four off, King seven off. In eight, four, I, I got two pair on the turn after everybody was racing into an ace on the thing. The other two, I hit trips as well. Every, nobody could ever imagine that somebody would be playing a three for a hundred, uh, for a $50 pre-flop or a hundred dollars pre-flop who plays three, six, who plays eight, four. It's just not in the thing. I, you know, my, my friend, Alan Keating, who I play with or Stanley Tang from door dash, you know, they have adopted this style because they have very big chip stacks. They do that in high stakes poker when you watch that Triton and stuff like that. So this is an analogy that's important. If you have a big chip stack, you're a founder and you want to

SPEAKER_153: really be disruptive. Um, go all in and play a lot of hands. You're playing a lot of hands and you're putting max pressure on the industry. I, I, I think Facebook's going to do exactly an AI, what they did in social media, which is find their WhatsApp, find their Instagram and go all in. And that's, this might be the first of a couple of acquisitions that are going to happen.

SPEAKER_79: Well, I'll be, I'll be very curious to see who they buy next as they pursue the Texas Mike pre-flop

SPEAKER_00: strategy. And if you don't get the poker references, well, that probably means you're going outside too much and touching way too much grass. So get back to your computer. All right. There was a really interesting story about a startup today. It's called traversal and it is building a AI, basically an AI agent that does site reliability engineering work or SRE work. Essentially it looks at all your systems and helps you ensure that your company's website doesn't go down. Your service doesn't go down. It's a good use for AI. And they just came out and they raised $48 million. All this is to say that they were in stealth for a long time. So long, they raised their seed and a while in stealth and only are now coming out, talk to the public about this. I'm really curious about the strategy. Cause I feel like going in stealth is a very 2012 thing. So I'm curious from your perspective today, one is being in stealth still something startups should consider. And if so, when, cause I was just kind of blown

SPEAKER_68: away by this seeming blast from the past. Yeah. Okay. So I'm going to have a weird take on this. Okay. Weird take on this. Different entrepreneurs, um, have different energies. There are some,

SPEAKER_20: I, I saw Adam from, um, core. I've always really respected him and I've always tried to talk to him. Cause I, I just think he's cool. Um, but he's incredibly quiet and he just stealthily operates in the world and Cora just grows. And then I was really interested in what Cora was doing. And I said, Hey, you know, Hey, what's going on? You guys doing like X, Y, and Z. I want to say what it was. And he said, that's exactly what we're doing. We're doing X, Y, and Z. I'm like, I never hear about it. He's like, yeah, no explanation, nothing. It just, he's quietly moving in the world, like an assassin doing assassin like things. Some people that's how they're built. They're not

SPEAKER_153: looking to get a bunch of kudos from the world when they, you know, wake up in the morning and they show up for work. Other folks like these building public folks, um, they release a new feature. They get their, you know, daily Stripe report or their Shopify report and it's public. And you know, the buffer app guy was like one of the pioneers of this. He was showing his like actual, he made a public version of his like Stripe or whatever. And you could actually see his revenue. You could see his churn rate. Now all the statistics were great. So it was good. And that made his team excited. It created excitement around his startup. People checked it out. It also inspired. I kid you not, there must be a hundred startups right now doing cross

SPEAKER_38: posting, uh, services because it it's all balkanized in social media and management because cross posting and management, there were like a small number of companies that did and it cost

SPEAKER_20: thousands of dollars. Now it's basically, uh, free, you know, like close to free a hundred bucks a month, 10 bucks a month instead of thousands of bucks a month. So yeah, there you go. And it looks like you had a bunch of churn during, uh, when, when did you have that chart? Was that, is this a yearly chart?

SPEAKER_220: This is, uh, this is all time for them going back to January of 12th. And so it seems like right

SPEAKER_79: around two, right around COVID. Yeah. They had a pretty big decline, but they've since, and I knew this, so going in, I kinda knew what the chart was gonna say. They're now at, I believe at all time ARR high again. Right. Um, but what's cool Jason here is they didn't stop sharing it when they were here. Right. That's the thing that gives me a lot of belief. When it went down, they didn't do it.

SPEAKER_225: Yeah. So it's like, it's like a fun thing to watch along, you know, the challenge with this is he has employees. So when an employee says, I need a raise, I got these other job offers. And he's like, you can't cry poverty. He's gotta have like a really significant discussion and say, you know

SPEAKER_226: what? You're right. You're worth it. Or, you know what? I got two other people here who are, you know, half your price. So I can get two for one. So I'm just gonna stick with them. Or, you know what? I automated your job. And, you know, so, but this reality has always existed. So in some ways, your job becomes pure. Uh, you can say to people, you, you can't obscurify the reality and then make that your tool for doing compensation negotiation. You could, you then have to focus on reality. You're making X market is X minus 20%. I like working with you. I'm willing to pay the extra 20% because you're awesome at your job.

SPEAKER_20: Or, you know, the market is now X minus 50% for your salary. And I'm gonna make a change because it's in the best interest of the business. So you and I had this discussion about Microsoft cutting 3% of their employees in a record quarter with record cash in a booming market where their

SPEAKER_153: prospects are incredible. Like what? This is against all Silicon Valley tradition. You always had extra staff. You cherish the people who stayed with you for a long time. You wrote them letters. You celebrated their 10 year, five year, 20 year anniversaries. And now it's just like, yep, we're in a different,

SPEAKER_20: it's a different world. Now folks, you're gonna get cut. It's a sports team world. There's 15 seats. And as Andy, Jesse said, like, we're gonna have less people doing more. So use AI every day,

SPEAKER_225: but best, best advice. And white collar work is changing. You know, like if you're good on podcasts,

SPEAKER_142: that's one thing, but you know, if you're a developer or a sales person or customer support, you know, it's just metrics. And can we automate your job lay, you know, I think this is why every

SPEAKER_79: CEO is like, please learn AI tools now. Cause what they're saying is please don't make yourself the

SPEAKER_00: first one that gets cut, but looping back to stealth. Some founders are quieter, moving in silence, like G's and lasagna, I think as Lil Wayne said, and some are more chest pounding. So essentially,

SPEAKER_79: then does stealth just come down to, does it fit the founders operating ethos?

SPEAKER_20: Yes. If you're a founder and it gives you more energy to build in public, go for it. I don't

SPEAKER_153: give your entire roadmap. Don't talk about things you haven't done, but it does create momentum and

SPEAKER_20: joy. And, um, you know, with your customer base, they can get more engaged, be in a community, but again, just be aware that everything you're doing, your competitors are copying and in a world where software can be built pretty fast. If your buffer, like that buffer chart, if the, I wonder if

SPEAKER_153: that buffer chart would be, he'd be at a hundred million in reverend right now, if he didn't inspire

SPEAKER_20: 10 competitors and competitors are constant, it creates downward pressure. You know, people know

SPEAKER_235: you have a big business. I'm going to go compete with it. So if you really think you found a really

SPEAKER_148: great business, maybe shut up. Don't just collect the money. Do you remember when Google went public and everyone collectively crapped their pants because they're like, they're making how much money?

SPEAKER_00: Now no one caught Google, but like that was, uh, you know, I saw on saying, please come here and take our business. Cause we're just printing money. It was crazy. You want to inspire your customers,

SPEAKER_20: your team building public yourself, get stoked. You want to inspire competitors, ghost health, careful out there folks. The both, both I've seen both strategies work. I think over time people maybe zip it and they just do their two yearly keynotes and get into a cadence for product releases.

David Friedberg: That's more like that. Uh, but we have some high performers and I think it puts pressure on them.

SPEAKER_153: Uh, but that since every company does go through an ass kicking at some point, you know, and this is what being public is being public is if you're an IP, if you're a public company, you are building in public. It happens to be a quarterly cadence, not a weekly or monthly. So in some ways, those folks are prepared for it. In fact, I think Facebook and Uber, Stripe, some other people started doing quarterly, releasing quarterly data, even as a private company, um, or maybe not being as protective of the quarterly data. Like if it leaks, it leaks, like we're just getting prepared to

SPEAKER_00: be out there as a public company. So if you've raised a series G, that means gosh, darn, just tell me every quarter what you're doing. Even if you don't put it through an SEC filing. Uh, we do have an office hours today, Jason, but I want to squeeze in one more thing because

SPEAKER_07: TikTok is the popular vertical short form video application. It is the brainchild of ByteDance, a major Chinese private technology company. Never heard of it. Never heard of ByteDance.

SPEAKER_247: Oh man. We know it. We know it. We know it.

SPEAKER_00: Uh, Congress passed a divest or get out law earlier this year. President Trump came into office. He has delayed that not once, not twice, but now three times giving another 90 days on this deal. Uh, my understanding Jason is that TikTok and its future in the United States has become wrapped up inside of trade negotiations. So it appears to be kind of a, a political football going back and forth,

SPEAKER_250: but I'm curious, uh, why this is so top of mind for you.

SPEAKER_20: We very rarely have consensus amongst all politicians on any item. And this is one other than spending money like drunken sailors. And this was one where they all agreed like this is a threat to America. Um, and there are a bunch of people who are saying like, you're hysterical. It's not a threat. Um, and very clearly is they've been caught multiple times, stealing consumer data, tracking journalists. The reports from inside of it is that, you know, they're slow rolling. Um, you know, the data rules that they've been given and, um, you know, the Chinese government essentially owns every company in China and they have the ability to have a golden share and be on the board. Putting aside why Trump

SPEAKER_37: has extended it now three days, Jeff Yass making big donations, all of that, you know, you can have all your criticism that this administration and the last were pay to play. If you make a donation,

SPEAKER_153: you get what you want. Uh, politics has always, unfortunately the U S had that, you know, um, aspect to what we do. And I think it should change, but Hey, here we are. I don't know if that's going to

SPEAKER_20: change. What's interesting about this is the fact that it's become part of the overall trade deal. What that should tell you is that this is a foreign asset. The fact that China will make this a key part of trade is that it's important to them to spy on us. I am 100% certain with zero evidence, but just my gut and number of years on the planet. I believe the Chinese government is using that software to read the phones of children of important people to hack into people's phones, to hack into people's data, to know their location and to feed that into their version of Palantir. And we are idiots for allowing them to do so in India. They just banned like on mass a dozen

SPEAKER_153: or two dozen of these Chinese apps. They were like, yeah, it makes no sense that you can track our

SPEAKER_38: citizens at scale. No, get out. Just no, it's not personal, but reciprocity. We don't have it. So now that this has been put into that, just pause everybody, you know, first principles thinking,

SPEAKER_153: why won't the Chinese spin this out and let it go? Why are they holding onto this? Like it's a, it's like they're nukes because it's more powerful than nuclear bombs. This is more powerful than

SPEAKER_38: nuclear bombs. They could get compromise on Trump, you know, on Nancy Pelosi. They could have hacked AOC's phone. They could have hacked all of their children's and grandchildren's phone and have compromise on them, know their locations. They can know the locations of special forces. And if you're wondering if that's at all possible or in the realm of possibility, the Israelis just did it in Iran where they were tracking people because of their WhatsApp app. It's an app. WhatsApp is an app. They were tracking, they've hacked those people's phones and they literally sent drones in to kill them. This is not science fiction. This is the reality of what this technology enables. The Chinese could come in and assassinate a hundred people deftly using TikTok data. I am certain.

SPEAKER_56: I recently joined TikTok again, because my boss asked me to, um, but what I'm really, what I really care is about.

SPEAKER_261: Law is a jerk. God, Lon.

SPEAKER_00: Oh, I wasn't, I wasn't talking about Lon. Um, what's, what's interesting to me is how much perspectives here have changed. And again, not trying to get political, but the bill in question passed the house by a vote of three 60 to 58. It passed the Senate 79, 18, which is an enormous majority for the American Congress. And it did seem like the technology industry was somewhat unified

SPEAKER_263: on the points that you're taking. Now it feels much more mealy mouth, much more, well, free market.

SPEAKER_00: Is the shift just people's political allegiances coming into the white house and therefore them following the leader here? Or is there a counter argument that we should take into account as we

SPEAKER_04: consider the next 90 day pause in this, uh, span?

SPEAKER_20: I mean, listen, it's all a raw shock test. Now if Trump's your guy and you voted for him, like the majority of the country did, you think this is 4d chess and he's awesome at it. So therefore part of the deal, he's going to figure it all out. If you are the minority of the country that, you know, didn't vote for him, uh, and, uh, you don't like him and think he's a dictator and that there's too much grifting going on, yada, yada. You'd think that this is all some way for him to enrich himself personally. Uh, and so the truth is probably the people who are major shareholders in TikTok have his ear and they can make very convincing arguments. He wants to do this grand reconciliation with them. I was talking with three, last night I had

SPEAKER_153: a conversation with three of the top Chinese entrepreneurs in the world, the top, top, top of the top. And we had a discussion about how the United States and China, the people of both these countries truly want to work together and we need to turn it around. So literally at this little conference and, you know, commerce, you know, we have two governments that are seemingly always at each other's throats. And somebody said, you know, this is all about Taiwan. As I said, what is this about? It's pretty much about Taiwan, a little bit about the trade stuff and making money from each other. And there are trust issues and whatever. And I said, what's the solution for Taiwan? I kind of liked the ambiguity, what we had before, like Taiwan's an amazing country and these people are amazing and they're a great part of China. It's like, wait, that's two different things. Are they a country or are they part of China? It's like, yes, yes to which one. I was just dodging bullets here. Yeah. They're an incredible part of China and they're very proud people and have a great country. The end, you know, like that, that worked so well. And a tolerance for ambiguity is just so important. When you're trying to get along with people, right? Not everybody's perfect or fits into your mold. Um, somebody said, why don't they just give a hundred year guarantee? Like the Hong Kong bang in a hundred years, Taiwan will be part of China. Like in a hundred years, like maybe that's the best way to avoid world war three. It's just to come up with like a,

SPEAKER_20: you know, a hundred year timetable, like kick it out. Like, and you know, Hong Kong is part of China now and Hong Kong is still independent and has its own sovereignty and like 80% of what they do. And okay. It seems to have worked out better than a war for Hong Kong.

SPEAKER_122: All right, Jason, we need to go talk to a founder. It's about time we did that.

SPEAKER_00: And today we are going to talk to Risha Gupta from Actuality. You may know him from the ninth Founder University cohort and the 34th Launch Accelerator Group. He is the man behind Actuality. I think it's a really cool company. I was just looking at it earlier today. Welcome to the show.

SPEAKER_226: Good to see you again. And so you went through our Founder University program in the ninth cohort and launched Accelerator 34, which is going on right now. You're in week eight, seven or eight. Yes. Yeah. So tell everybody in the audience or show us what you're building and why it's important.

SPEAKER_280: Absolutely. So my name is Risha and I'm co-founder and CEO of Actuality. At Actuality,

SPEAKER_281: we are building an RFP response platform specifically for the construction industry. We use AI to automate RFPs. Essentially, the advice is important. Most construction companies have to respond to a bunch of RFPs to bid on projects and get sales. And responding to these RFPs is time-consuming and resource-intensive are often taking more than 200 hours, $12,000 to respond to just one RFP. And many times they don't respond to more than 50% of eligible RFPs out out there. So it's a great idea to streamline it so that they can respond to more RFPs and without increasing headcount. That's the essential idea.

SPEAKER_226: Yeah. It's an idea that's going to help people do more with less. They're going to be able to respond to more RFPs, which means they're going to get more work. The quality of the RFPs, the quality of the bids will be better. The win rate will be better and it'll be cheaper. So you've made something better,

SPEAKER_281: faster, cheaper. Yes, absolutely. Yes. And the customers are loving it. We started working on this in November of 2024. So since then, we have an onboarded six design partners. One of them started paying us and they're loving it. They're using it on a regular basis. One cool thing, recently they came back to us and they said their customers told them that this is one of the best responses response they have seen in many years. Because it was very personalized. What ends up happening is people use templates and they just fill in their data from previous RFPs. But using AI, everything gets very personalized for the actual bit they're responding to. So hallucinations, making sure you

SPEAKER_226: don't over promise in an RFP or make a mistake, this is critically important. So what is your instruction to your customers in terms of best practices and how, you know, you obviously could just start firing things off and flood people. And some people are doing stuff like this. I get pitched all the time on people who are like, we'll make a thousand, literal a thousand clips, you know, with AI from your podcast. We'll put, you know, 10 bucks behind each one. We'll tell you which ones work. We'll get a map. And I'm like, that sounds like spam. And they're like, no. So yeah, maybe. So what is the current state of affairs? You know, how much time do people have to put into polishing them, checking them? And what are your instructions to customers? And then the follow-up to that,

SPEAKER_20: obviously, is how is that progressing? If it was going to get better, at what pace would it be better

David Friedberg: and to the point at which there would be no changes needed or virtually no changes?

SPEAKER_281: Our responses, when we started working on this, initially, they were like, you know, 80-85% accurate. But over the past six months, they have got into a point where they're 90-95% accurate. The way we are selling this to our customers, we are saying that it is a co-pilot. Like, you know, it would create the first draft, you have to like, you know, it's your responsibility to go through it, check each and every item on this. Because like, from liability perspective, we want to shield ourselves. And this is a revenue generating document directly affecting the revenue. So you want to be mindful of that. One of our customers, like, actually, like, you know, all of our customers that came back to us and said, we, they're spending right now, they're spending approximately five to eight hours every time, like, reviewing the RFPs. And that's a lot. So, so we, we, we want to like, you know, closing the gap from 85 to 95 was, was like, okay, like, I would say easy, not easy, but like, you know, but 95 to 100 is going to be difficult. And it is difficult. So my question here, Jason is, I read a, read a, like, you know, article from Paul Graham saying that two things that don't scale. So I was wondering, like, you know, would it be a good idea to, to probably like have from someone from our team, at least like, you know, doing something for them included within the cost

SPEAKER_226: for each RFPs for now? 100%. Yeah. Human in the loop is a great idea. The product Grammarly, which I'm addicted to had a feature early on that they've deprecated where you could write in Grammarly and then you could press a button and have a human editor review it remote. And,

SPEAKER_20: you know, this was crowdsourcing or, you know, virtual war mechanical perk back in the day, we had a bunch of different ways of phrasing it, but, you know, a little reinforcement learning or human in the loop here would go a long way, I think. And I think they'd be happy to pay for it.

SPEAKER_226: You could make it a feature, not a bug, which is, would you like to pay for a human to review this draft? Or you could just absorb the cost yourself and consider that R and D and have that person, you know, training the AI and saying, you got this wrong. This could have been better. This was good. This is great. And just explaining to them what's good and great. And depending on the sophistication of your customer, you could let them know that you're doing that, not let them know. Um, I wouldn't, you know, if you're raising money and you present this to shareholders, uh, future shareholders, you do have to make sure that you're clear with them

SPEAKER_225: what AI does it and what humans do it. There was a company, they were, I think, having a bunch of people in India write code in the co-pilot. So people thought they were out of coding co-pilot, and it was actually just really brilliant Indian engineers. Now. Okay. Points for creativity,

SPEAKER_296: but yeah, not cool. Um, and it was builder.ai over in the UK that had 700 engineers that were fake AI

SPEAKER_298: and, uh, didn't go well in the end. Yeah. Now hiring those people to do reinforcement learning.

SPEAKER_226: Awesome. So, you know, you could, when I did Grammarly, it was, they were explicit about it because people know what a copywriter is. I'm curious, what is the person who responds to RFP

SPEAKER_280: called? Is that a specific title in the world? Uh, it's RFP, uh, RFP manager or RFP specialist.

SPEAKER_281: These are the two titles. Yeah. Which is the higher title? Manager, like, you know, uh, yeah, proposal manager is another one, like people call them. Uh, so what is the junior person called?

SPEAKER_153: Specialist. Okay, perfect. So I think, you know, actually your RFP specialist, your AI RFP specialist, you could present it as we are going to be the specialist, but you still have to manage it. And then you could hire a manager, uh, to be part of the stack. So we have managers available to polish this for you. Would you like that for $200 per unit? Are you charging people per seat, per unit, per co-pilot? How are you charging right now? What's the experimentation like?

SPEAKER_281: So, so we are charging based on per RFP. Um, I was from this industry, most platforms charge per seat, but I think like, you know, charging per RFP is a better way to do that. Uh, yeah, like, you know, we can, we can increase, we can say like extra 200 bucks will get you human reviewing it. So you would

SPEAKER_79: say eight hours. Richard, I'm really curious about the verticality of this product because I did not know

SPEAKER_00: about the architecture engineering and construction or AEC sector until I was literally prepping to have you on the show. So my question is, has the technology that you've built to handle RFPs in that particular industry, does it translate to other sectors or is there a lot of back-end work to get the right data pipelines in place that would make it hard to kind of copy and then paste this into a

SPEAKER_281: different vertical? There are a few companies which are already doing a good job of like, you know, servicing other industries, but there are not many companies which are doing a good job specifically for the AEC industry. And the reason for that is, uh, is that the tech stack for the construction folks, AEC folks is very different from like, you know, rest of the companies. Uh, they don't integrate with softwares like Procore. They don't have the ability to read architecture drawings, all these kinds of things. These are the, these are the things which we are building and we want to specialize in. And that is why we have chosen this vertical rather than like, you know, going horizontal to, to, to service many, many industries. The other big difference is, uh, the way this whole industry works is many, many, uh, people within the sector, like an architect would form a syndicate with a general contractor to collaborate together to bid on a bigger RFP. Uh, so that type of collaboration is not allowed in many of the other tools. So we are, we are building those subcontractor collaboration workflows, uh, which is very unique to us. Oh, well, that's awesome. So the answer is

SPEAKER_00: not particularly copy pasteable, but also there's good alternatives elsewhere, but you're solving a harder problem and that'll form a nice moat around the company. Exactly. And that's, that's

SPEAKER_308: exactly, uh, the idea. Well, now I know why Jason invested in you. That sounds, that sounds great.

SPEAKER_281: 10 points. Yeah. So I used to respond to a bunch of RFPs. Uh, I've responded to more than a hundred RFPs within the AC sector. So, and we tried using all some of the existing tools, but none of them works that that wanted me to like, you know, work on this. Got it. This is going to be a great

SPEAKER_249: business. Is there any way I can be more helpful? Uh, anything you're struggling with? Obviously a

SPEAKER_226: lot of, uh, firms in your stage or in your accelerator, looking to fundraise, I'm sure you're going to crush it in fundraising. Um, and I don't know if we've introduced you to VCs, but, uh, I'm curious how the fundraising is going, or if you're considering a fundraise. Um, and then also on

SPEAKER_313: that, I'm curious, uh, if you have any questions for me or things I can be helpful with.

SPEAKER_281: Yeah. So fundraising is going well. We have already started forming good relationships with a bunch of VCs through your, like, you know, network chase and thank you so much for that. Um, and, and we are like holding off a little bit. We, we, we have, we have started building relationships. There, there are a couple of customers, couple of design partners who are about to become paid. So once that happens in the next couple of months, that's when we like, you know, uh, go, we'll go full throttle and, and, and fundraise. That's the idea. Um, on the other side, like, you know, from, from, we, we do, we are, we are facing a little bit of a dilemma, uh, regarding product side of things. So some of our costs between like how to prioritize, I would love some help, uh, how to prioritize between, um, data privacy, uh, building new features or like increasing the accuracy of the responses. This was my question for you, but we already talked about accuracy. Uh, but like, you know, uh, how do we, how do we,

SPEAKER_226: how do we prioritize that? So your customers care about the proprietariness of what they're doing and they really want privacy. They bring it up. Yes. They bring it up. Everyone does it

SPEAKER_281: some more than others. So we are working with a really big multi-billion dollar company. For example, they do 6,000 RFPs a year. Uh, so on the, on the last call, they said, is there a way that we can deploy it on prem? So, yeah. Okay. So, yeah. So this is, I think,

SPEAKER_226: the enterprise version is how most people do this and it's contact us for pricing for enterprise. And it just is at a different rate. It's much bigger rate. It's at a six figure per year rate. And so you should really explain like how much extra work that is. And you know, you're giving up the reinforcement learning, but we understand that's key. There are people, you know, I now pay, uh, because I needed to have security for, you know, as a finance firm, you know, so now we can, for Google docs, for Slack, we pay for higher versions of this, specifically to have access to APIs or specifically to have like backups of everything or monitoring of everything. If you like have the standard Gmail as an employer, you can't search through all the emails in your organization, but you can like, I guess, change somebody's password and then go individually and search in it. You work at a bank and you need to know if somebody is talking to this person or there was fraud going on, or somebody did something inappropriate. You have to search all emails for compliance reasons. That's like, that's not $8 a person, $50 a year. Now you're at, you know, $300 a year. So look at the pricing Slack and, uh, Google docs charge for it. It's like super enterprise. And you can probably find that by talking to an LL about like, what, what is the pricing and what are the features, but those, and the way to sort of back into it is look at the compliance industry and finance. Cause they have the most acute, the government forces them to keep all communications. Like,

SPEAKER_20: because if you were to make banks, Goldman Sachs, trading desks, they need to have every phone call recorded, every trade, every trade recorded to make sure they didn't like sell somebody something they didn't. And if somebody says, I didn't buy that, you know, when they're on the trading desk and everybody's ordering, they have a recording of it. And they're like, well, here's like a playback the recording for you. Like all records are all phone calls on the trading desk are recorded every tax to make sure they're not doing insider trading, all that compliance work. So compliance

SPEAKER_216: is a way for you to understand how to charge because it's very similar to that. And I think

SPEAKER_153: it will be analogous in our industry is people are willing to pay. I'm going to think it's four times as much, three times as much. I don't know what the multiplier is, but I think it'll be four times as much to have that. And then on-prem might be, you know, 10 times as much if you want that privilege of we're going to run our software in the cloud for you and this thing. Again, as an

SPEAKER_226: entrepreneur, you only have so much work you can do. So the version that's cheap, we get the reinforcement learning. The version where we don't get the reinforcement learning, you know, and we don't even see what you're doing, that could be a different price.

SPEAKER_323: Okay. And 10 times more that I would have never imagined, but like, no, thank you. Thank you for

SPEAKER_326: telling me that. Great job. Uh, and I can't wait to see you in person. I'll see you at the graduation,

SPEAKER_226: uh, demo days, barbecue, everything. All right, everybody, this has been... Thank you so much, Jason. Thank you so much, Alex. Yeah. Thanks for coming on. This has been another amazing episode of This Week in Startups. Do us a favor, subscribe, rate, comment over at Spotify, Spotify, iTunes, everything. Uh, just let us know you're watching. He's at Alex on the Twitter, x.com slash Alex. I'm x.com slash Jason. And we'll see you all next time on This Week in Startups.