This Week in Startups - AI-generated South Park, “Holy Grails” of AI & more with Sunny Madra | E1785

Episode Date: August 1, 2023

This Week in Startups is brought to you by… Embroker. The Embroker Startup Insurance Program helps startups secure the most important types of insurance at a lower cost and with less hassle. Save up to 20% off of traditional insurance today at Embroker.com/twist. While you’re there, get an extra 10% off using offer code TWIST. OpenPhone. Create business phone numbers for you and your team that work through an app on your smartphone or desktop. TWiST listeners can get an extra 20% off any plan for your first 6 months at openphone.com/twist Roots. Invest in the only real estate investment trust that creates wealth for you and its residents at investwithroots.com/twist * Today’s show: Sunny Madra joins Jason to discuss Fable Simulation's new AI showrunner (5:51), AI-powered note-takers and CRMs( 37:25), and more! * Time stamps: (00:00) Sunny Madra joins Jason (2:27) Using AI to analyze large datasets (5:51) The WGA strike and Fable Studios’ AI-generated South Park episode (12:29) Embroker - Use code TWIST to get an extra 10% off insurance at https://Embroker.com/twist (13:45) AI-generated South Park continued (16:58) What is lost with AI and the Modern Seinfeld Twitter account (23:40) OpenPhone - Get 20% off your first six months at https://openphone.com/twist (25:10) The pace of innovation with RunwayML (31:06) Production studios sharing the wealth & alternative plans (36:00) Roots - Head to investwithroots.com/TWIST to sign up and start investing today! (37:25) Sunny showcases Folk app’s CRM (45:29) Sunny demos Circleback (51:32) Using Claude to break down an investor presentation and Sergey Brin back at Google * Follow Sunny: https://twitter.com/sundeep Check out Circleback: https://circleback.ai Check out Folk.app: https://www.folk.app Check out RunwayML: https://runwayml.com * Read LAUNCH Fund 4 Deal Memo: https://www.launch.co/four Apply for Funding: https://www.launch.co/apply Buy ANGEL: https://www.angelthebook.com Great recent interviews: Steve Huffman, Brian Chesky, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarland, PrayingForExits, Jenny Lefcourt Check out Jason’s suite of newsletters: https://substack.com/@calacanis * Follow Jason: Twitter: https://twitter.com/jason Instagram: https://www.instagram.com/jason LinkedIn: https://www.linkedin.com/in/jasoncalacanis * Follow TWiST: Substack: https://twistartups.substack.com Twitter: https://twitter.com/TWiStartups YouTube: https://www.youtube.com/thisweekin * Subscribe to the Founder University Podcast: https://www.founder.university/podcast

It's like going running alone or going running with a group of people. You run with a group of people. You run further. You run faster. You run better. You run alone. Two out of three people who run alone. And I'm in that group.

I'm just not going to perform at the highest level. When I'm in an office, I used to do 20 meetings at a day with founders. Yeah. 20. I would go for 12 hours, 20 to 30 minute meetings. That's crazy. No breaks. Eating at the table with a founder.

I don't know. When we did our accelerator, I probably do 30 meetings in a day. I would do three groups of 10. And I would give each one 15 minutes. So it's for an hour. And we were on a clock. And the clock was in the corner. I mean, YC does this too.

Everybody does this. But the point is, performance goes up in a group, in a room. This week in Startups is brought to you by and broker's startup insurance program helps startups secure the most important types of insurance at a lower cost and with less hassle. save up to 20% off of traditional insurance today at a broker.com slash twist. While you're there, get an extra 10% off using offer code twist. Open Phone brings your team's business calls, text, and contacts into one delightful app that works anywhere.

Get 20% off your first six months at openphone.com slash twist. And, Roots, invest in the only real estate investment trust that creates wealth for you and its residents at invest with roots.com slash twist. All right, everybody, welcome back to this week in startups. I'm back. I'm back in the United States. And so is Sunny, some deep madra, also back. We happened to be in Italy at the same time.

That was nice running into you. Just randomly. Yeah, it was great. Spend a couple days together, play a little bit of cards, saw some friends. But here we are. We're back. And our voices are a little shot. I don't know. I don't want to attribute it to anything, but nice to get a little time in Italy.

And, you know, on this week in startups, we'd like to do this weekly AI recap. We tend to do it on Monday or Tuesday. So consider this this week in AI with my good friend, Sandeep Madra. If you don't know his company, it's called Definitive Intelligence. They do AI data analytics, basically pointing all this incredible AI at your data sets to make you smarter and make better decisions. And do the work for you. And do the work for you. But also you're combining it with public data sets too.

Yes. Which is super fascinating. So you could have your internal data set of, let's say, all your sales internally, all your marketing internally. But then you could put it against, say, Amazon's reviews or Google Trends or the stock market or weather. And this is the holy grail that we never really got to. Why did we never get there, Sunny?

When we look at how data, we've been talking about big data for 20 years. We've been talking about the cloud for 20 years. We've been talking about data science for 30 or 40 years. You know, talking about in corporate America. Obviously, these disciplines have existed for even longer than that. But it's been front and center in corporate America. But it never really came together the way it is now. Why?

Well, I think it was like a few factors. and we've had a perfect storm. I think it was for a long time, companies were resistant to the cloud. And companies that were doing this type of work natively were cloud native, meaning for the longest time, Facebooks of the world, Google's, Amazon's,

despite having their own clouds, they were using sort of cloud technologies for a long time. And so it took a long time for enterprise to get comfortable with the move towards cloud. So that was the first shift that was needed. then the second ship that was needed was an ability to basically make it easier for a broader range of folks to access this data. And I think that's what's really kind of come together with AI. You know, if you had a lot of data, which many companies did and had it in a data lake that you spent millions or hundreds of millions of dollars putting together,

but usually all the requests from the CEO down get funneled into a team of maybe, five or, you know, 50 analysts. And those are trying to answer all the questions for every person in the business. And that wasn't scalable. And I think that was the choke point, if you were. That was the choke point, exactly. So you have this incredible data lake. You invest in it.

But only like there's a little tiny dock and like five people can sit on the dock and fish from this incredible data lake. Correct. Now everybody's out in the lake, wave runners, boats, people are swimming in the lake. Everybody can go drink from this incredible lake and then the insights happen. And that's really why this is a revolution. And the thing I saw this week that I thought was while we were on a little bit of a break there, having All-Star Summer.

I don't know if you knew this, but we dubbed this summer, All-Star Summer here on this week in startups. And man, we just had the CEO of MongoDB on, and Ryan from Qualtricks on. We'd be just having crazy CEO-level talent on the pod to keep everybody's summer. really, really hard. Even Darmesh from HubSpot came on. Thanks to all my friends coming on and supporting me during the All-Star Summer here. And if you're one of my All-Stars, reach out. Let's go. Let's do it. Let's do a quick conversation. Actually, you know who I'm having on tomorrow is the Woop CEO. I don't know if you know whoop. Everybody's into this Woop wristband.

Yeah, the band, yeah. Yeah. So I'm thinking about getting on that Woop because I'm, you know, I'm getting on the health kick. So Will will be on tomorrow. Amazing. But while we were gone, a couple of things were going down. Obviously, we still have the strike going on by the Writers Guild. And the thing I found very interesting about that was I have been getting tapped for some entertainment projects. And I find myself using ChatGPT to do my research, even to do some proposals. And it's wonderful.

Absolutely works great. There's no doubt in my mind that anybody who's a writer on a TV show can get benefit from using these tools. and I don't think it replaces people, but on the actor side, I do think it does replace people. And that was shown to me perfectly, crystallized for me so perfectly. Not with the Star Wars, you know, Princess Leia, Luke Skywalker cameos, trying to cross the Uncanny Valley, not even the new Harrison Ford. I don't know if you saw the, did you see the new Raiders? Yep. I mean, that Harrison Ford, I did not realize that was DH.

I thought that that was from when I first started oh that's like a previous movie or is it outtakes or something I think we even talked about it we talked about it right we saw like great when they took like a

we did it like a demo of someone taking it older Harrison Ford and making them younger right so we saw that yeah yeah so I mean the thing that has crystallized it for me is a little TV show called South Park and the reason this crystallized it for me

yes is because when South Park came out Trey and Matt Matt they when they created it used if you ever seen the first episode that they did I think in college

or show after college they were you know doing stop animation basically with cardboard cutouts so they used this lofi so lofi

stop motion would be the first thing that computers can just perfectly replicate obviously so sure enough somebody created an AI

generated showrunner, basically, but not only did they nail the animations, which is easy to do, they nailed characters, location, and plot. Now, I'm using the term nailed. Perhaps that is a bit of a moving target, but I would say enough to be fascinating. So explain to us exactly how this has manifested because this is the first multi-dimensional AI effort that I've seen for entertainment specific. Yeah. And J-Cal, for this one, let's pull Nick in because, you know. Producer Nick, yes.

Yeah, because producer Nick has really done a lot of good background work on this one. Okay, here we go. So, Nick, why don't you lead out on this one and then I'll chime in with you? Yes. Hey, everybody. Yeah, you have a luxury boy, too. Were you in Italy? No, I wasn't.

So, yeah, this South Park demo was done by a team called Fable Simulation. They've worked on VFX projects for Pixar and Oculus in the past. And right now, yeah, they're legit. This new model they built as a generative TV and showrunner agent, it's called Show One. Show One. You can write, animate, direct, create voices, and even edits. And what they did was they basically put a prompt in, which I can pull up the prompt right now. basically the prompt was,

researchers trained an AI robot pig to write infinite TV shows but it goes horribly wrong. Cartman is a big fan of AI generated TV show Westland Chronicles which is basically a stand-in for Westworld. The Writers Guild is on strike because of AI. And from that,

it created this basically 20-minute South Park full episode. And I would say, honestly, in terms of the plot, it's like 60 to 70% there. And just as a general like plot overview, what happens in the episode is Cartman, Kyle Stan and Kenny are all debating

this new show Westland Chronicles that's AI generated. It's right. Kyle and Stan think that it's wrong, that writers are now being generated by AI instead of- Just so meta. Yeah. Well, exactly. It's crazy.

And then so the show, and this is where it's like so, it's so nails like the Genesis and the joke writing in South Park. And South Park always does as well, which is great. So the show is being produced by a company called Bisney, obviously, standing for Disney. And in the episode, these Disney researchers are creating a, generative AI pig, which is basically you could think of the pig as Mickey Mouse. It's called Met Porker, which is funny enough, like a combination of Matt Stone and Trey Parker's names. Met Porker is the pig. And they create this like generative AI pig to, and they start training

it on like different racial stereotypes. And it goes horribly wrong and the pig turns racist. And then it becomes a big news story. But that is almost exactly what would happen on a South Park episode about generative AI, I think, right? I mean, I'm not like falling out of my seat like hilarious laughing. So I think your 60% sounds about right to me. It gave some good bones to an episode. It obviously has, as part of their model, my understanding is they put in every script, every location, every character that's ever been on.

And in other words, they took the Wikipedia or the Wukopedia or Wikia, whatever, you know, those vertical wikis are. South Parkopedia. South Parkopedia, if you, if you, if you, just South Parkinpedia and it knows these are characters and it knows these are locations and then it knows the themes. It knows they like to do laboratory experiments. It knows they like to send up corporations and Disney.

So the AI is going to get you 60, 70% of the way there. But what's very interesting about this is the script would normally be over here in language model land, right? Sunny. The images would be over here in stable diffusion land. Yeah. So what's happened? here is stable diffusion land and open AI land have been merged, which we knew it happened.

But here we are in month 10 of this revolution or month nine of this revolution. And somebody stitched those two things together, didn't they? It was bound to happen. And even Open AI, right, has Dolly. And so, you know, it's not doing full animation yet. And we'll talk about runway and some other things in a bit. But that intersection is already here now, right? And so you can assume that this type of functionality, especially with the ability to set like context windows, in this case, like, you know, the entire South Park training set is coming really, really quickly to anyone.

And that's going to be a big threat to the writers and this whole strike that's going on. Listen, I work with super early stage companies at launch, like literally year zero. They haven't even incorporated yet. And then we hit the Series A. People have thousands of dollars in MRR. And maybe they've only raised a couple of hundred thousand before that Series A. And they don't have their insurance set up. And in fact, we recently had a great startup that didn't have DNO.

And we had to really stop everything because they were having board meetings. They were making massive decisions. There were legal issues. And they didn't have the basic D&O insurance that protects directors and officers. So we send them right to Embroker. Embroker is business insurance built specifically for startup. A single application will help your startup get four quotes for four lines of coverage in 15 minutes. Think about that.

Four quotes, four lines, 15 minutes. And they're going to connect you with one of their expert brokers for unmatched service that goes beyond your policy. We use it at launch. It's easy, peasy, lemon squeasy. It's easy, breezy. What more do I need to tell you? I use it.

I love it. A lot of our startups use it. They love it. Try and broker today with the code twist. And you'll get 10% off their startup package in broker.com slash twist. That's E-M-B-R-O-K-E-R.com slash twist, and use the code twist for 10% off. Okay, let's get back to this amazing episode. And the truth is here, on the animation side, Nick, you don't need for South Park to send this to, I think they were doing it in South Korea.

Wasn't that the big animation capital fraud? I remember reading somewhere that the way they turned over the animation so quickly in South Park, because you'd be like, wait a second, this Wednesday's episode is about, you know, I'm getting indicted because of the video cameras, right? Yeah. And you'd be like, wait a second. Didn't that indictment just drop on Friday and you got it by Wednesday? So there was some compressed thing that they, that was part of the magic of South Park was how did they do it so fast? And here we have, they could be working on it on a Wednesday afternoon and drop it Wednesday night.

Yeah. So South Park could be commenting on the day's events. With a. With high fidelity in the animation. And then you need zero animators. So, I mean, if you're an animator for South Park. park right now. This is Code Red. They don't need you. Now, if you're an animator for, you know, Ashoka or Clone Wars or Pixar, they still need you. So this is the beginning of the episode,

where Cartman and Kyle, Stan and Kenny are talking together. Get this, Bisney is using AI to make the entire fifth season. Ha ha, that's right. Artificial intelligence, baby. What? That sounds like a terrible idea. How would AI even make a good story for Westland Chronicles? Oh, I'm sorry, Kyle. I didn't realize you were an expert in AI technology. How about you let the super smart machines do their job, and we just enjoy the freaking show? Whatever. You guys can keep watching your lame non-AI TV shows while I enjoy a thrilling new season of Westland Chronicles. Fine, you enjoy your AI generated season. We'll stick to shows written by actual humans. I would say the voices, because I forgot to bring that into play here. Uh, you know, doing AI generated, uh,

voices are perfect. I don't think I could tell the difference in a show. I think you can a little. The inflection is a little. The inflection is a bit of all. All right. But they could probably smooth that out. Oh, yeah.

That's just a matter of, I don't even think that's a matter of technology. There's probably just a little more work. More tuning, yeah. More tuning. So let's just say they got the voices. The dialogue is kind of lame. It's not funny. It's boring.

So it's halfway there. And the animation is obviously 90% there? Yeah. I mean, most of the are even looking at the screen anymore. Yeah. So pretty amazing.

Yeah, you're right. The animation is the closest. And I would say the general, like, the sting of the joke is probably the furthest thing from what they have. Right. But as we showed, I think two weeks ago, we were sitting here and we were putting memes up and it was explaining memes.

So I think if you did some reinforcement learning here and said, that's a funny South Park joke. and the South Park writers punched up the jokes in real time, this thing after having 500 jokes punched up, would have enough of a data set to punch up its own jokes. And joke, yeah, yeah. If that makes sense. So with reinforcement learning,

and I have to say this is inspiring, I think, as well, because you could look at this through the lens, Sunny, of what's lost. Okay, the animators are losing their jobs 100% they're done. The people who do the voices, they're losing their jobs. You're completely unnecessary to have Matt and Trey and whoever else is doing voices do the voiceovers. Now, you may still want to have the animators. You still may want to have people do their voices. I get it.

But I'm saying necessary. In the next year or two, completely not necessary. And then the writers are the most necessary, but do you need 10 or do you need five? think you probably can get rid of two or three and let the AI brainstorm and then next year it's fine. Okay, so people are going to lose their jobs or you don't need as many people. What are the things you gain? Well, here's the things you're going to gain. You're going to be able to do this in real time. You're going to be able to do it tonight based on today's news. You're going to be able to do 10 times as many episodes because the cost is going to be so much lower. So there's no

reason for them to go on hiatus. You could have a South Park every week. You could have South Park every night if you wanted to do that. Now that might dilute the brand. You've got to be thoughtful about it. Well, the other thing you could think about is, let's just say there's a historic show that had writers, actors, the whole kind of infrastructure required for a show. Do you remember the, well, I guess it's called the X account now? The X account, Modern Seinfeld.

Do you ever follow this one? Oh, yes. There is an account on Twitter where they take the characters from Seinfeld and they put them in modern day scenario. So a modern day scenario might be using open AI or flying in a V toll or cryptocurrency. So it's like when they do, they do this for Sopranos. There's a guy on TikTok who will do Christopher Maltesanti talking to Tony Soprano about NFTs. Trying to order an Uber or yeah. Yeah, got it.

So modern day situations for classic shows. And now you could actually do that. And now you could do that, right? And so what I think it does is it basically says for creators that have, you know, strong basis in content, there'll be an ability. Because I would, like, I love Seinfeld. I still watch Seinfeld. And I would love to watch some modern Seinfeld shows. Jerry is never going to do it, right? They're never going to come together. But I would pay for that. And I think everybody that was involved in that infrastructure should get paid. And so I do think, you know, there's going to be a real big opportunity. And what we might have. end up with is not like 10 seasons of a show. We may end up with one or two, let's call them human-powered seasons. And then beyond that, that's enough of a seed for people to continue

with it because they enjoy it. In other words, you know, like General Hospital in some of these soap operas, I mean, they've been going on for decades. Yeah. And if you take the General Hospital model where they produce, I don't know, a couple of hundred episodes a year, you could have the Sopranos come back. You could fill in all the episodes in between. You could just say, hey, here's the foundation. Here's the canon. Now we can do non-canon stuff and fill in in between.

Here's what was happening in New York. Here's what was happening in Miami. Just fill in. So that's a really incredible opportunity. It takes some show that people still love. I mean, you could go far back. You go back to MASH. You can go back to Cheers.

You go back to, you know, you know, whatever show. And so I like to look at the opportunity. And then I think what you said there is very interesting in terms of a new model for those actors. So those actors made all their money in the final seasons, right? Yeah. They negotiated together as a group.

They got a couple million dollars. Everybody makes their money on the last two or three seasons. Yeah. But what if, you know, some show like the Sopranos, with the less than nine seasons or something? Something like that. Yeah. It was like seven, but the last season was split into two, so eight, really.

Eight seasons. So you have seven, eight seasons. What if that entire group could vote or they had in their contracts, hey, for every additional season, you're going to get paid, even if you're not on it, a minimum of in today's dollars, 80% of what you got paid for being there. And now, if I'm an actor and you give me 80 cents on the dollar for something that I have to do zero work for, that seems pretty reasonable to me. In perpetuity?

Yeah. And then what does that mean for the network? Well, the network can say, well, we can keep putting out sopranos at 80 cents on the dollar. Pretty great. And even after, you know, Gandafini has passed away, you know, and a number of the other cast members, tragically. People can still enjoy this in the family. I mean, I think that's the very interesting opportunity. Now, this, there's another rub here.

So we've identified some really great opportunities here for everybody. And it could be win-win. And people could opt out of it too, right? You could be Jerry Seinfeld and say, you know what? I created this. Do not train. You're not allowed to train models off. That should be like the artist, at the artist and the producers and everything should

have to come to that agreement. But here's another opportunity. You could personalize this stuff. And so knowing what it knows, just like Cameo does personal stuff, I could watch a Seinfeld or you could watch a Seinfeld that could tackle some topics that you really like. I could watch a South Park that was funny and included things in my world or you could regionalize it. So you could have South Park, but it could be based on the politics of Brazil, or it could be, you know, have an Italian aesthetic to it, right? And they did regionalize shows,

right? You'll have shows like the office that came from the UK to the United States, but now you can have the office in this model for every single location around the world, you know, for every dialect. So, you know, there's just so many wonderful opportunities here. And as long as the studios are not greedy, pig pigs, they could share the wealth. Yeah. Or if I just wanted to learn about like an event, like a recent event, you could have a show created about it and learn about it that way, right? Many of the things that we would see on those shows where... Give us an example.

Well, let's just say, imagine like a Seinfeld episode where somehow Kramer's involved in the Trump indictment. Perfect. Yes. Incredible. Which indictment? I mean, I haven't refreshed Dredge Report in four hours. So it's probably another 50. Kramer was in that room with the box.

Of course. Yeah. Of course. No, he was, you know what? He was the guy who they said, go, you go erase the videotapes of the big boss, watch her to erase the videotapes. It's crazy. Are you still using your personal phone number for your startup in 20203?

Is your sales team? Really? Your ops people? Customer support? This is a huge. mistake. You have your precious customers, your clients, your partners in one of your team members, personal phones. And team members last, what? One year to five years on average. And then all that knowledge is gone forever. And maybe they go to a competitor. Well, Open phone will solve this

problem. Open phone has rethought every detail of what a modern business phone will look like. It's just magic. You download an app or you log in on your desktop and you're done. It's amazing. My sales team, my ops team, they use it every single day. And we've recently found that it's valuable for when we do conferences. We have a phone number for when we do something like Angel Summit. All of the communications goes through there. If anybody needs to call us, our VIPs, they call that. And then it can do like a round robin kind of thing where it rings multiple phones at once.

So different employees can field those calls and nobody gets to the fourth, fifth ring. It's the number one rated business phone on G2 for customer satisfaction. And that's really hard to do. If you're in the SaaS business, you know how hard it is to beat. competitors in the ratings game. Open phone is already affordable to you at a starting price of $13 per user per month. But Twist listeners, you always get a better deal. You get 20% off any plan for the first six months at openphone.com slash twist. And if you have an existing number with another service, open phone will port them over at no extra cost. Head to openphone.com slash twist to start your free

trial and get 20% off. All right. So now, building on this, you have another company you wanted to show. Well, it's just pace of innovation. We've got runway back up and we're just starting to, you know, we've seen some of these things before. So I'm just going to point out some very specific examples. So this is runwayml.com. Ramayml.com raised a good, nice check of money recently. They've been crushing it. What I'm showing here is a prompt that says aerial drone footage of a mountain range and it creates a little clip that has that.

what I'm going to go forward here on is something a little bit more interesting. And I think this is pretty cool. This is the example. So here, someone is trying to create like a storyboard. And J. Kel, you've probably done this way more than I have ever even come across. And the storyboard on the left-hand side is someone taking their iPhone and stacking up some books. But really what they want to have is those books represent a shot that is, of skyscrapers.

Yeah, it's a flyby. Yeah, it's a flyby. Like a drone shot. Yeah. And so I really think this is incredible coming back to, you know, we were talking about the animation and where it's coming to and how you can create flyovers and screens. And last week we demoed the, you were drawing your bulldogs. Yeah.

What do you think of this, Jacob? I mean, so it's taking a shot on the left and then it's enhancing it. and making it into what you want, which is kind of like somebody walking into the design studio, like George Lucas comes in and says, hey, I made some drawings, and here's what Darth Vader is going to look like, and they pull out a mood board,

or here's a picture of a samurai, here's a picture from sci-fi, and here's my drawing. And that is all part of the creative iteration process. And if you can make the process go faster or increase fidelity, you're going to have better outputs. It's like somebody coming into a room. There's this great video of a producer pitching Jay-Z on a bunch of different tracks,

and one of them is dirt off your shoulder. Yeah. You know, that famous bomb, pop, pop, bomb. Anyway, and you see Jay-Z kind of like reacting to the first three or four, you know, loops? And he's like, he's kind of digging it. But then he hears that one, the iconic one. And he just starts spitting, and he's just this. is it, you know, like you see the look on his face, like, whoa, this is going to be iconic.

Yeah. And that's the moment everybody's looking for, which is how do we take, and you're showing it right now with that turtle, you know, a human turning into a turtle. It's like, these are incredible. You know, if the actor can emote it or the director can emote it and then you can make it into an animation, I mean, I think some of the great animated films in history are, could be created today, you know, like The Little Mermaid, or. or something or Beauty and the Beast.

Yeah. It feels like that 60, 70s into the 80s genre of animation from Disney. Yeah. Just like South Park. All of those, the super friends when we would watch or He-Man, that fidelity is going to be available to consumers to make their own shows. And so I think this means creative people. You could have some kid, you know, some young girl in Afghanistan is going to come up

with the next South Park or the next iconic superhero. And that should be inspiring to everybody. Now, I know for some people, it's scary because, you know, they get to play superheroes on TV or they've got a lock on it or they've spent their careers. I look at, well, what is the opportunity for the next generation? The opportunity for the next generation is, I think, more people are going to be able to make the next iconic superhero, you know, rap song, logo, poetry, novel, and more power to them. I love it. Yeah, I think we're just on the verge of an explosion. of content creation.

Yes. And each iteration has made it easier and easier. Like, think back to the very first filmmakers. They were limited to the, maybe the few hundred people on the planet that had access to the camera and the film, right, Jekal? This was, I mean, look at what we're doing right now. Zoom and webcams and all of a sudden enable everybody to have a talk show.

In order to have a talk show previously, we had to beg some local news station like Oprah did. to get, you know, a time slot to get cameras, $150,000 in cameras, $10,000 in tapes every time you recorded an hour, you had to have 20 people on the set, and now the lighting and the makeup is all done by software right now. So, you know, I love the fact that more people, when you lower costs, you increase accessibility. Yeah. That's it.

It's that simple. And so I feel bad for, you know, actors and writers. who feel like maybe they're going to be losers in all of this, but I feel grateful a long tail of content creators who are going to be enabled by this. Yeah. And I think they're the seed, like just to give them some hope.

Like, I think you don't have to do 10 seasons or, you know, what we were talking about earlier. You can do two and you can put that out to the community and then they can create from it. And I think what I'll be amazing is we'll see a show that maybe was like a B plus or a B in popularity.

in its original run, but then it's derivative. Its AI derivative becomes like a top five smash hit. And we've had that happen before. You know, you've had things that had a cult following, you know, on air. They didn't make it to season three. I think there was like Firefly or some of these shows. And then all of a sudden people are like, wow, that was such a great show. Let's bring it back.

Let's give it another shot. Make a movie out of it, whatever. What else is going on in AI? All right. Now that we've got, we beat the entertainment. We're going to keep cruising. here. By the way, for the studios, you really just have to share the wealth and come up with

the system. That's fair. So the actors and the writers have a point here. Instead of the actors having to say no AI and you having to say, like, we're going to do AI, just come up with an economic model. It's very sad. And they had an economic, but they had residuals. Yeah.

So, like, I think this is on the streamers. I think this is literally the technology industry is causing this problem. I think Netflix is the key. bad actor. Netflix does not want to play residuals. Netflix just wants to buy it. They don't want to share the data. I put the entire industry's problems on Netflix's shoulders.

Maybe double-click into that, J-Cal. What's the reason for that? Netflix is a technology company. They basically want to pay everybody a flat rate, own their content forever. What the studios did for a long time was they were like, oh, well, we'll have you make this. And if we sell it again, we'll take Seinfeld and we'll sell it to Fox for reruns and we'll sell it in this region for reruns. And whatever money we make, we're going to chop it up, takes a little bit of accounting. Yeah, there's a little Fugasey accounting on the margins, but everybody's going to get residual checks.

Everybody's going to get some piece of the pot. What Netflix said was, you did Orange's the New Black. Those actors got paid a pittance and they've been pretty vocal about it. They got nothing. And Orange's the New Black was one of the seminal shows, right? That felt the Netflix's original IP catalog. Why did they get paid so little? Because the streamer said, listen, we're not reselling it.

We own it. We're going to put it on our service for a subscription fee. So give us a pass on this. We'll pay you more. Netflix lied. They said to everybody, or they eventually lied. They told everyone, we'll pay you more on the way in and then, but we're not going to do residuals because we're just going to have a content library.

We're going to subscribe it. It's a new model. What they should have said was, hey, we're going to put you into this new model, and you're going to get equity, some number of shows. shares in Netflix. And if you choose to sell them, you sell them. But as part of your deal, you're going to get, you know, 10,000 an episode in cash and 5,000 an episode in stock.

If you want to keep making money, don't sell your stock. You're going to have, you know, as one of the actors on the show, you're going to have 100,000 in Netflix shares. Keep it. We suggest you keep it. Or they could say, we're going to give you RSUs. So for the next 50 years, you're going to get 10 shares a year of Netflix. And the notion of syndication went away.

The notion of syndication was it. Netflix, Orange is the New Black is not available on HBO or whatever. So the whole industry now is going to a library that you subscribe. So how do you chop it up? What they should say is we're going to pay you for the show, the other possibilities, pay for the show. And then every time a show is watched on air, when it hits a million views per year, we're going to give, we're going to put $1,000 into this kitty to share it when it hits $10,000, another $10,000, whatever. And you'll just build it up by the number of views. So if something like Orange is the New Black or Ozarks is a perennial thing that people

keep watching, great, if not, great. And that's why you see HBO Max. You notice how they took shows down. Yeah. And everybody's upset. Like, why you take shows? That's because they have to pay residuals on those, is my understanding. So because they have to pay money, they're like, well, what's the point of having this

on air? Not enough people are watching it. It's costing us money. We need to save money. So they take all that stuff down. And they're like, why will you take that down? Yeah. It's like, well, because we have to pay back.

on it and we don't have the money. Wow. So there's a balance between these two things. I believe the right system, if you want to read a great book, something like an autobiography by Akira Kurosawa, one of my top three directors. And Kurosawa is worked for Toho Studios. And in Japan, you worked full-time for Toho Studios. So imagine if Spielberg and a great writer and the actors all worked for the studio full-time,

they had a full-time position. They made a million dollars a year, a half-million dollars. a year, whatever it is. And then they would send them to an onsen and like a, you know, for a retreat. When they had the retreats, they would write two or three screenplays. They would act them out. Then they would say, let's go with this movie. Let's do this one. We're going to do a samurai film. No, let's do this noir one. And the group would kind of like, almost like a little startup factory, would make two or three films a year. And you, I would work with you on the samurai film. Then I would work with you on high and love, the noir film. And then we do an experimental

film. And everybody was just full-time employed. Everybody was part of the system. And it wasn't this like begging for scraps nonsense. Which is what's killing people now. So I think if I'm Netflix, I just hire 100 best writers in Hollywood for a million dollars a year each or, you know, on average, 250K each. Give them some, 250KK. Like, look at them like developers. That would work amazing. Hey, everybody. Today I'm joined by Root CEO, Dan Dorfman. Dan, welcome to the show. Thanks for having me, Jason. everybody here in the audience. What is Roots and what makes it different than the other real estate investing platforms? I'm a complete neophyte. Roots is a reed with a little twist. Sorry,

I had to do it. We are the first real estate portfolio that we know of that builds wealth for both our investors and our residents. And we've created a unique win-win model that creates partners and not tenants. We built this model because I've spent my entire career in real estate investment, 13 years. And what you always hear from people is, hey, location, location, location. But location at the end of the day, doesn't actually pay your bills. And location doesn't let you know when there's a small leak that will create mold in the future. The people do. And the people who rent your properties are really the people who generate your profits. And when those people are your partners, it really creates this amazing scenario. And it's,

this model that's helped us grow our fund over 36% in the last two years. Fantastic. How can people learn more about this opportunity, both on the tenant side or on the investor side? Check us out at investwithroots.com backslash twist. All right. Everybody go to invest with roots, no spaces, no dashes.com slash twist to sign up and start investing today. Jacob, I didn't set this one up as a demo because it's a CRM. Oh, I love this. And I want to talk about it because they've done some really incredible things.

And as I was actually doing some prep work with Nick, he mentioned that you guys had done some heavy CRM work recently. And so I didn't want to create a whole fake CRM. Customer relationship management, if you're new to the game, is a way of putting all your contacts into a database to manage them. Those can be startups you're investing in. It could be customers at your restaurant. It could be people who buy your SaaS product.

Yes. So this is folk. app. Okay. Kind of a new CRM software. But, you know, they have some AI twists, which I thought were really, really cool. And so I just have their video up. And I'm just going to forward through a couple of sections that I find it really

interesting. And so, you know, you can do all the stuff about sucking in your contacts and all that. So what it's doing here is it's authenticating with your LinkedIn, authenticating with your Gmail. And then it's putting everybody on the left-hand navigation as investors, journalists, media, potential sales. So prospect, suppliers, your team, etc. So it's automatically putting people into what looks like an error table or a Google sheet or, you know, HubSpot or pick your CRM. Exactly. And then they have a Chrome extension. Sure. That as you're cruising around the

internet as well, if you come across someone's page, you hit this ad as potential SaaS investor, and it adds them to that CRM, which I think is pretty neat as well. We have a company lead IQ that has that exact functionality. Great. And I think those are a little bit table stakes. Like I'm not like, you know, we've seen. Those have existed for a long time. Yeah.

Where it starts to get interesting is as they go further down the process here, and they're going to start, and in this particular case, they want to do some follow-ups. And you can see here they want to write a personalized icebreaker to each person in their particular list. And what it's doing here. is it's created this magic field, which is the personalized fall, and it, you know, I've done it with AI, which is the personalized icebreaker to each person in that list.

Love it. Yeah. And so, and then, you know, obviously you go down this typical CRM workflow here to create your template, and it obviously integrates the email, mail merge type functionality that you're looking for. And I know, as I was showing this to Nick, Nick told me, that he spent about a week customizing a bunch of outbound emails. Yeah, when we invite people, like we go to the All-Star Summer, we make a target list,

and then we'll use mail-merged technology to send a first email, a follow-up email, a follow-up email, and a follow-up email. And so that, but you don't want to send it like, hey, Joe, join us, or hey, Darmash, join us for this. It's like, okay, let's be really thoughtful.

We got a thousand people we could invite. Who are the top? 100. Okay. Now let's write a very specific email to that person of why we want to have them a show. Hey, Darmesh. The last time you were on the show was 11 years ago, episode this. So that initial sentence that you should do. And so it's basically doing that for you. Who knows how good it is or the fidelity of it. But, you know, that is definitely the holy grail is to write that AI sentence. Yeah. And so obviously here you have all the different folks in the different stages of your pipeline and where they're at.

And so standard CRM. Standard CRM stuff. And so. Yeah. That's great. Yeah. And I think, you know, this is a great example of taking an existing tool and just how you add in AI and almost make it really seamless to the workflow that is challenging

to us every single day. And so I thought these guys did a really good job. I mean, this is the thing I talked about earlier with you, which was, hey, I want to be able to find startups that I'm unaware of and then send them a note, let them know JCal AI is operating. And I think this is like one of the holy grails of what is something we do every day?

Well, we try to network with people. Well, how can we more intelligently network without it being spammy? And so if you're an investor in companies, if I know I have 10 meeting slots next week, and then 12 and then 12, and then 15 and then 30, if I could set an AI loose on the world,

the baby GPT, to go and say, find me anything that's like these companies in our portfolio that have become wins, like Uber, like Robin Hood, whatever, DoorDash. These are companies that we know are winners.

Just go email people who have companies we think might win and pull it up on pitchbook, crunch base, you know, TechCrunch, New York Times, whatever, could be compelling. in the future. So I do think that everybody wants that.

And then it just has to not be spammy. And the thing I'm finding is fidelity. Yeah. I'm working on a bunch of AI projects. And even things like tagging a blog post I was talking about in our previous thing and having the right tags, even getting that fidelity to human, I keep hitting 60, 70, 80 percent. And one of the things I'm getting a little concerned about is, is this technology going to hit 60, 70, 80 percent? Because for me to trust it, I really needed to get to 100. Yeah.

And so I think that's where the industry might have been oversold a little bit. And I'm really focused on closing the gap between 60%, like we're talking about with the South Park stuff and 100%. Or 99% or the equivalent of humans, which might be 98, 97, 98, 99, so that a human can come out of the process. It reminds me of the CafeX coffee machine that we were investors in. Yeah. The reason I invests, I knew robots would have a profound impact on food. I looked at Zoom Pizza, I looked at the dishwasher robot, I looked at a froyo, I looked at a tea, I looked at a salad one, I looked at the Momentum Burgers one that I think Costa had invested in.

All of those, Sunny, would automate 30, 40, 50, 60, 70% of the process. And if you look at that, you're like, oh, wow, it can make 70% of french fries. You're like, great, but that doesn't make the difference that CafeX hit, which was this thing's making you you know, a dirty chai latte macha with foam, iced, warm with some cookies on the side with zero human intervention.

That last 20, 30% of self-driving or making a pizza is the hardest part. And I suspect for knowledge work, it also is the hardest part. But when you close it, right now those two machines in SFO, the two Kaffax machines, are the highest square footage

of revenue of anything in the United States in a Now, it's not as high as a jewelry store. Like, I think actually the highest per square foot is either Tesla or Tiffany. Well, Apple store is high, but I think it's Tiffany, Tesla, and Apple are the highest per square footage. Those cafex machines are such a small footprint. And the two of them are, I think, are going to hit, like, they're getting close to a million between the two of them. And it take one hour of human clean up a day to put more beans, put more milk, you know, take out the garbts.

take out the garbage kind of situation. So, man, I think that has to happen for the knowledge work, too, whether it's CRM or tagging something or making a script. Well, my comment there is anything that has to do with language, I feel like we're in the high 90s. And what we feel high 90s, huh? Oh, yeah. 60s, 70s.

Yeah. And we need to understand. So one of the things that we should do offline is look at your use case and understand where it's coming short because we use it in a lot of different places. anything that's to do with language, it can be definitely in the high 90s. Now, that requires a lot of tuning, but we'll get there. So let's continue with the demos, J-Cal. I know you love these.

Oh, I love your demos. I need demos. So what we did here was we added the circle to our conversation about Netflix there that we had, J-Cal. And basically, it gave us an overview. What's his product called? Circleback. Okay.

Yeah. Oh, look at this. It's giving, this is like Vinnie's demo. Exactly, exactly. Which is, hey,

we just have this conversation on Zoom. We're talking about this episode and it says Netflix compensation model. Jason Calacanus criticized Netflix low payment to actors and creators citing Orange is the New Black as an example. Calacanis proposed an equity-based payment deal

where actors and creators receive a certain number of Netflix shares. He suggested that residuals could be linked to shows viewership on the platform. That's unbelievably accurate. And producer Nick, this is what I've been begging for for the three or four years you're here is like, how do we get the solution? And it turns out it's right here with AI, really well done to the team at Circleback. And so they do, you know, they kind of summarize the conversation and then, you know, I say,

hey, we were demoing this at the end. What I really like is, JCal, it also drafts an email. to send to folks that weren't there. Thanks for your great insights in yesterday's meeting, where we discussed the pressure shift in compensation models within the streaming industry. We tackled some controversial topics, including criticism of Netflix approach to payments.

The impact, I mean, this is uncannity. You can do all the formal. Actual formal, you can change the tone. Exactly. It's got the action items that, you know, and look, we were just, we didn't know it was there. like I had Nick kind of added in the background when we were talking about that Netflix section and then the transcript and so look I think you know this continues to just show how fast the

space is evolving amazing yeah unbelievable and I think you know look this this is really good we should have any come back on right and talk about when he's back on with us like I mean this is going to change everything. And this is why in our investment firm, I started recording everything. So I said, I want all of our investment. Because you know, like Bridgewater, is that Ray Dalio's firm? Yeah, it is. Yeah, and he records everything, and then

they transcribe everything, and then they insult each other and raid each other and they rate each other, all this. Yeah. I mean, it sounds wonderful to me. Yeah. Yeah. Because I'm a lunatic. I'd love to have somebody sit there and be like, Jake Al, nine years ago, you said no to Twitter, and here's what you said

and just put it in my face and like, make me a better decision makers, so I make more money. Love Bridgewater Associates. Loonacy. I want that lunacy. And so I started saying, hey, every investment team meeting and every program's team meeting, we're now recording. And I told everybody, assume that we're going to get hacked at some point or Zoom gets hacked and our files get released. So don't ever insult a founder on a call or call them dumb or say they had a piece of spinach in their teeth. It's like just basic blocking and tackling. So make sure young people don't do something stupid, like talk out of school. because people do do that sometimes.

It's a mistake. Don't ever do it. So I assume it's going to get leaked. But what's really interesting is now we've had this wonderful thing where we're transcribing it, bullet points. And I'm like, man, I can't wait till next year. When I can point AI at this and say, tell me which of those companies we passed on, who was in the we should invest and who was in the not invest.

Yeah. And then let's have a, you know, let's see if we can get better. be better decision makers how on earth would you ever coach a venture capital to be better at what they do you know like nobody's recording this stuff it's all ephemeral everybody re-does their

view of what happened in the past you know course yeah would you guys start using this one j-cal like uh would you show you're sure yeah i mean right now we're using notions AI we've tested obviously chat chpt4 and I guess

Slack doesn't have any kind of API built into it, but we did take Claude. Yeah, Claude, yeah, yeah. So we just started playing with Claude. I mean, one of the things that's happening right now is the AI companies are releasing so fast and they're leapfrogging each other that I feel like I'm buying a mobile phone 10 years ago. Yeah. When it was like, oh, my God, the Blackberry Bold has this. And then the Nokia has this.

And then the iPhone has this. And then the, you know, Android has this. And you're like, HDC has this. It feels like that moment in time where you just can't even keep up with the innovation. I don't have a demo, but we'll do it next week. Jacob, but we just added Strzok to our company Slack. Ah. Yeah.

And so we added Claude to our Slack. And so you can kind of ask Claude to join the discussion kind of situation. Is it the same kind of thing with Strz? Same kind of thing. Yeah, exactly. But what is Strzok's model based on? Do you know, is it Chatsy before or is it their own model? No, that is probably one of the bigger, larger models.

I don't think it's their own model. So, you know, the interesting thing here is, like, do I want my Slack being fed into a language model or even giving access? So this is where it gets a little dangerous because we're talking about proprietary in our Slack about people's companies. And so, you know, what if we're talking about Uber in year one? And then some other venture firm has this thing plugged in and they're asking about it. And then it spits out some Uber data that ingested, right? Yeah, I mean, Open AI, and we can, I'll pull up the tweet here. Open AI has been very, very explicit the last few months about if you have an enterprise

relationship with them, they do not train against your data. Nick, tell us about your 45 page investor presentation you dropped into Claude. Yeah, a couple weeks ago, we were going to cover it all in, but we never got to it. Sam Altman spacked a nuclear company that's doing any nuclear reactors all around the country. module. It was a 44, 45 page investor presentation for their stack. And I just dropped the PDF and Claude

to see if it could do it because I remember reading that Claude could, I think the big thing they were referencing. 100,000 tokens, yeah. Yeah, they could do the whole novel of the Great Gatsby the whole novel and explain it to you in a couple of bullet points, right? It can ingest everything. So I tried the investor doc and the investor presentation and it was perfect. Wow.

Explained it in bullet points perfectly. I could probably pull it up, actually. I have it in my history somewhere. No, so good. I mean, I think we get it. It was really impressive. Yeah. Apple also cut off employees access to chat GPT, which makes sense. So, yeah, I mean, there's a rumor now that Apple's going to get in on this.

It would be amazing if Siri actually worked in relation to like how well BARD and chat GPD are working. All right, everybody, this has been. And any other demos or anything we should get to or we'll put it up for next week? I think we do it next week. I got a couple more, but let's queue them up, right? Let's chew them up. All right, everybody. Oh, did you hear Sergei's back at Google too? Yeah, I heard multiple days a week.

He's in the building, multiple days a week. Brad had a, like, he does a yoga barn series. Yes, the yoga baron series. Yeah, and he had, Sergei showed up there to the last one. Yeah. Oh, whoa, that may be speaking at school there. No, I mean, Sergey's super smart. And, yeah, great 10-year hiatus.

And now he's back in, I mean, I think everybody wants to go back to work. I am getting this space in San Mateo and I'm getting my garage. And I'm done with remote work. I'm never going to hire. I'm never going to hire another remote worker I decided. I mean, I'll grandfather everybody in. I got great remote team members. But I think I'm just so over it.

I just want to have people in a room. There's just too much that's lost and all due respect to remote workers who are crushing it. I think one third of remote workers do a better job at home. and I think the other two thirds do less. Yeah. That's just my, I would say two out of three, don't do their best work. One out of three, do better work. You know what's happened?

As good or better. You know what's happened? What? With all the AI enhancement, I know you've been pushing your team. That has a, I was just going to put this out there. That has a further multiplier effect on those that are in the office. because they can go over each other's shoulders and walk people through what they're doing. Yes, exactly.

Versus when you're at home, you're kind of just using it as a, like, you know, replacing yourself and that's it and doing your own job. Right. Versus in the office, everyone's pushing each other to go to, you know. It's like going running alone or going running with a group of people. Yeah. You run with the group of people. You run further. You run faster.

You run alone. Two out of three people who run alone. and I'm in that group, I'm just not going to perform at the highest level. When I'm in an office, I used to do 20 meetings at a day with founders. Yeah. 20. I would go for 12 hours, 20 to 30 minute meetings. That's crazy.

No breaks. Eating at the table with a founder. No, when we did our accelerator, I would do 20 meetings in a row. Well, actually, no, we did more. I probably do 30 meetings in a day. I would do three groups of 10. Wow. And I would give each one 15 minutes.

So, you know, for an hour. And we were on a clock. Yeah. And the clock was in the corner. I mean, YC does this too. Everybody does this. But the point is, performance goes up in a group, in a room. Yeah.

Yeah. So I think it's, I think, and that's what I think is going to be the big rebound here is, I don't know if you're seeing like all these AI hackathons happening in San Francisco. Yeah. Like, everybody who's in AI is super motivated. Yeah. If you're a founder right now, I'm telling all my founders, move to San Mateo, move to Oakland, just get to San Francisco.

Yeah. Because right now funding is tight. Yeah. If you want a 10x your chance of being funded, come to the bed. Yeah. Get a tiny little apartment and rom in it and your chances will go way up. I saw your tweet today, though. You got $25,000 for anybody who's just...

25-000. No, I'm literally, you know, I got this launch fund four. It's going pretty well in terms of raising it. And, you know, we have this founder. university, which is a pre-accelerator is for people who are working on MVP. So I'm going to start doing it in person. Yeah. And, you know, we are on our six cohort.

We had a thousand people apply. We're accepting 250. And then of the 250, I'm giving 40 of them 25K each. So if you're, you got a one in five chance, basically, of me give you a 25k during the program on the way into the program. So if you show up with an, if you show up as a developer with a designer or two designers in a developer or a growth hacker in a developer, if it's any too smart. people, I'll give you 25K for 2.5% million dollar valuation. I'm your rich uncle,

rich aunt. Yeah. Call me Auntie JCal. Yeah. I'll give you the 25k to just incorporate. Because I'm meeting people. There's a lot of young people who don't even have the money to incorporate. They don't have the 25K to just even pay their lawyers or whatever. I don't care if you lose it because I'm going to do 400 in this next fund, launch fund for, I'm targeting, I'm interested in your feedback on this. I'm targeting 300, three to 400 of these 25K bets, 7.5 to 10 million. If it's a $100 million fund or a $50 million fund, then it will be either 7.5% to 10 to 20% of the fund, something like that range. But as a group of first bets, and then I have an option to put more money into the company.

So I asked, I'm like, hey, just save $2.50 for me in the next one, save me a slice of the next round, which is like an official term. And, you know, if I can hit one in 100, I'll, one in 50, whatever, you know, whatever hit winds up being. this would be an incredibly performing portfolio. So this is my kind of vision is to just get more people to start companies. I mean, it's like what Toby talks about with Shopify and e-commerce. You are doing. And I think the intersection with AI is going to be your real 100x J-Cal because it's much easier to start these.

And they fall much more into the flavor that you can help at in those phases as well. I think it's incredible. Well, you know, if your first stamp in your passport is J-Cal, like when you open your passport at the next stop, they're like, oh, you know J-Cal? I'm not being a narcissist here. I just know that I know everybody. Yeah. And if I don't know somebody, there's a 50% chance they know me. And you've seen a lot, which means you can help them because you can tell them, hey, I've seen it.

I'm just even making it even more basic. If you're a known quantity, it's like, oh, you know J-Cal, you know. great, what's he like? Or, oh yeah, I know Jake Al. We're in this company together. So when you get 400 investments like I have, you know, you're kind of a known quantity in decade number two. So I'm more than willing to put that stamp in your passport. You know, Jason at Calacanus.com for a life email me.

Listen, Sunny, you're amazing. Thanks so much for educating us every week. And I sent you an email while we're talking to meet with the team at inside.com. I want to show you what we built. Okay. Yeah, let's do it. All right. Thanks.