How Apple Stumbled Into AI Hardware Success, Salesforce Revamps AI Pricing, AI Homebuilder Company

22:36And then we just announced Casper, which is in the U.S., very interesting. It's from Melio, you know, our newly acquired subsidiary. And that is a ledger agnostic AI assistant, meaning it will work across any ledger, not just the zero ledger. So we have a range of features. And I think the most interesting thing is for our customers who really don't know, let's say, how to use AI that well, they're taking embedded features from us. But right now on our APIs, we have 6x the volume, 6x since January. So that's why products like Zero Force are really important as well to really cater to that sophisticated customer who wants to build right on top of Zero.

23:14Aaron Tilley:So, Kendra, let me ask you sort of a higher level question here. You've managed a number of different companies in your career. You've also done a lot of venture investing yourself. I mean, we talked a year ago when you came on the show about how you have brought the company to generating free cash flow sustainably. The growth rate is healthy. If you look at the stock price over the last year, I mean, it's down about 47 % in the last month, though. It started to creep back up a bit. It has, yes. So maybe some narrative shift here with the SaaSpocalypse. What's your read on what investors are reacting to here?

23:53Aaron Tilley:Is this all SaaSpocalypse concerns? Because, I mean, look, financially, the company, look, it's growing. It's a Rule of 40 company. So what do you think investors are? It's predominantly SaaSpocalypse. And as you're saying, like people are struck, a year ago we hit our all-time highs. And then, of course, the company, alongside others, suffered, I think, in that time period where people could not tell the winners from the losers. Now, to your point, if you think it's zero, we're a full-stack player. We have infrastructure. We have data. We have compliance rails. We have applications. We have an ecosystem.

24:26And then we, of course, have got a GTM layer that operates across 180 countries and multiple channels. This past year, we announced in May we grew 31%. We generated almost$800 million NZ in free cash flow. We're rule of 48. So I think, look, I think this favors patience. You know, you can see this, you know, I think this week with Salesforce's earnings, people are starting to talk about the SaaS-sans. So the SaaS renaissance or call it whatever you want to call it. Oh, interesting. Okay. Look, I mean, I think the most important thing right now is to keep executing. For our customers, we are the builders.

24:56and we're going to give them everything from embedded AI features to the ability to build low-code, no-code agents on top of zero. And then, of course, I think you know we're integrated with Microsoft, OpenAI, Plot. So really we are an open staff with a lot of proprietary assets and an offensive play. We're going to keep playing offense on AI because that's what our customers need us to do.

25:19Aaron Tilley:Right. And so to anyone who says that, oh, I can build a software myself using these AI tools, What do you think? Yes, I would say, well, that's just one layer of our stack, right? So if you want to build applications, go ahead. By the way, I don't know about you, my friend. I've been an entrepreneur three times over. Let me tell you, the thing I never thought I would code as an entrepreneur is my accounting software. If you ask me where I want to spend my time in building businesses. So I don't think, and then maintaining the compliance. Remember that we're connected to thousands of banks. We're connected to payment rails.

25:51We're regulated. We're connected to tax authorities. We file your sales tax, your payroll taxes. By the way, if you want to maintain all that infrastructure yourself, knock yourself out. We're going to keep doing it privately, securely. We're going to open up that infrastructure for first-party applications that we build, but also third-party applications that people want to use Xero for. So bring it on.

26:09Aaron Tilley:Great. Well, Sukhinder, I want to thank you for coming on. It's a great conversation as always. That is Sukhinder Singh-Cassidy, the CEO of Xero here on TI TV. Meta agreed to a landmark settlement last week, agreeing to pay roughly$18 billion. Now the big question that remains is how other social media platforms will follow, but the repercussions of this settlement on the product that is social media are what might be the bigger impact here. To help us assess whether or not they'll help address the issues at hand, I want to bring on Ben Moore, U.S. Managing Director at BeReal. Ben, welcome to the show.

26:45Aaron Tilley:It's great to have you here. Thanks for having me. So we'll talk about BeReal in a minute here because it's a platform we haven't actually talked about on the show, but I want to get your reaction to the meta settlement last week. What was your thoughts on it? Not really a big surprise, to be honest. I think the market has been really feeling it coming for the past few months now. We've been talking about addiction over social media for the last few years. Everybody knows that social media has no longer been social for over a decade. And I guess they had it coming. So it's more of a wake-up call for the rest of the industry.

27:24Platforms that have been optimizing for time spent knows that it's legally liable to do so, and they need a change. And everybody wants a change, including teens, parents, and most of social users.

27:39Aaron Tilley:So I'm looking at the platform changes here. I mean, we've got default daily limits, age assurance, new tools for parents. Meta has agreed to implement some of these changes. We're now waiting to see what other social media platforms follow with and if they agree to similar changes. Do you think that those changes go far enough? No, absolutely not. I think these are remedies that could help limit the amount of time that those teens, those users spend scrolling for their feeds, but it's not enough. I think when Facebook came out, they had a limited feed. They had no algorithm and they were essentially being used as a real social network.

28:27People were logging, checking on their friends and family, people they care about. And then that's it. They were going back to their real life. And I think those remedies are just, you know, product specs that they could have implemented a long time ago. So I think there needs to be a real change, a change that can actually limit the ability for people to physically limit their amount of time scrolling through their feeds. And there's only one way. It's by changing the architecture of the platform, which should be definitely limiting the feed and getting back to what social media was intending to be.

29:07Aaron Tilley:So are you suggesting changing the algorithm? Is that what you mean? I'm suggesting changing the algorithm, removing it if possible. On B-roll, there's no algorithm and it's not a problem. But we know that the impact for Meta will be about their advertising revenue if they do so. That's really what's at stake here. So this is kind of interesting. I mean, removing the algorithm, I don't know that that's something that Meta would ever consider, given that that is really the big moneymaker here. I mean, in the absence of them removing the algorithm or adjusting the algorithm, are there other ways that you have figured out at your company to curb addiction, you know, more creative pursuits around parental controls, daily limits?

29:58Aaron Tilley:I mean, these are the changes at hand. How are you addressing it outside of the algorithm? By a very simple way of doing it, by having a finite feed. I think that's really the real remedy here, the real solution. Just bring back a finite feed, a feed that ends after you're done checking on your friends and a hand-off feed message, just like on B-Roll, that says, hey, it's time to get back to your real life. And I mean, B-Real, this is, again, this is a platform that was very much in vogue a couple of years ago. The platform as it is today it's still that whole idea that once a day I take a photo and I see, is it both sides of the camera?

30:46Aaron Tilley:Is that the idea? Yeah, that's what made B-Rail such a success for Gen Z for the past five years now. Still very much the same DNA and product specs. One notification, one moment to share your real moments of your life and then a feed that ends, no algorithm pushing content from stranger, no control and no filters. Right. Do you think that the other social media platforms, TikTok and YouTube, I mean, Meta was very, they sort of spun it from a, hey, we're getting sued message to, hey, we're taking a leadership role message with their encouragement that other social media platforms follow their lead on these changes.

31:35Aaron Tilley:do you think TikTok and YouTube will make similar changes? Will they go further with addiction? What's your view on it? I think they should. Do you think they will? I think they will because of that open letter that Meta sent out to them over last week. I think they're different though. They've always been an entertaining company. People are going on YouTube and TikTok to actually consume content, whereas on Meta and Instagram, they want to connect with their friends. But nowadays, they just, you know, consume content, media the same way they can consume and scroll through the content of YouTube and TikTok.

32:21So it's two different companies. And I don't think, you know, we should put everyone in the same basket. I think Instagram has been really optimizing for time spent, but also vanity metrics for the past 15 years now, which is a problem. I think -

32:38Aaron Tilley:Like what? Like what? Like algorithm pushing, you know, content from strangers that keeps the user scrolling forever, optimizing for the level of views, the level of likes, the reshares, all of those metrics, those KPIs that are so important for content creators. have been really addicting to them. And that's, I think, the source of the problem as well. So, you know, I mean, I sort of hear you on the finite feed is one way to, it is, I think, a smart way to curb addiction. I don't know that it happens. What do you think all this means for the current moment in AI and maybe lawsuits that we could see coming with respect to use of that product?

33:27Aaron Tilley:I mean, I don't know that addiction is the biggest issue. I mean, there's certainly, there's the data component, there's the misinformation component, there is all sorts of issues. What do you think we learned from the social media case? I think this is one of the biggest concerns, you know, those platforms, those big tech platforms should also address. We all know that the AI slope has been inundating those feeds for almost 18 months now. And this is something that needs to be addressed, that needs to be solved for, because quite frankly, most of the social media users can't tell the difference between a post generated by AI and one generated by a real human.

34:13And that's something we need to also take care of because who knows what's going to be appearing on the streets tomorrow. Can social media be trusted as it is today? The answer is no. And so that's really something we need to talk about. We need to legalize and we need to change, optimize for.

34:37Aaron Tilley:What do you propose? Do you propose like watermarks or what's your preference? I think on B-Rail, we have one solution that really solves for that, is the ability for users to actually not upload any content from their camera roll. And so that's how we fight the AI slop. Everything on our platform is being taken, captured within the app. You can upload any content from outside the app, which makes a big difference. and eventually people on B-Real never have to question if what they see is generated by an AI or a real human. It's the place where they never wonder if what they see is real because of the nature of our platform.

35:25Aaron Tilley:Right. And so last question for you, Ben. You know, I've been thinking a lot over the weekend about why Meta actually settled because this case, it would have gone on for a long time. I think, I mean, there's a pretty good case against them. So I've been trying to figure out, did they think they were going to lose? Was it the headlines? You know, the monetary value is really just a fraction of what they were being sued for. Why do you think they settled in the end? Because, yeah, as you said, it's just a fraction of what they were being sued for. It's just a check for them. And I think it's barely 10 % of what they make every year.

36:06So they could definitely settle for that. They wanted to get away from the noise and focus on their next product roadmap, their next updates, and they want to move away from any litigation so that people don't really point the figure at them.

36:25Aaron Tilley:Right. Well, Ben, I want to thank you for coming on. That is Ben Moore, U.S. Managing Director of Be Real here on TI-TV. There is a lot of talk about humanoid robots, but it is the robotics companies with the more tempered ambitions that are more likely to become commercial successes in the near term. To that end, the information exclusively reported today that Reframe Systems, which is using robots to build modular homes, raised$40 million in a Series A this week. I want to bring on CEO Vikas Enti for a conversation. Vikas, welcome to the show. It's great to have you here. Hey, Akash. Great to be here.

37:03Thank you so much for having us on.

37:04Aaron Tilley:Okay, so modular homes. Tell me, how does this robot actually work? Yeah, so it's a whole system play. We're building microfactories across the country to build high-performance homes. We're effectively chunking up homes into smaller sections and components that we produce in a factory, ship them on a flatbed truck, crane them into place. And the whole goal is, how do you recreate this experience where a home is magically getting built overnight? And we're able to do this by moving a whole bunch of this, the construction into a manufacturing setting in a microfactory. Our innovations are applying some really cutting-edge software and robotics to make the process of manufacturing significantly faster and significantly more cost-efficiently in the factory context.

37:46Aaron Tilley:Okay, so let me just get some of my silly questions here out of the way. So you, well, first of all, how many pieces does it take to make a reframe house? Yeah, a key part of our technology is how do you design a factory to produce everything from a small ADU to a large apartment building? So depending on the product type, the number of blocks vary. So for an ADU, an accessory dwelling unit, we could deliver that in a single module. A module here is a large shoebox for all intents and purposes. So it's about 14 feet wide by 40 feet long. And that's like a living room and a bedroom type of thing?

38:26Aaron Tilley:It's like a one-bedroom apartment? Exactly. So it's a studio with a kitchen and a bath. In fact, I'm sitting in one of those units right now. Okay. As we start going to a lot of... I mean, I'm not going to make you do it, but next time you come, I want the tour of the MTV Cribs version of you taking us inside the house. But okay, so it's like a shipping container type of unit. You could imagine the form factor to be of that shape. But once it comes together, we're able to create multiple types of structures. So our larger projects end up having anywhere from six modules, including gable roofs, so that the shapes aren't necessarily constrained to be cuboidal.

39:11We have a five-story apartment building coming up later this year that's going to have roughly 80-plus modules. Some will be trapezoidal. So we get to mix and match other types as required by the project.

39:24Aaron Tilley:And so just to confirm, so you are making all of these blocks and components yourself? You're manufacturing them from scratch? Yes. So we buy regular building materials. We're buying studs and sheet goods from your lumberyard or from direct source from manufacturer. And then our software and everything we've done to model a building to the nearest path, near pipe, and wire allows us to then sort of translate all of the stuff and instructions that go to our CNC saws and routers, to our robot, to take care of all the structural systems. That's where we're framing our walls and ceilings. Is it all made in California?

40:01Aaron Tilley:Where are these components made? So today we have one operational microfactory here in Massachusetts. So I'll show you, I'm sitting in, is right next to the factory. What we're also announcing is we're launching our production scale microfactory, which is called FabWatt, which is also in Massachusetts. We have done work in California to help with the wealth of Rebolt. For those pilot projects, we shipped product across the country. But we're currently in discussions with a couple of developers who opened up factories out on the West Coast and we'll announce those later this year. So tell me, where does the robotics component then here come into this story?

40:36Aaron Tilley:I mean, is it insofar as how you build these blocks? Is it assembling them? Where do the robots come in here? Yeah, so today it's in component manufacturing. So our robots are actually fabricating the wall panels and ceiling panels. And we have a roadmap to start doing higher value work in addition to getting insulation, getting your wiring components, and getting plumbing into the walls. So over time, we'll see about 65 plus percent of work content robotically fabricated. And we'll still have a highly variable human component. Our goal is to be able to have builders of all skill levels. So we've had high school co-ops come in and do plumbing and wiring work.

41:15so the goal is how do you generalize the rest of construction so that we can bring in a new workhorse in that's following screen-guided instructions instructions that we print directly onto our wires onto our studs, onto our sheet codes so that they're assembling things it feels more like assembling Lego bricks and assembling your IKEA furniture but built to a much higher quality built to code homes that allow us to then truly transform what has been figured things out on the field construction work for the last few hundred years into high-tech manufacturing. We've got teams of people and robots working alongside each other.

41:50Aaron Tilley:And so tactically, I'm imagining, I mean, this is like, it basically looks like a series of arms, basically robotic arms, the same way that Amazon has its arms and its warehouses and factories and stuff like that. That's kind of what it looks like, the facility. It is. And we've also taken a stance where we have no concept of an assembly line. We believe that this is a high-mix manufacturing problem. A lot of prior attempts have misclassified this as a mass production problem to try to really get to a very capital-intensive mega factory. What we've done instead is picked traditional off-the-shelf industrial arms, but built our own vision system.

42:27A lot of vision algorithms that allow us to have a$200 ,000 robot perform the work of a million-dollar robotic system in a much smaller footprint. And you basically have a collection of these robotic arms able to build different components off the home over time.

42:40Aaron Tilley:And is that because of it? So when you say your own vision system, does that mean that you guys have created your own model that the company is, that the robots are running on? Or is a vision system different than a model? I don't know the details here of the software components. I mean, all the cool kids have models today. We're in the process of still training our vision systems and classical methods till we get ample data sets to eventually say, okay, now we can actually train our own full-state model. We're not limited by a model's capability to achieve our objectives in our near term. Our end-state objective is we're creating enough body of work and enough data sets here that we will get to build the right type of system that allow us to continue leveraging off-the-shelf specialized hardware but be able to perform a whole bunch of tasks that weren't designed to do just because we've created the right environment to do that.

43:36Aaron Tilley:And so last question for you, Vikas. I mean, I've asked this to a couple other robotics founders on the show. Insofar as how you see robotics developing and I'm sort of specifically talking maybe about the model layer here for robots and, you know, world models, real world AI, physical models, stuff like that. You know, we sort of live in this world where we saw the big AI labs dominate. They are still dominating. Now we are in a world where open weight models are becoming more popular. How do you see robotics software playing out? Do you see there being a couple of big giants that have the models that everybody uses, that they're, you know, closed source models?

44:23Aaron Tilley:Or do you see it being a much more fragmented story for physical AI than it has shaken out to be with the current AI era? Yeah, that's such a good question. I don't think I have the perfect answer for it, but this is my framework. A robot's a tool to solve the problem. So we're really obsessing over the problem, which is how do you industrialize home building? And then the goal becomes how do you build the right set of tools to achieve this outcome? And today, the problem set-search requires us to constrain the problem in the physical world. So you actually have high leverage before you're constrained by what's available in the digital world.

45:01So if I sort of forecast this over time, I think there's going to be a lot more work done in defining the problem appropriately. Picking the right solution architecture and then looking out to see, are there ready-to-use models that allow you to save a whole bunch of enduring time on solving the vision problem or the manipulation problem? And you pick one of those models. or you then decide actually there isn't a model to do that and now you have to own your outcome and you're going to spend the time building either tuning an open source model or building your own model from scratch. So to me, there's a whole set of outcomes out there.

45:32The reality is, right, take this analogy of a construction site where you'd expect just one single manufacturer to be successful in sort of getting all your earth moving equipment, right? It's a well understood problem. has been happening for 100 years, but you still have a dozen different manufacturers, even the models of components that have vary significantly. Like it is fragmented by nature because the use case actually becomes really nuanced based on size of project and type of work you have to do. In our case, we're sort of taking a lot of the world model constraints out because we're really constraining the problem into a factory environment.

46:10And the next piece for us then comes down to play is to say, what else do we need to constrain next? My whole experience and own experience from Amazon Robotics has been really constrain the problem to have maximum leverage in applying automation.

46:23Aaron Tilley:So you're saying basically the diversity of the problems in the real world that robotics has to solve necessarily means that it will be a very diversified field and a very fragmented sector of models. Is that the idea? Short answer, yes. I think it'll stay fragmented for a while. And the primary reason for it is that the robot's not the solution. It's a tool in the tool stack. And depending on who's designing the overall solution, you'll end up picking the right set of components to come together. Great. Well, Vikas, I want to thank you for coming on. That is Vikas, Antti, co-founder and CEO of Reframe Systems here on TI-TV.

47:09Aaron Tilley:That does it for today's show. A reminder, we are on this stream Monday through Friday at 10 a.m. Pacific, 1 p.m. Eastern. If you can't make it then, episodes are available on theinformation.com, on our YouTube channel, or wherever you get your podcasts. Make sure to follow us on social media, on X, on Instagram, on TikTok, and on LinkedIn. I am already excited for our next show tomorrow. Have a great rest of your Monday. Bye-bye for now.