Musk’s xAI Loses Two Co-founders, Why AI Automation is Different, Wealth Management Stocks Fall

0:13Welcome everyone to the information's TI TV. My name is Akash Pasricha. It is Wednesday, February 11th. First up today, the exodus at XAI continues. Two more co-founders of Elon Musk's AI company have left. We'll bring on our Elon Musk reporter to break down what's behind the departures and what they signal for the company. Next up, our financial analysis columnist has a piece out about ServiceNow. She has taken the view that the company is undervalued at the current share price. We'll talk to her about some of her analysis. We'll also talk with Matt Schumer about his viral article on X about why AI represents a different wave of automation than we've ever seen before and what the implications are for labor and for work.

0:57And finally, the SaaSpocalypse seems to be spreading to the wealth management sector. A new AI tool is spreading fear across that arena. We'll speak with wealth management executive Ross Gerber about how he thinks all of this will shake out. It's going to be a fun show, so let's get right on into it. Two more co-founders of XAI have left the company. Tony Wu and Jimmy Baugh said on Tuesday they were exiting Elon Musk's company. It's not uncommon for there to be a lot of turnover within the Musk empire. Joining me now to break it all down is Theo Waite, our Elon Musk reporter. Theo, welcome back to the show.

1:32It's great to have you here. Hello. Who were, who are Tony Wu and Jimmy Baugh? They were two, you know, everyone at XAI has the title member of technical staff, pretty much. So technically, they were both just members of technical staff. But they were two of the 12 founding members of XAI. And they were, until yesterday, relatively senior in the engineering org. And they're now gone. Now, there were like 12 co-founders to start with, though, right? There were 12 and we're down to either six or seven currently, depending on how you count it, because one of them is technically on leave. But 50 % of them are working day to day there.

2:21Okay. So 50 % of the co-founders, which it was a pretty big group to begin with, are gone. What do we know about what's going on here and why these people are leaving? So my understanding is that, you know, around the same time that Elon decided to merge SpaceX and XAI, he expressed some dissatisfaction in pretty sharp terms to the XAI leadership about the state of progress on Grok. um they released grok 4.1 which is their newest model late last year they were supposed to release grok 4.2 um you know at the beginning of January or end of December ish and that did not happen uh unclear exactly why you know whether it's training delays or or it was ready and you know it just wasn't good enough to release um but for whatever reason that got delayed and And Elon doesn't like things getting delayed.

3:19And when he is angry or focused on a project, heads will roll. And it seems like that was the case here, which is not totally surprising, but kind of a quick reversal because Tony Wu, who was one of the two co-founders who left, was promoted and got way more responsibility less than a year ago. So there's a lot of ups and downs and twists and turns with leadership at this company, even more so than his other companies. Well, but I was going to say, I mean, there have been departures at other Musk companies this week as well, right? Yeah, there was the departure of Raj Jaganathan at Tesla, who was there for 13 years and recently had his role expanded to include sales, which, you know, was kind of odd because he had not had sales experience before.

4:13And, you know, that could have been why he left. but you know tesla has been generally struggling with with a downturn in demand for evs in the us and overseas and increasing competition so uh you know being head of sales at tesla was going to be a hard job no matter who was doing it now all of this has sort of prompted the question for me who are the executives that musk may actually not be able to afford to lose in some some ways i mean And members of the technical staff, again, researchers, they certainly are hard to come by in terms of the top talent. But, I mean, Gwynne Shotwell over at SpaceX is someone that we have written about extensively here at The Information.

5:00Last night in the briefing that you wrote, you sort of highlighted her as someone that Musk could not afford to lose. Why is that the case? Yeah, I mean, you know, to be clear, I have no reason to expect that she would— immediately leave or anything like that. But with the SpaceX IPO, you have to think about who are the critical people at this company besides Elon, and she is by far the most important one. She's been the president and COO for a very long time. And by all accounts, is someone that knows how to take Elon's grandiose impulses and overstatements and enthusiasm and energy and direct it into an actual structured organization.

5:46And, you know, when you're trying to take a company like SpaceX public, that's more important than ever. So, you know, I think she's somebody that if she were to ever leave, that would be a big crisis for the company. Great. Well, Theo, I want to thank you for coming on. That is Theo Waite, our Elon Musk reporter here at The Information. One software company whose shares have dropped drastically is ServiceNow. The company's stock price has fallen roughly 50 % in the past year, but the information's financial analysis columnist, Anita Ramashwamy, has a piece out today arguing that it has fallen too far and this could be a buying opportunity for investors.

6:24I want to bring on Anita to help us understand how she is thinking about all this. Anita, welcome back to the show. It's great to have you here. It's great to be back. Why do you think that ServiceNow is undervalued? So, Akash, I think ServiceNow is undervalued because if you just start by looking at the numbers, it is trading at the lowest multiple of expected forward sales in the past decade. That's also true if you look at its forward EBITDA multiple. And so the market has really sort of left ServiceNow for dead. But at the end of the day, the fundamental metrics that ServiceNow has been reporting are really promising.

6:59Yes, there are some concerns about their progress in AI, but at the end of the day, they're making inroads, they're increasing revenue from AI, and they are pretty well positioned when it comes to the AI stack, which I can get a little bit more into. So I will say that on this show, we had Tomáš Tunguz on, I think last week, after ServiceNow Report Earnings. Him and I had a discussion about exactly this, and we both sort of took the same view that you did, which is that the stock has fallen too far, And we didn't have a lot of the data that you had, which is what I want to get into. But, you know, I think that there are some arguments to be made.

7:34Of course, I want to sort of look at this issue comprehensively from all different sides. And so walk me through some of the data that you found to back this and some of the common pushbacks from investors that you've heard over the course of your reporting that have concerned them. Yeah, absolutely. So I think the first important thing is that ServiceNow is growing really fast. They've managed to maintain a 20 % plus top line growth rate for several quarters now. And that's even as they're sort of transitioning to including AI capabilities and AI products in their stack. I think perhaps the most impressive metric, if you look at ServiceNow, is really the free cash flow margin.

8:12So they have a free cash flow margin of 34 % if you look at last year. And that was about$5 billion in free cash flow over the year. And I looked at 70 different companies in Bessemer Ventures Cloud Index. So these are all publicly traded software companies. And ServiceNow only had five other firms that ranked higher than them on the free cash flow margin metric. So they are printing cash. They're doing it really efficiently. And I think a lot of the common pushback that investors tend to levy against ServiceNow is that they're not doing enough in AI, right? They're this big enterprise sort of legacy software business.

8:48And there is this concern about how are they going to be able to adapt? But we've also seen that their customer retention numbers are looking really good. We've seen positive trends on all of those fronts. And I just think it's far too early to discount and to write off the idea that ServiceNow can get ahead in AI. I mean, they have the balance sheet to do it. They have the firepower. They've been making acquisitions. And they've also been investing a lot in selling their existing products with that sort of AI overlay. What has CEO Bill McDermott said about the acquisitions that the company has made?

9:22Because they've come in succession and that's one reason that some investors, analysts have sort of said, hey, why are you doing all these deals in this very experimental time for AI? What has he said about all that? Yeah, so he sort of addressed this on the last earnings call. I mean, we've seen a couple of large deals from ServiceNow recently. We saw their purchase of Moveworks, their purchase of this cybersecurity firm, Vesa Technologies. And we saw that they bought this cybersecurity firm called Armis, which they announced that deal right before Christmas and said they were going to do. And so they have gone on this acquisition spree recently.

10:01And what Bill McDermott said about this is last month, he said that he didn't really see ServiceNow having to fill any other, I believe he said, large white spaces when it comes to their platform vision. So he was basically trying to reassure investors like, look, we've already done these acquisitions. There is going to be incremental growth that comes from us having these products, and we don't necessarily have any other big spaces that we need to fill. And I think that's valid to some extent considering that they've just gone on this acquisition spree, and we have to see how that's going to contribute to their top line.

10:30And it is kind of funny to me that this narrative has come up recently that, oh, ServiceNow is making acquisitions. That's a scary thing because in another time, you know, if the markets weren't what they were, I think you might look at a company with these metrics and with this balance sheet. And you might ask the question, well, you know, why are they making even more acquisitions? You know, like this is a company that that, as you said, growth is strong. Profitability is strong. i almost feel like maybe the company feels like we have to take our foot off the gas the acquisitions just to sort of satisfy the the broader investor concerns right now but you know i think maybe there could be more coming um you know i want to i want to ask you a little bit about the broader sas sector right now so this is one company that you identified as having fallen too far what are your thoughts on the the sell-off more broadly across the sector i think that you you know, the sector has been painted with a really broad brush.

11:31Everyone is just, you know, kind of willy-nilly selling off the software names because they're concerned about this AI issue. I mean, to your point with ServiceNow and acquisitions, I just wanted to sort of bring it back to that for a second because part of what makes ServiceNow so well-positioned compared to all these other software names is actually the fact that it's really hard to do good M &A and ServiceNow has done good M &A over and over and over again. And so, you know, when it comes to some of these upstarts, you see these companies like Serval or companies that are competing in ServiceNow space that are really hot.

12:03And at the end of the day, which software companies are going to be well positioned to either build or to buy that technology and make sure they're adapting to this new world? I think ServiceNow is certainly one of those names. And I think when investors are looking at software companies, that's exactly what they should be asking themselves. Is this a company that has really fallen behind and it's too late and they don't have a clear path to be able to catch up in AI? Or is it a company that they have the existing advantage because of the distribution that they already have of their software? Do they already have customers that they can then use those AI capabilities and sort of roll those out, increase revenue per user, that sort of thing?

12:39So I think that investors are painting software with too broad a brush, and I think they really need to peel back the curtain and look individually at which companies are investing in the right areas of growth for AI and are generating free cash flow. And last question for you, Anita. You talked a little bit about this in your column, but it has come out across the sector too. This question of, well, if people are employing less at these software companies, the software companies that are buying the software, if the headcount is lower, what does that mean for the number of seats, for example, that they have to buy subscriptions for?

13:16What have investors told you about that concern and how worried they are about that? Look, Akash, that's a very real and valid concern. A lot of these software companies, including ServiceNow, have historically been reliant on selling on a per-seat basis. So fewer employees, because AI takes over a lot of more corporate work, means fewer employees to actually buy that software. So fewer seats for companies like ServiceNow. But first of all, the companies are transitioning. ServiceNow has been working on this dynamic sort of hybrid pricing model and introducing more and more features that are charged on a consumption basis.

13:49What I've heard from investors is that that consumption-based pricing aspect could actually be more lucrative for companies like ServiceNow because a lot of times customers will sign up and think, oh, we're not really going to use this technology too much. But if it's good technology and if it's actually driving business improvement, which at least in theory a lot of these AI tools should be, then they can actually charge more. And at the end of the day, the company will end up making more than they might have in a seat-based world. So we have yet to see whether that'll play out, but that's the thesis that a lot of bullish investors in service now are taking.

14:22Great. Well, Anita, I want to thank you for coming on. That is Anita Ramaswamy, our financial analysis columnist here at The Information. One of the buzziest articles on X over the past 24 hours has been an essay from Matt Schumer, GP at Schumer Capital. Matt took a step back looking at this moment for the AI boom, why people might be underestimating the impact of it and what people can do to get ahead. And I want to bring him on to talk all about it. Matt, welcome back to the show. It's great to have you here. Thank you for having me. So, Matt, why did you write this piece? And also, more importantly, when did you write it?

14:56Because I texted you last week and you said you were sick all week. You were stuck in bed. You couldn't even get out of bed. I told you I said immunity broth would be the solution. I don't even know if you got it. But when did you write this essay that has gone completely viral over Twitter? Yeah, so it actually started on Thursday night and I continued it through the weekend. The reason actually was I was home with my family and I was trying to find the words to explain what was going on to my parents because frankly, I do feel like we have hit an inflection point. I was struggling and I started looking online for things I could send them.

15:33So it wasn't, you know, their son talking to them, but like, you know, folks that they might know or folks in the industry that, you know, they can look at and say, this isn't just my son talking to me. And frankly, I came up empty. I was immediately drawn to Dario Amode's essays. But while they're utterly fantastic and like probably the best things written in the industry on this stuff, they're not meant for the average person who isn't in tech. And I realized that I needed to find a way to explain this to them that they would understand without having to learn all the jargon because people shouldn't be forced to learn all the jargon to know what's going on.

16:04They shouldn't be forced to understand the dynamics of training, for example. My idea here was really just, can I put something together that my mom would understand, my dad would understand, that takes the ideas from a lot of these pieces that are utterly fantastic and views some of my own thinking and my own ideas and how I think this is going to move forward and what I think people should do and how they should prepare, and then explain it in a way that just makes sense, that isn't forcing them to sort of understand extreme levels of jargon and similar. Yeah. So we're going to get into the different elements you talked about here.

16:36but if you were to just summarize the entire essay in a single sentence, what would be the message you would be trying to send here? Yeah, it's tough because there's a lot there. But I think it's just, it's going to impact everyone differently and different timelines and sort of in different amounts. Right. But people should stop and think at a minimum. Right. This might affect me. Let's just take it seriously for a second. If it does, what will it look like? What does it look like if all this hype around AI not working actually isn't true, just in case it isn't? What should I be doing? What will my life look like in three years if I do nothing?

17:15What will it look like in three years if I decide to do something? I just want to get people to think because all of the content right now is for the in-group. And that's not a good thing. It means that the folks that aren't in the sort of, you know, in AI already aren't hearing this message. And I just wanted to spread the message, right? It was originally written for my parents, but I realized this could probably help a lot of other people. I didn't expect people to take it as well as they have. I frankly thought it was still a little bit too technical. Clearly, it's kind of broken out and gone way further than I expected.

17:47But the message is just take a second and think. This may not affect you. This may affect you. Just in case it does, what should you do? Why do you think that we are in the middle of another inflection point? You pointed to OpenAI's GPT 5.3 Codex and Anthropix Opus 4.6 as two pivotal releases that changed the game in a way that even people who say, well, I've tried AI before, it wasn't that good, in your mind, makes that argument obsolete. Why did these two releases change the game for you? Yeah, it's in particular mostly 5.3 Codex. I've been extremely impressed by the model and I had early access to it to test it.

18:33And it was hard not to talk about it because it was the first time I felt like, you know, and I write about these models every time they come out. I often am very impressed because they're always better than the last versions are usually better than the last versions. But I'm still kind of like working with the AI and iterating with it and finding issues. 5.3 codex felt like the first time it could just be like, okay, here is the thing I want you to do. Here are the tools I'm going to give you access to to do it. do it. And I will literally come back hours later and it's just, it's, it's, it's perfect.

19:03And it's, it's a little terrifying for me. And I think I don't want to sort of like amp up people and get them feeling sort of negative, but I think it's important to know that this stuff is, is coming in the way that I think about it. Right. AI models are kind of like a year ahead in coding of everything else where it is in coding today is where sort of everything else will be a year from now. And if we're seeing this today with code, right. We expect that about a year from now or possibly even sooner at this rate of progress, this is going to translate to everything else. It was just such an eye-opening moment to see that model go from sort of just simple prompt to this thing just works without me in the loop at all.

19:37It just tests everything on its own if you give the right tools and it just builds incredible software. I think that same thing is going to happen for most other industries over the next year or so. That being said, it's going to take longer to actually sort of proliferate through the economy just because there's a capability doesn't mean it's actually going to sort of be made real in terms of like enterprise adoption immediately. But in terms of capability, it's going to be there. But you made the argument that AI, for this reason, what we're seeing with Codex 5.3 and with Opus 4.6, you made the argument that this is different from other previous waves of automation.

20:12Why is that the case?

20:18It's a little hard to put into words. A lot of this is. But the way that I think about it is previous waves of automation sort of were like productivity boosts. And early versions of AI, I mean, I would say up until like, you know, a couple versions ago, pretty much productivity boosts. Though they've been increasingly moving towards the sort of end of the spectrum that I've been talking about. Now we're getting to models that can do what a human can do on a computer. And it's not 100 % there yet, but it's getting there, right? Right. You know, if you were displaced in a previous wave of automation, right, you can move and do something else on a computer.

20:54Right. You know, there's there's so many opportunities for that. But when an AI can do anything a human could do on a computer, and I think that's coming, it's it's a very different questions. What can we do that an AI can't do? And the answers are slimming down each time there's a new release, if that makes sense. Okay, so what can we do then? The essay paints a bit of a stark picture on what people will do for work. And so, and I think I know you to be a pretty optimistic person. So what can people do then in this environment? I think it's different for everyone. And again, the idea of this article was to just get people thinking.

21:29I don't think there's a single silver bullet. I don't think there's a single pen of steel. Like it's just, you need to think about what it means for you and your job. And I think people can't go into it blindly. they can't say, oh, this isn't going to do what I do. They have to think of the sort of sense of like, this will, so what should I do? And if it doesn't, great. But just in case, they should be thinking about it. Right. So in terms of what people should do, I would say, just take your time to think about this and then say, okay, if this world is what this sort of like article paints, what should I be doing three years from now that can differentiate me?

22:03Should I be owning more? Or should I be trying to start a company so that I can benefit from it more than if somebody else owns and I'm working for them? Should I be thinking about building a social media following so that I can distribute things that I build? There's a million different answers. And I think the question can go in a million different ways for a million different people. That's why it's really hard to answer. But you made the point. You said people need to learn how to be using these tools even more than they already are. And one thing I thought was interesting was you talked about education.

22:34and I don't think you were trying to say that this should be a substitute for education, but I think the point you were trying to make was that really, I mean, you know, kids need to learn how to be, how to integrate these things into their everyday lives so that we don't fall off in terms of skill sets. Yeah, and again, no one knows what's happening. And I think that's part of the point and part of the problem. Like it's easy to articulate where we are today and where it's going to go over the next, you know, couple of years, but it's way harder to say this is what the world's going to look like in five years because it's just changing so fast and you can't see through the fog as a lot of the leaders in the space like to say.

23:08In terms of education, right, I think it's still very hard to know what the right thing to do is, but then you can only guess. And if I had to guess today, I'd say kids should be learning to be extremely adaptable, learning how to use these tools to make what they want to come true, come true. And, you know, I think until we have more clarity on what this world is actually going to look like for real, not just in hypotheticals, it's hard to make really concrete decisions, right? But I think you can start thinking about it today as like take current education and start kind of optimizing it more around like what is going to be possible, right?

23:41If a child is learning without a calculator, right? They're sort of teaching themselves for a world that's not the real world, it's an old world. And I think the world is going to be, this is going to be everywhere. You need to learn to use it, to work with it. And again, you may not need it. It may not become this big thing that we all expect it to become, but you want to be prepared in case it does. And last question for you before we let you go, you had something like 15 ,000 people reposting this article. What was the funniest response, the funniest message that you got from folks? It's funny.

24:15I don't actually think there was anything too crazy with this one. Normally when I post something and it goes relatively viral, and this is a whole other level and I did not expect that, I get a range of things. Like from congratulations to death threats. and everything in between. This has been more of a consistent, wow, I didn't realize this was happening or wow, I need to share this with the people in my life, which I think is a good thing, right? The whole idea of this was not to scare people. It was just to help them understand this is real. And I actually think that it hit the mark there, which I'm very, very happy with because - Well, and you very much put into words, I think a lot of the uncertainty that people are feeling in this moment.