This Week in Startups - Best of CES recap with Sunny Madra | E1880

Episode Date: January 16, 2024

This Week in Startups is brought to you by… Vanta. Compliance and security shouldn't be a deal-breaker for startups to win new business. Vanta makes it easy for companies to get a SOC 2 report fast. TWiST listeners can get $1,000 off for a limited time at https://www.vanta.com/twist MEV. Tired of the dev shop rollercoaster? Mev is your reliable technical partner, offering a well-established software development process designed to consistently deliver unparalleled value to their clients. Get $30,000 off your first three months at http://www.mev.com/twist Imagine AI LIVE is an AI conference where you'll learn how to apply AI in YOUR business directly from the people who build and use these tools. It's taking place March 27th and 28th in Las Vegas, and TWiST listeners can get 20% off tickets at http://www.imagineai.live/twist * Today’s show: Sunny Madra joins Jason to explore CES 2024 highlights including the buzzworthy new device Rabbit R1 (3:57), Elli Q, a new AI companion for the elderly (23:44), the launch of the GPT store (49:22), and more! * Timestamps: (0:00) Sunny Madra joins Jason (2:11) Review of CES, highlighting AI's omnipresence. (3:57) The new device 'Rabbit' at CES, generating significant buzz! (10:26) Vanta - Get $1000 off your SOC 2 at https://www.vanta.com/twist (11:34) Evaluating the Rabbit R1 and purpose-built devices. (19:42) MEV - Get $30,000 off your first three months at http://www.mev.com/twist (21:04) The role of AI in integrating services via Large Language Models (LLMs). (23:44) Sunny demos Elli Q, a new AI companion for the elderly. (32:03) Imagine AI LIVE - Get 20% off tickets at http://www.imagineai.live/twist (33:37) Sunny showcases a new smartphone designed for kids. (49:22) The launch of the GPT store (43:23) Sunny checks out a GPT called “The Associate” and pitches a fun startup idea. (45:57) Exploring the brave new world and disruptive ability of the GPT store using All Trails as a case study. * Subscribe to This Week in Startups on Apple: https://rb.gy/v19fcp * Check out Definiteive Intelligernce: https://www.definitive.io/about * LINKS: Verge Article on CES and AI: https://www.theverge.com/2024/1/13/24035152/ces-generative-ai-hype-robots Bill Gates / Sam Altman Interview: https://www.youtube.com/watch?v=PkXELH6Y2lM Rabbit AI**:** https://www.rabbit.tech/ Elli-Q: https://elliq.com MMGuardian article: https://venturebeat.com/games/mmguardian-debuts-ai-smartphones-for-kids/ * Thanks to our partners: (10:26) Vanta - Get $1000 off your SOC 2 at https://www.vanta.com/twist (19:42) MEV - Get $30,000 off your first three months at http://www.mev.com/twist (32:03) Imagine AI LIVE - Get 20% off tickets at http://www.imagineai.live/twist * Follow at: X: https://twitter.com/sundeep https://twitter.com/Jason LinkedIn: https://www.linkedin.com/in/jasoncalacanis * Great 2023 interviews: Steve Huffman, Brian Chesky, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarland * Check out Jason’s suite of newsletters: https://substack.com/@calacanis * 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

We saw some big companies show up with advancements in AI. So let's start with the biggest announcement, I believe that got the most coverage this week. This is this new device called The Rabbit. You know, it's the size of like a, remember back in a day that T-Mobile sidekick thing? Yeah, like a handheld device that's smaller than a phone, maybe a little wider than a phone, but essentially it fits in the palm of your hand. Yes. Exactly.

This has a very iconic orange color with a screen on it. and it feels like a real hardware device, like a form factor, right? It's got a form factor to it that's very unique. Exactly. And it definitely got the most press and it got the most buzz. Now, I don't know why it got the most buzz, but let's get into it. This week in startups is brought to you by Vanta. Compliance and security shouldn't be a deal breaker for startups to win new business.

Vanta makes it easy for companies to get a SOC2 report fast. Twist listeners can get $1,000 off for a limited, time at vanta.com slash twist. Mev. Tired of the Devshop rollercoaster? Mev is your reliable technical partner offering a well-established software development process designed to consistently deliver unparalleled value to their clients. Get $30,000 off your first three months at mev.com slash twist.

And Imagine AI Live is an AI conference where you'll learn how to apply AI in your business directly from the people who build and use these tools. It's taking place March 27th and 28th in Las Vegas, and Twist listeners can get 20% off their tickets at ImagineAI. Live slash Twist. All right, everybody, welcome back. It's Madra Mondays. We're here on this week in startups with my guy, Sandeep Sunny Madra.

And we got a lot to demo, don't we today? We do. And 2024 is coming at us really, really fast. We, the breakneck pace, AI everywhere, and two big things last week. And we'll start with the first one, which is CES. And so before we get into it, like at a high level and there was a bunch of different articles that came out. And people were talking about at CES this year, AI was everywhere.

And, you know, just kind of pull out this one article. I think it was from the verge and saying there was almost too much AI. And I don't think that's the case. You know, we've been through these arcs before. we were there in the mobile revolution that kind of kicked off with the iPhone. And we've been there with cloud services. And now we're there again with AI. We're really seeing like huge advancements in a lot of different ways.

And let's just kind of get into some of the big ones. Yeah. For people who don't know, consumer electronic shows been going on forever. They typically just show gadgets. They went into a little diversion where they got into cars for a bit. I think the cars may or may not be back. So if you think about CES, it happens in Vegas, it happens in January, tens of thousands, if not hundreds of thousands of people now. I used to go, as a matter of course, because of the Engadgett blog we had started.

And then it kind of receded a bit, but it now seems to be coming back post-COVID. Yeah, there was big crowds and a lot of announcements of new products. And it's usually, though, hardware. Consumer electronic is the key to this, but software now is trumping the hardware, I guess, in many ways. It is. And one of the things that they're evolving to is as new technology comes out, they really kind of bring together the combination of hardware and software. And obviously, in its core way, but then they also bring like a great place for companies to launch. We saw a really good launch is there. And we saw some big companies show up with advancements in AI.

So let's just double click in. All right. Let's start with the biggest announcement, I believe that got the most coverage this week. And it's going to be worth a nice discussion here. This is this new device called The Rabbit. You know, it's the size of like a, remember back in a day? There's that T-Mobile sidekick thing. Yeah, like a handheld device that's smaller than a phone, maybe a little wider than a phone,

but essentially it fits in the palm of your hand. Yes. Exactly. And very simple, a camera that can kind of flip around and one scroll wheel. A scroll wheel that, yeah, you can flip up. up and down, kind of reminiscent of the round one that was on the eye pod, which then disappeared. So this feels, this has a very iconic orange color with a screen on it, and it feels like a real hardware device, like a form factor, right? It's got a form factor to it that's very unique.

Exactly. And it definitely got the most press and it got the most buzz. Now, I don't know why it got the most buzz, but let's get into it. It got the buzz for three distinct reasons. One, the company behind it, they not only released the device, but they came up with their own large language model. And what they did with their large language model is that they've kind of tuned and customized the model to be very good at interacting with your existing services. And so this device sort of sits with its language model on top of your Spotify account, on top of your Gmail account, on top of your Uber account. UberDash. DoorDash as well. And they brought all those examples in here.

And so what they really talk about is this is the first time that there's been a hardware device coupled with its own language model. So not building on top of any of the ones that we've seen and basically bringing that all together as like sort of one cohesive experience. So that's the first thing that's unique about this is it is a large language model. that they made that interfaces seamlessly with your favorite apps slash services.

Correct. Got it. Okay. And so by doing so, and you nailed it with your kind of food example. They have a lot of these examples we're not going to play their 45-minute video. But if you go here,

I'll just play this short one here. Get me a 12-inch pizza from Pizza Hut delivered to here. The most ordered option on the app is fine. Ordering a 12-inch pizza from Pizza Hut. And so you can basically understand. what happens here. And so what I really think what's interesting is this combination of custom hardware

with custom LLM that lives on top of all your services. And in many ways, what this does is it starts to take, I guess, an angle of attack to this training problem that we've been talking about the last couple of weeks with going after people's data. In this case, if I explicitly connected to all these services, then I'm not really running into that problem in the same way that we've seen it before. Got it. So the first thing is the large language model.

What are the second and third components of this that you think got at the buzz? I think the form factor. Okay. That's number two, clearly. Is unique. Okay. And then I think, you know, sort of the price point, you know, it's $200. I think they really kind of came in at like a reasonable price.

Yeah, exactly, quite, you know, quite affordable. It's absurd. And there's no subscription for each other language model, which also is confound. founding. So it's got an interesting foreign factor. They built their own purpose-built

large language model to specifically help you navigate services. And this is kind of what we hoped Siri would be. And it's kind of what

we hope chat GPT app would be, but GPT's app does not let you connect to apps. And Siri does, let's face it,

a modest job, I don't know, 10 years in, When was Siri launched? Yeah. I mean, it's been around for closer to 10 than five years. Yeah. Oh, yeah, definitely.

You still try to get Siri to play a song on a playlist or to navigate using ways or to book a reservation. Like, none of that stuff works very well. I will say Spotify does work now. You know, if you ask it, hey, play this playlist on Spotify, I would say two out of three times it works. Yeah. But, you know, that started October 4th, 2011, I think, is when it was launched. So that's when Siri was released. So we're in year 14 of Siri not working well.

Yeah. And the fact that a small startup is able to create a device and the back end experience shows how much work Apple has to put in, or maybe they're putting it in and we'll see it in their next keynote in order to give us what we've always wanted. The question for me is always, you see these keynotes, and those are canned examples. It's probably a canned, not-loyal.

live demo. And as we know, when we review stuff here on this week in startups and we do our little reviews here and we will get a, we will get these and we will do live reviews, what they show is something that is a sniper shot. It's a bullseye. And then when you get it in the real world, if I ask it, hey, can you get me some sushi? Give me the most, you know, ordered stuff. It's going to send me eat miso soups and three at amamis and no sushi because it misinterprets that, right? All right. So you know what's crazy is basically, uh, it's sold out. And when it sold out, I think it had 10,000 pre-orders, they ended with 30,000. But then a couple of really clever engineers and, you know, someone we've demoed some of

those up before, Alvaro here, he basically went and rebuilt it as an Eiffel app. And so backended by the LLM of your choice and, you know, operating. And so I'll kind of just run this quick demo where, you know, he kind of does this here. So we'll play this. Call Tom and confirm our reservation. And so it's going to click, it's going to click out. Sure. Give me one moment while I call him.

And so what's incredible, and he does a few different other demos in here. And so really, you know, what is this saying, JCal? Like I already have a device. It's super powerful. It already can run applications. And so to me, when I saw this, that's going to be the challenge for them, unless they kind of go and offer both to everyone and say, if you really want like a smaller device because you don't want to be distracted by your phone,

you're going out on the night, go and do that. Or you don't want any of that. You don't want any of that expense. Just come and get our iPhone app and pay for it. All right, listen, selling software is hard. It's hard right now, right? 2022, 2023, it's been a grind. 2024, it's going to be hard too.

Everybody's making very thoughtful decisions. And the last thing you need is to slow your sales team down because you don't have your sock queue dialed in. So if your SaaS or services company that stores customer data in the cloud, you need to check out Vanta. Vanta will get your startup, SOC2 compliant easier and faster. Vanta makes it really easy to get and renew your SOC2. On average Vanta

customers are SOC2 compliant in just two to four weeks. Compare that to three to five months without Vantta. Vanta can save you hundreds of hours of work and up to 85% on compliance costs. And Vanta does more than just SOC2. They also automate up to 90% compliance for GDPR, HIPAA, and more. You can't afford to lose out on major customers because of silly stuff like lacking compliance. Just work with Vanta. Get your compliance automated and tight, tight is right, and close those big deals, the lighthouse deals that send all the other customers to you. Here's the call to actions, very simple. Fanta's going to give you $1,000 off at vanta.com slash twist. That's vanta.com slash twist to collect $1,000 off your sock too. I love a purpose-built device when it is for a specific vertical and it really helps

solve that problem. The one I can think of is the GoPro or the Insta 360. You know, if you're going skiing if you're a surfer, you know, your iPhone, yes, you could buy different things to, you know, strap your iPhone or smartphone to your body, but those purpose belt devices that have motion stabilization, et cetera, they work really well, correct? Yeah. For that specific purpose. So does it mean that 99.59% of videos taken in the world are going to be just perfect on your iPhone, but it does mean for skiing and for surfing, you know, this is a much more preferred

device. And so the question is, is ordering your Uber going to be distinctly better because of the form factor? And I don't think that answer to that is yes. So what is this form factor? And I would put the humane, now I haven't used it. So I always like to reserve judgment until I use it. The humane pin that you wear that you put your hand out and like that's supposed to be better. But you know, you see it's kind of heavy. So I love the idea of wearables. But I don't know if that one actually is purpose built for that. Now, I will say my ultra watch, which is the first passable Apple watch

in my mind, I think the other versions were all garbage, is absolutely purpose built for me to look at my ski runs or my height, exercising my heart rate, to take my heart rate, it does something unique in this form factor. And it's also great for getting a message without taking the phone out of my pocket in a meeting.

You can leave the house for the entire day with your Apple watch. you can maybe you can you've done it I'm gonna try it I'm gonna have and put my I haven't turned on a service because that's like 10 bucks extra a month or something 15 yeah yeah you got added to your 18 I didn't add it to my bill yet I'm gonna do that I just yeah for some reason I was like I don't need it but I think actually that would be good if my battery died on my iPhone so yeah you can I do I do know people go running with just their watch you can pair your AirPods with it your AirPods you can then do all your voice to text and all that

Like you can do it. Yeah. It can happen. Yeah. So I mean, that's, and I see that more as a backup than a primary, but I do see the health stuff on the watch and the working out stuff as a primary, not a secondary. So is this primary for anything? I think the answer is no.

The question I have is why is everybody so afraid to consummate a transaction with an LLM? I think it's because of like who's responsible for the wrong thing being ordered. Yeah. You've been able to use Alexa to order coffee for a long time. People don't do it. The concept of getting the wrong thing and buying the wrong thing is just too scary for people. Just like ordering on the internet was too scary for people. So I don't think people feel confident in it. But eventually saying reorder the coffee I ordered last time on these devices would be good.

Or, hey, just get me four cheeseburgers, you know, from in and out. No onions on any of them. You want the Philhomewis? Order the Philhombole. Yeah. Give me four things from the melt. And I'm only going to eat half of each. So I'm only eating two sandwiches. Why do you think that is? Why do you think people or, app developers are not consummating transactions and taking that leap of faith that it could be done. Amazon did it, but Apple won't. And chat GPT4 won't.

Yeah. I think what happened is, you know, there's three layers to what happened. I think it just wasn't as good in the ecosystem prior to LLMs. And I think LLMs, when they do it, it's magical, but they still hallucinate. So before you had this problem of like, it just wouldn't really work. And it was like too structured. Like if you wanted to get your order, you'd have to be like, Alexa, from DoorDash order. You know, you have to kind of really structure it like a robot, right?

And then Siri would never just do it. Now LLMs can pull it off, but they require kind of extra fine tuning or it's kind of like creating like a custom model. So I think we're very close to getting that magical experience where who is going to be a responsible for it, the language model or the app developer and the e-commerce provider? Is it Uber Eats and Doradesh's job to get this right? Or is it chat GPT, Bard, etc.,'s job to get this right in your mind? It's a great question. I'll answer in a slightly different way. I think the app developers are going to get it right before the, you know, the large language models. The large language model folks do. Because they have the data. Exactly, right? They have the corpus of data.

they know how people are going to want to interact with it and how they mess up and how they do all that. So here's what I want. Stanley, you're listening from DoorDash and Dara from Uber Eats. Here's what I want to do. I want to come in and say, show me dinner for our five and our family, which you know is two adults and three kids, two of them who are seven, one who's 14. and give me some options for Japanese food. And just show me a cart already filled with stuff, right? And then I say, but never give us onions and never give us spicy stuff because or only give us only one of us like spicy.

And so then if it was like, hey, here's your cart from your Japanese restaurant. Here's your cart from this restaurant. Here's your cart from this restaurant. And then you could start with a filled cart of items and then maybe say, you know, oh, I look. at that, you know what, swap out the uni, I'm not a fan of uni, but give us more salmon, you know, double a salmon, less of salmon, less

or give us less exotic pieces of sushi, give us more basic sushi for the kids. And it just did that. And it just did that. And it has a great job now, because it has like that reorder and they have that all kind of, they've got all the data, going back to our point. Yeah, exactly. Just like you could say to Spotify, hey,

give me a playlist of 70s road trip music. And it's going to find a 70s road trip playlist. It doesn't exactly understand that, but, or maybe it does, I don't know. Because didn't Spotify have their own assistant built in at some point? They created their own Siri. You know, everyone is tried,

but in the same way that Siri has failed, in the same way Google assistants never worked, the effectiveness that we get with an LLM-based one is we haven't seen it fully coupled together yet. Got it. All right. Well, this is probably some, this is the opportunity for 2024.

Yeah. I think I give a lot of kudos to the rabbit team for being creative. I don't think it's going to work. I think it's going to be like one of these devices that historically, you know, just doesn't find a niche for itself. It's just too, but I do love the fact that they're bold and they put it out for $199, which means they're going to probably lose money on this thing. There's no subscription to it.

I guess you can put in your own SIM card. It just feels to me like this thing's not going to work. And everybody's going to want it to work like a phone. Yeah. So why not just have the phones? up here. Yeah. Yeah. In terms of grading it, I think the way that they've built the service, at least the way they showed in the demo where you can log in and connect all your apps, I really give them an A for that. Like, they really deal that. It's well done. No one has done that

at all. Like you open AI or BARD or anyone else, right? The authentication of your Spotify, your DoorDash and your Uber works well. Can I ask you a question? Was it clear if that happens on your desktop with a keyboard so it's easy to do? Yeah. Yeah. It happens on your desktop. I was trying to find that in the video. I just couldn't pull it up here. I watched the release video on 2X and I was scrubbing back and forth kind of looking for it and I couldn't find it. So that to me is a really great idea is I set this thing up on my desktop where I have a full keyboard and I can say, hey, these are my preferences in order. I usually want an Uber black. These are my 10 favorite restaurants. This is my two go-tos for pizza in order. These are my three

Japanese in order. If you can start to do that, that's kind of cool when you start thinking about it, right? Like, if I order a hamburger, I want to go with five guys, then Shake Shack, then, you know, whatever. Oh, I found it. This is the screen. Oh, there it is. Yeah. Yeah. So you connect your Spotify, Apple Music, Uber, DoorDash, eats, you know, all your services. A lot of founders are great at going from zero to one. This takes vision. It takes creativity. And most of all, it takes a bit of hustle. But those same people can often struggle going from one to a hundred to scale efficiently. You're going to need process. You're going to need process. You're structure and that starts with your product. So if you need a more structured engineering approach,

you need to check out MEV. M-E-V. M-V helps businesses build and maintain their product faster and more effectively. They'll make your product more stable, scalable, and secure. They'll also set up custom infrastructure for you and they can even help build additional features. For each of your need, MEV organizes an entire tech team comprised of senior engineers, delivery managers, DevOps, Q&A and of course designers to make it look really, really top-notch. Mev has been in business for 17 years, and they've helped the following companies build spectacular products, Cartier, Intuit, and the Ozempic maker Novo Nordisk. Shout out to Ozempic. So let Mev help you increase product velocity and make product engineering more sustainable. Here's your call to action.

Mev is going to give you $30,000 off your first three months. That's right. Get $10,000 off per month right now at mev.com slash twist. That's Mev.com slash twist for $30,000 off your first three months. I want chat GPT4 to do this. I want Google to do this. I mean, I want my phone to do this. Like, this kind of exists a little bit on your iPhone when you pick your browser, right? You pick your default browser.

And I think you also pick your music player now. Oh, I didn't see that yet. Anyway, that's what these things should do. You should pick your music player. You should pick your ride-sharing app. You should pick your food ordering app. You should kind of say this is my number one food ordering app. This is my number two, right?

Just so when I shop from Amazon, I target Walmart. This is my go-to. All right, very cool. I'm going to give it a B-plus. I'm going to give it a B-plus. I thought it was very bold. I understand why people liked it. You know, the combination of price, it had wow factor.

It looks stunning. And it was $199. You know, if they ship three, at $200, if they said six million of those, $30,000 of them, they're going to lose money on that, I think. I don't think this, I mean, what is the building materials on this? it's got to be $199 shipped. And then all the returns and stuff like that. This is very dangerous for startups.

It's a smaller screen. They're not dealing with the, you know, you got to bring your own SIM. I don't know. It seems like, you know, it's probably... I don't know. Just one of the things I see with hardware products like this is the founders underestimate the true cost,

which includes returns and shipping and, you know, delays and other problems. And so with hardware, it's so hard, I think they put it out a price point that is, They should have said it's going to be $199, but we're selling the first $10,000 as the signature $10,000 for $3.99. So if you want one of the first $10,000, it's $3.99, and it comes with a number on the back of you have your thing. This way, at least they have a little extra money to invest in the software, right? If it was actually, it would be better to sell $10,000 of them at $500 than to sell...

Well, you know who nailed this with Elon? With the Teslas, right? with the roadster, the model S, the X, and then the three and the Y and what we get. And having the signature version. So it was 100 signature roadsters and a thousand signature model S's. How much were those? Were they like 150? 150.

They were probably $20,000 more, $30,000 more than when the actual products got off of the signature series. But I think the signature series historically are going to go for twice as much as a non-signature. So if you own a signature series, I think you probably have. has a little extra oomph in it. Probably pays for itself. All right, let's keep moving here. You touched on something, and it was actually, it's quite important. Last week, there was a video release between Bill Gates and Sam.

Oh, yes. And I want to just touch on that really quickly. I'm not going to play the video because it's a bit too long. But basically, he talks about, like, what we're going to see for GPT5. And in GPT5, it touches on exactly what you said, which is video capabilities, which we know are coming, like kind of more multimodal generate video. videos, boosting reasoning and reliability. And then the last one is enabling user customization.

So I do think for all the startups out there, you know, including Rabbit, they are going to kind of build that same interface that we were just showing there. It's a 30-minute talk between them. It's launched on Bill Gates's pod. I suggest everyone listen to it that's building stuff in and around AI. It was really good. The fact is he's super excited about Chap-C-T-5. And 4.5.

And he also did a talk at Y Combinator, his alma mater, and where he worked. And he seemed to think like these new ALLMs are going to be quite powerful. So there's a little bit of buzz that like they're going to drop some really. And he thinks AGI. I think there was another part of this where he sort of signals that he thinks AGI is coming sooner rather than later. Yeah, he does touch on that there. So we recommend everyone watch that one. I'm going to jump to another thing, J.

I'll get back to Amazon in a second. Okay. I thought this was pretty exciting. This is a good application. it's basically like a sidekick for older folks. And so it's like a little device. And it's kind of interesting as like a lamp. And then as a, you know, like a LCD display, a speaker.

It looks like a charging station for your iPhone, but it's got like this little lamp. And this is designed to be a companion, this L-E-Q? Yeah, it is. Yeah, and it's designed to be a companion for seniors. And, you know, loneliness in seniors is a really big problem and, you know, keeping active. And if you look at what they did, you know, it can basically, and this is, you know, kind of powered by LLMs, it can chat with you, it can play games with you, it can do all the, you know, kind of interactive

things that I find are, you know, really, really interesting. And I like kind of how they put that little light on the side there to not make it like, fully that you're staring at a screen. So it's like a... Yes. But it also is not fully like a robot. So I thought this was one of the more interesting things out of CES this year as well. So there was a movie about this, robot and Frank, that came out in...

Oh, really? It's a comedy. It came out in 2012. In the near future, an ex-jewelief receives a gift from his son, a robot butler programmed to look after him. But soon the two companions try their luck at a heist team. Now, I haven't seen this, but I got to see this because this is like, it's got my friend Peter Sarsguard in it and Frank

Langella. He's a great actor. And so this looks like a pretty compelling movie that did this. And then, of course, there was the film that my daughter got obsessed with seeing, I didn't let her see it, but Megan, the robot horror story, remember the viral dance? Yeah, yeah, yeah, yeah. But this idea that an AI companion would

relieve loneliness or do a little bit of an AI, of course, there was the kid in AI, I think it was Haley Joel Osment, played the kid in AI, yeah, who'd become friends with the little boy, and then he has the teddy bear that is his friend and just sort of multitudes of artificial intelligence, you know, helping people be less lonely. And so it's dystopian in some ways, but if well executed, I could see it relieving a bit of stress for people. And so, yeah, interesting. And so the one thing I want to talk about here is that

like what this does, and going back to, you know, because CS, we're going to stick with a little bit of the hardware theme on some of these. What this allows companies to do that know how to make great hardware where, you know, this integration of hardware and amazingly difficult software became harder. And you'd see stuff out of CVS, but only the big folks were winning the Samsungs and the Apple's. Oh, Apple doesn't go to CS, but like the bigger company's LGs. And now what you're going to see is the ability for companies that have, you know,

strong strength in standing up hardware, but that can leverage large language models to create incredibly compelling experiences without having to have, you know, the depth of the team required to build like an entire operating system or something along those lines. And this has some really cool goals in it. Like I'm looking at the copy on the website. Always important for founders to understand writing good copy and having a really good value proposition that you communicate well on your website to incentivize people to have product market fit and to buy the product.

Build healthier habits with goal setting and encouragement. You know, like taking your pills and exercising. Bullet point number two, exercise with easy-to-follow video workouts. So I guess it will talk to you and say, hey, you know, let's do, put your hands above your head, stretch, whatever. Three, challenge yourself and stay sharp with cognitive games, right? There were a whole bunch of cognitive games. I forgot the name of it. I got very popular.

There was one app that, was it Lumosity that was doing. Oh, yeah, it was. It was. So Lumosity was like, hey, there's this concept of, you know, exercising your brain essentially. And I think wordal and crosswords, people put in that, Sudoku. There's like a bunch of things that people do. And the New York Times has gotten into this. Like, you get people into a habit.

So, you know, having wordel or word challenges or jeopardy in this would be incredible. And number four, relax with various mindfulness and stress reducing exercises. That's calm.com, right? So this has lumosity in it and wordle, essentially, like some cognitive games. And it's got an exercise app in it. It's pretty good idea, I think. If you put all that in one and you make it a really great. And then some other things, set reminders for errands appointments and medications.

Okay, that's really good. notify loved ones there is need for support. So that's that I fall in and I can't get up, you know, the alert me kind of stuff. Search resources in your local area. There's your yellow pages. And then stay current with news, weather sports updates. Hey, that's your Yahoo or AOL or turning on CNN. So this actually, if executed really well, could be amazing.

Here's some more things you can do with it. Easy video calls from your friends and family. Okay. So that's your iPad kind of thing. Send and receive messages and photos of loved ones. Okay, that's like your smart frame. share your favorite memories and life lessons through digital memory. Okay, that's your ancestry.

And send virtual greeting cards and special events and birthdays. That's your Blue Mountain Arts, if you remember that, if you're in there. So whoever made this, I have to say to the team there, I'm giving this an A. Because I think this team, and this is this weekend startup. So I always like to point out when somebody does some really good startup best practices. I just read 12, not three, 12 bullet point items in one package. If this thing actually is done well, there's like 12 items in here that are, you know, other businesses in and of themselves. So really well done to the LQ, L-E-Q, E-L-L-I-Q.com.

What did you give it? For me, I'm in the same. I think it was an A. The design, the experiences, the same thing. I went through it all. And just a real proper integration of good hardware and the latest in L-L-L-Ms and user experience and collapsing a few of those businesses into it. So I thought it was incredible. Now, I would like to see this in the form of a frameless TV, uh, picture frame, smart picture frame. This could also work in that format. So I don't know if you know about this like Samsung, uh, frameless, uh, or they call it the framed. The frame. Yeah. They basically, and I stayed at an Airbnb that had one of these. And I didn't realize it was a TV. Yeah. Because it has like a border on it and they put wood on it. And it looks like it's a piece of art. And I guess you can buy art.

but it can also work as a TV. So I just aesthetically love this concept. And I just think a smart TV that had a great interface. This is why Apple should have launched a TV. And there was a rumor for a long time, and I had heard it from people who were in and around Apple back in the Engadgety days that they were working on a television. Instead, they came out with Apple TV to slap on the back of TVs.

And I guess you can get Apple TVs built into some TVs or a light version of it is. This would really have been amazing. If you buy any smart TV now, it has like an Apple TV. Apple TV in it. It's like they've done that. But these guys have really nailed, and I just have it up here, but these guys really nailed. I've seen a couple of these as well. And they're really well done. And they kind of make them like a matte screen. So when they're, you know, when they're not in like TV mode, they really do look like a piece of art. And it looks like you can get like, enjoy 100 pieces of art from some of Disney's greatest stories on the frame. All right, everybody, 2023 was obviously the year of AI. We saw all the amazing productivity gains that can be happy. but in 2024, it's all about adoption. Are you really using these tools every day? Because if you're not, then you're falling behind. But here's the hard part. How do you separate the signal from the noise? There are a ton of AI tools out there. And while some are revolutionary, others are just parlor tricks. We all know that. Here's one way you can start. Head to Imagine AI Live in March.

It's a conference taking place on March 27th and 28th in Las Vegas. And when you go there, you're going learn how to apply AI to your business directly from the people who build and use these tools. We'll witness live demos of AI products that can make your business run more efficiently. And you'll hear founders and CEOs explain how they're using AI to reshape their companies. The founders of this conference are huge fans of this week in startup so Twist listeners can get 20% off at ImagineAI. dot live slash twist. That's 20% off your tickets at ImagineAI. Dot L-I-V-E slash T-W-I-S-T. So again, back to product market fit.

You know, if you can show apps, if you can show use cases and people go, I want that use case. When they get to two or three of the use cases, you know, people start hoovering over that buy button. I know I've been hovering over the buy button on one of these. I'm just figuring out, like, what TV am I going to throw away to put something gorgeous like this in? And I was thinking of my new set. I'm building a new office here at the house. Oh yeah, it should be there. Yeah.

And I was saying about having one of these and maybe just having kids, photos, skiing, whatever, which from pieces of art, but have it rotate, you know, be kind of cool. Awesome. Let's keep rolling. Keep moving. We've got a couple more of these to go. Okay, so the next one, which I thought was really interesting as well.

And this is like a touchy subject for, you know, parents with kids at certain ages, which is smartphones for kids. And the smartphone controls are just really lacking today. because one, the Apple's UI doesn't really allow you to do it in any kind of significant way that's usable. And two, even if you have some of those controls, you're not really getting any data out. And so what these folks did, MMGuard, is they introduced sort of an AI-powered smartphone, which gives you summaries of, it gives apparent summaries if, you know, kind of like bullying is happening or, you know, the content that's being used is inappropriate. or you can see the type of messages.

And the beauty of this is, this doesn't have to integrate with all the apps. So what they really do is they use the power of multimodal LLMs to just kind of constantly do a screen capture, send that back. And basically, they don't have to integrate with all the apps. They don't, like, the developers don't have to do anything. So the end, you know, your child gets the phone experience they want. But you as a parent can get all the detailed metrics, plus you can get the alert. are on bullying and things like that as well. So you can have monitored social media or not monitored social media.

You can monitor their texting or not. But it's using an LLM to know if the language contains, I'm looking at the website here, cyberbullying, sexting, drugs and alcohol violence, or a predator or suicidal ideation, all this stuff that parents worry about. So, you know, those words change over time. We had words for marijuana, Mary Jane, you know, or something, or four, You know, to hide that from our parents or whatever, to speak in code or whatever. Kids do this.

And so, yeah, they're going to do those kind of secret codes and text or whatever. You do want to know and you want to get an alert if, God forbid, somebody starts sending, you know, they start sexting or something or sending, you know, inappropriate pictures or they're getting bullied. All of these things can be really gnarly. In fact, you know, I was reading a bunch of kids got in trouble because they were sending pictures of themselves. and just to be delicate here, you know, some kid forwarded one of these pictures of somebody who took a picture of themselves,

you know, with less clothes on, and then they got arrested for child pornography. And so this is like really deadly serious stuff. If you're trafficking, even if you're underage, in pictures of other underage people, you know, this could result in somebody going to jail,

getting expelled from school. It's a basically, it is a crime. And then young people can basically commit these crimes without ever really fully gronking what's going on here. And so this is really important. The platforms today don't do a good job.

No, what's their incentive? Yeah. Back to incentives. Yeah. And so I really want to give these folks of kudos. Like this is a real coming together of, you know, cell phone's been around for a while. Kids are using them.

They're all using them. And now this allows someone to build software that's really advanced that can give you, you know, sort of the insight into what's happening in the messages. and sort of the analytics on how they're using the device, they don't do a good job at all. So I really like this one. I thought it was, you know. You need to have a third party do this.

You can't rely on the headset manufacturers, the iOS manufacturer or the Android. It really has to be done by a third party who sells this product to a parent so that their customers, the parent, not their customer's iOS or iOS's customer. or iOS's revenue stream or Facebook's revenue stream comes from addicting kids or getting people to use the phone more. Yeah.

But the key here is also, you don't want to limit the kid to be using some, you know, what's that old people phone? You know, they've shown it. Flip phones. Yeah, no, not even flip phones, but like they sell one now to them as well. It's got like an aim for it, I think. But you don't want to give someone that's like totally limited and neutered because you've taken over the whole phone.

This can just be a regular Android phone and basically can use all the apps in the app store. but there's some features for the parents that I think are really important. I'm going to give this an A as well. Yeah. Oh, wow. Look at the gate. Like, amazing. And you're saying in this case, because I see they, you can download their apps and have

software, but in this case, they're making actual smartphone. Yeah. Well, I think it's a Samsung phone with that built in. That's what it is. Yeah. So it's kind of built in. Yeah. These are smartphones designed for kids.

It's built in to the phone. So this is a collaborative. effort between MM Guardian and Samsung. Okay, great. Just such a great idea. Such a great idea. I would like to see, you know, Apple maybe get a little more. Doesn't Apple do something like this with images? Wasn't there some blurring of images that occurs?

They do. They've said this, not plain, but they've said that they are monitoring eye message for inappropriate photos. They have said this. What I would say is given what I've seen, the amount of controls that Apple has and allow is very, very limited. Like, once your kid has an Apple device, it's the wild, wild west. Or it can be totally locked down, but then you can't use the phone for anything.

Like, you can make it like no apps, no messaging, and then, you know, then the phone becomes. If communication safety determines that a photo or video your child has received or is about to send, appears to contain nudity, communication safety blurs the photo or video displaying a warning that it might be sensitive and offers ways to get help. So kudos to Apple doing sensitive content warnings. I know people feel that censorship or whatever, or they're getting too involved and it's scary. But, you know, it's a feature that people can opt into by being, you have to turn it on, number one.

And number two, if you're in the Apple ecosystem, you're opting into that to protect your kids and to have a bit of censorship. You're opting into having curation. So one person, you know, censorship could be another parent's curation. and that one person's like safety is another person's freedom of speech. So, you know, just you pick your operating system.

If you want to have like a more contained experience, you can use Instagram. If you want to have a more free flowing experience, you can use X formally known as Twitter. Yeah. Great. Awesome. Yeah, I'm giving this an A for sure.

Time has come. Yeah, great job. All right. Awesome. A lot of strong stuff coming out of CES. Yeah. It was, you know, I think it didn't get the credit it deserved.

But like to me, it felt really reinvigorating. I was only there for like 24 hours, but it was pretty incredible. All right. Okay, we're going to switch gears here and a couple more things. So last week we also got the launch

of the GPT store. And so... Right. Yes. So we got the launch of the GPT store. So you go over here and to explore GPTs. In your sidebar on the left

is explore GPTs, yes. Yep. And basically, what you can see here is featured like a classical app store, trending, and you can click through those. interesting what we'll analyze this in a second by offered you know by open AI themselves

and then a couple of different categories and you know what I've got to say I think they've really nailed this they have done a good job and you know I'll just start with kind of one of these that I did before and then we can do some live ones and you can pull some up as well I did this Canva one and so I said how about an inspirational quote graphic for social maybe I want to post something on X or or Instagram and it said, hey, you know, the typical thing, Canva comes back. This is what do you want to say? I wanted to say at Carpetium. And basically, it gave me these two graphics right here all in line right out of Canva. Yeah. I mean, we were talking about the Canva apps in the early days, just not working particularly well. And these integrations

not being well. It just feels like this is their second swing at bat. And it's going to be a little bit tighter. Do any of these include revenue and subscriptions yet? So no, they haven't. The only thing that they talked about, and they talked about it at Open AI Day was like sort of a Spotify model. And so what everyone is doing right now is sharing their usage. And so the idea, at least based on opening idea, is that you'll get a RevShare based on how much they're being used.

I'm looking at the, you know, they seem to be really promoting the All Trails app. And I use the All Trails app. Yeah. And it's pretty great. So which trails near me are docked. friendly and it actually you tell it, hey, Bay Area and it just starts giving you really great summaries of those, you know, places. And it understands the categories, the average duration, the difficulty rating, the elevation, the length, all that kind of stuff. So this does feel like

a great front end to all trails. Now, based on what they've given me, why would I subscribe to all trails? So I'm wondering what all trails is doing here with this product, because this feels really like you just basically cannibalized your entire business. So maybe you could talk a little bit to that since there's no revenue here. I guess maybe they guaranteed all trails a certain amount of revenue to put this on. What do you think is going on here in the background? Well, like I said, they have said they're going to share revenue. So in order to use these, you have to be subscribed.

And so you are paying $20 a month, which is, you know, more than what you're paying. Explore GPs. I'm paid. Yeah, I'm not paid. You can't explore GPD. Yeah. Yeah. Yeah.

I think like that's how what they're going to do is like very similar to Spotify, right? Where, you know, you spend your whatever Spotify is at now $14.99 a month. And that gets sharded out to the artist. You know, while we kind of think through that, I had this funny one. I think you'll like. Let me share this other one I found, which is called VC Associates. So it says, pitch me your startup and I'll enhance your web shirt. Yeah.

So we should do it. and thinking about building a company that takes ice from the poles and sells it as expensive ice to put in your mixed drinks. So it basically kind of goes through the whole process. We saw a few similar things like this, but there's a lot of fun ones in there. So I recommend everyone take some time, go in, double click in, mess around. and send us ones that they're finding. There's what I read on the launch day, there's three million of these.

Three million. Yeah. So I'm guessing there's thousands of high quality ones. Yeah. You know, in terms of how this typically works, so people know, when you're kickstarting something like this, yeah, you're going to do the revenue share.

But what you do is you go out to the top brands, like all trails, and you say, listen, would you build an app for us? And they say, yeah, you know, it's on our list or whatever. whatever, but and they say, okay, listen, we're going to share revenue with you. And then if AllTrails is savvy, they would say, well, you know, we got to put three developers on this

for a year. That's a half million dollars in our developer time. And they say, okay, yeah, we'll guarantee you a half million dollars in revenue. We'll pay you 250K advance. And then we'll guarantee you another $25,000 a month minimum, right? So if you're a developer and you have a high quality app, the question is, why would I put this there? What am I getting out of it? And then what am I giving up? Because I can tell you right now, using all trails here, the use case it gave me, I would not use all trails. I would cancel my all trails subscription. And so now it might be, do I do all trails through my chat chippy T subscription or do I give them money directly? So chat chpetee just got me to cancel my all trails.

So I just think that's something to think about if you're a vendor to these kind of things. You've got to be careful about cannibalizing yourself. I probably 80 or 90% of the value of all trails is in the chat GPT, the GPT, all trails. Yeah. So the ability to look things up quite quickly. That's all I used for. And in fact, the experience is probably easier here than however you were doing it before. I mean, it's a pretty nice interface here.

So I have to say, like, this is very, very disruptive. And so this has got me thinking about all my startups. I think as a startup, you have to start wondering, like, you, this also is, you you don't have your customer's information. So now you lost me as a customer, but you also lost my email and my credit card number. Well, not that you ever had my credit card number because I went through Apple,

which was what all trails had to decide when they built an app. So every time you build these kind of apps or other experiences, you have to just wonder, am I distancing myself from my customer base? Because it didn't give my email to them, right? Didn't ask to give my email to them.

So they can't upsell me, etc. So, yeah, it's interesting. The chat interface might be better. than the app interface for some of these experiences. But I wonder if they're just going to use that same thing inside of all trails, right? Can I have a chat chip-y-t inside of all trails? Yeah, what will happen? That's what you would think, right?

I think the real thing here is like you create a limited experience that, but in your case, it's funny because they got your basic experience done, which is a little bit scary. So you have to really, as founders hearing this, you got to be pretty thoughtful. Again, I give, you know, chat-chip-tis now. I think, you know, the quality is going to be. going up B plus. I mean, the whole concept, it's an A plus concept, obviously. I think the execution is somewhere in the B range, B plus range right now for me. And it's getting better significantly. And now that there's a sustainable business model, I could see people investing in it. Putting aside

the all trails caveat, you know, like it didn't offer me to download the app. So that was what I was expecting is if you're going to do this, it should have upsold me on the app, et cetera. It does give you like a little picture of the all trails, but it doesn't, you know, I'll show you. So, you know, it gave me this little logo here, and it's giving me all the different trails with dogs. You know, I just put Bay or San Francisco, and it's like, here's the Lanzan Trail. Here's the Rodeo Beach Trail. I know that trail. I've been to that trail.

I've done that hike. Wow. This is really nice. This is one of the nicer ones. Yeah, it's gorgeous. You know, it gives you a nice little, and it gives you everything you need. The only thing I don't see here is the map. Click on the Trail. So it's telling me to click on the trails to learn.

more about the trails. Oh, I guess that's this link here. So I guess if I click the link, it does take me to all trails. Okay. So I guess that's what the all trails people are thinking is like if we give you. Like Legion. Yeah, maybe. Maybe we give you a little lead gen.

Yeah. Be careful here, folks. You could. Yeah. Yeah. And then I wonder if that all trails data is now going into the language model. Yeah. Or not.

So I have two thoughts here. One, I really kind of agree with you on the ones that are done and featured are really good. Unlike the app. store, which has a app approval process. There's no approval process here. And the thing that I've seen is, and that's, you know, why they have $3 million, is the quality is very low. So even when you kind of just scroll past the top 10, and I wonder how that plays in,

because one of the things Apple was able to do, and I obviously lived this very close, both on doing featured apps for folks and then building apps, was that they were able to keep the quality bar quite high. And one of the things I've experienced in the last week in trying a bunch of these, because there's no approval process, the bar is really, really low. And you do kind of do something and waste a bunch of time. You're like, oh, what a waste of time for me. And I wonder how they're going to be able to navigate that. It's a brave new world. I think if I'm the owner of the business, if I own all trials, I'm putting a limited version inside a chat GPT to learn, provided they're not sucking that data in. I'm putting everything on my website behind a paywall now with a no scraping rule, you know, no training rule.

And then I'm going to sue anybody who put my training data into their large language models to protect myself. And I'm only going to allow my LLM to do like, you know, three results and then say there's 17 other results inside of the app. So you get to skim the cream. But for 17 more hikes, click here. And to get the, you know, geosolm. coordinates of the height, you know, and to track, I would just really be powerfully upselling people on getting the full app experience. And so I do think they're going to need to authenticate.

So if I could authenticate my All Trails app, that would be pretty powerful. And then it let me do full on, you know, queries. Like we saw it with Rabbit. Like Rabbit, yes. Yeah, yeah. That, I think is the ultimate solution for all of this. I think Sam's a great deal maker. So this whole thing with the New York Times would be solved by to get any New York

Times wire cutter wordal results. you have to have, you have to authenticate what you're in your times. And then that makes you both subscriptions more powerful, right? So yeah, great job. I think Sam Waltman's, you know, really cooking with fire here. iOS needs to study this and then figure out if their app store is going to get disrupted. That's what Open AI is trying to do. They're trying to make better apps than the app stores.

But also, don't you think these developers should just be putting language models into their app? So when you open an app like all trials, it lets you search or do a question. and then you don't need to go to chat, GPT. I think that's the big push this year. And that's what I'm kind of, what I would be betting on is what we said earlier. If you have the data and if you can put the engineering effort in to get a model, to understand your business, your data, and create a user experience, I think that will win right now before kind of the aggregators can.

And I think Yelp is trying to do that. So Yelp launched a series of features to enhance the consumer experience. and they're trying to let you do searches on Yelp, I think. You know, so here, let me show you this. Because, you know, there are some people who have very robust businesses. Yeah. You know, that exist outside of... So here, I'll read it to the audience.

With the best amount of rich user-generated content available on Yelp, we've long used AI to power solutions across the platform, including DISH recommendations, organization of photos and automating content moderation among others. The recent advancements in LLMs, we are continuing to invest further enhancing the search and discovery experience on Yelp to make it even easier for you. And so we're announcing a number of new search features today,

which build on our investment in LLMs to surface smarter search suggestions and power insights to help you find the right business by leveraging these techniques. PR speak, PR speak, PR speak. And so, you know, pulling out, you know, these kind of experience, I think is going to be really great. Like actually being able to query, as you can see here, they're doing some unique queries. Let me just put my glasses on for a second.

Breakfast. We ordered the All-American breakfast and the eggs were cooked just right. I got the pancake flight with the side of bacon. So, I mean, it's like just, that's the keyword breakfast. What I want to do is say, who has the best Peking duck? And, you know, what does it cost? And put that into a list for me and put it on my to-do list, you know? And that would be the better experience.

So, you know, it's very 1.0, but I do think there are, you know, some really interesting. Give me the highest rated sushi restaurants that are two stars and tell me what the top three dishes are at each. Boom. That's like the kind of stuff people want to do, right? Tell me the places that have the best, you know, crispy. We're seeing this everywhere. We even see it with, you know, kind of grok and Twitter, right? Where you want to just ask, like, kind of some of these questions and people are so working to do that.

Call us at definitive. We'll help you do that. Yes. We'll help you do it. We'll help you do it. Sundeep at definitive. com. Great job today.

And we will see you all next time on this week in service. Bye-bye.