Musk vs. Altman: The $109B Legal Threat, xAI’s Desperate Coding Push, Will Meta License Gemini?

16:43Drew:Yeah. So, well, let me put something else out. I think that the next wave of consumer products, it's not going to be about who has the best model. It's going to be about who has the best taste. and especially now, everything can be vibe-coded. The barrier to building is going to be super low. And so to build something really great, you have to really understand the consumer. You have to understand what they want and you got to make something 10X or 100X easier, right? So a good example of this is, you know, taking a look at what LMs can do. They can read and summarize a lot of stuff. It can now start to take action, which it couldn't do 18 months ago.

17:20Drew:and then you're going to see some of these companies coming out, like editing down the workflow and the product into something that is so tasteful that people want to use it. Again, at the end of the day, consumer is about appealing to the mass audience and appealing to the mass audience means that you have to do so to the lizard brain within all of us, which means you have to make a product so simple that in hindsight, you're like, oh, that was obvious. Of course, Instagram should have existed. There needed to be something to share visually. It's obvious in hindsight, but it took Kevin and Mike's taste to really get that product exactly right, editing all the stuff out that doesn't need to be there to create something that people love using.

18:01Drew:And that's what we're going to see this year is that there's going to be a lot of companies coming out where the model is not the differentiator. It's the workflow, it's the taste, and it's the choices in the product.

18:12Laura Mandaro:But let me ask you this. I agree with you that the models are, once we stop talking about the models and how good they are, that's actually when consumer AI can take off. And I agree with you on the taste part, but are the models good enough yet to really have the mass adoption, again, outside of chatbots, but I'm talking about agents doing the things that we really are asking them? Yeah.

18:37Drew:I think you have to, if you're building a consumer company today, you have to assume that the models are going to get better and you're going to start building and skating to where the puck is going. So I'll give you an example. I work with a company called Extra and they are building sort of this amazing new experience for email. And they're treating email as sort of a backend of your life and things that come in and the communications. And they're able to apply this agentic layer on top of it to start doing things for you based on your email. and the founders are have incredibly high taste they were the early team at pinterest they understand what it means to build for you know the masses and sort of all the rest of us and you know right now the product is in beta i've been using it every day as my primary um evil email app right now and it really just helps me manage my life and you know when you get into that product you really start to see that the models are capable of a lot more than you think of but it's the creativity of these entrepreneurs who are building in consumer, who really understand how to use the model in a way that just feels magical.

19:39Laura Mandaro:So in that example, with that company, there's not been a scenario where they have wished that the model could do a little bit more to achieve that taste that they have.

19:51Drew:Oh, of course. Every day you wish that the models can do a little bit more. Is there an example that you can give? Yes. And let me just finish the cross real quick. I think in the past, you can see like, yeah, they wish that the models could do more. But then the next week or two weeks later, the models can do that. Right. I think that's what I mean by the entrepreneurs. And I'll give you a specific example in a second. But that's what I mean is that the entrepreneurs building consumer need to be skating to where the puck is going. And Sam Altman's right. Like you have to build assuming the models are going to get better.

20:24Drew:And then if you assume that and you're able to figure out what is the hardest you put around that agent um it can do some really interesting things so um you know every every you know every morning i actually get this push notification from extra that's like hey here's what's interesting about your day that's not just uh uh you know hey the road take my kid to school it's like this is the one special event like reminder you have to you know you have to be on tv today early that type of thing um can can be you know done now because you can start to understand what is the pattern of what's normal?

20:57Drew:And then what are the things that are actually special on top of that? And the company, it's early, but already I talked to Naveen, the co-founder, on Monday, I want to say, and we were already talking about the next set of features. And I probably shouldn't say too much, but the models are definitely getting good enough and you have to just go and assume that they are.

21:16Laura Mandaro:Right. I want to ask you about Sam Altman and about just the OpenAI product portfolio at large. I mean, look, there are some people that say that, hey, it's a bit unfocused and there's too much going on. Do you agree with that?

21:32Drew:I think that the greatest entrepreneurs are the ones that are maximally ambitious. And it's hard to say that, you know, Sam's anything but ambitious. Also the team there having been there is exceptional. All my friends that are still there are just like, you know, kind of top of their game. And so if you take a look at the team, the talent and the ambition, I believe they're capable of amazing things. And, you know, the other thing is that I think people often see, you know, when big companies, you know, put something out there and pull it back, you know, it's like a sign of defeat or whatnot.

22:04Drew:But honestly, I think the best companies are the ones that experiment and are willing to, like, be bold and try something. And then when they go out and fail, they just say, great, on to the next one. And so I'm not saying that there are failures. I'm just saying that that iteration speed is what's going to make any company a defining company.

22:23Laura Mandaro:Let me ask you one more question about the AI ecosystem at large here. I mean, we have all of these consumer tools that people are working on. We have the chatbots. And I think there's a question around how many different chatbots we need and whether or not that is the way to get adoption. On the other side, you have the data center buildup going on. And we had Josh Wolf from Lux Capital on the show yesterday. and I was talking with him about the data center build out and the threat of overcapacity and maybe the likelihood that, hey, maybe some of these data centers don't actually need to get built.

22:55Laura Mandaro:And of course, the hyperscalers will say, there's going to be enough demand. That's not the problem. I mean, do you see it that way? Is there not a threat here that there is not enough demand?

23:07Drew:There is plenty of demand. I was doing a reference call with the CTO at a Gen AI company, I think on Tuesday, and it's very clear that he's like, oh, the value of these chips are going up because there's just not enough of them. So absolutely, there's enough demand. We haven't even seen the peak of the demand when it comes to video, right? We're talking about text primarily. And video, as the models get better, they're not great yet, but they're getting much, much better as the models get better there's going to be increasing demand of uh for for ai generated video in in all different respects and i think that's one of the things that sometimes people underestimate is that the demand today is based on today's technology and what is available but going forward the demand uh is going to expand and the market is going to change in terms of what it's demanding and what it wants um and i think that that we haven't even seen the tip of the iceberg in terms of um what folks demand in uh in video so another one our portfolio company is runway AI I was talking to their uh co-CEO a couple weeks ago and it's it's very clear that the business is not only doing really well but uh I think that the way that people are perceiving um uh sort of the the the need for for more visual content is increasing and we're only

24:27Laura Mandaro:kind of getting started there so and so I I imagine that you're an investor in runway sounds like you're excited about all things consumers. So Sora and the concept of social media with AI-generated content entirely, you're bullish on that.

24:48Drew:I'm not sure if I'm bullish on exactly that sort of vector of a video. It's more just that I think that we are going to... Well, I think that it was a really interesting experiment to see sort of what happens when you're able to generate these videos. I think that one of the things that's interesting about consumer actually makes it really difficult, makes consumer really difficult, is that our bar as humans keeps on increasing. What we find interesting today, we may not find interesting tomorrow. And that's been what's really difficult and why I think platforms like Facebook and Instagram have been enduring is because they've been platforms on which sort of different tastes and different, you know, memes can come and go.

25:31Drew:And so I don't really have an answer for exactly what humans are going to demand. But once you take a look at sort of the ever-evolving bar, I'll give you a specific example. I think, you know, two and a half years ago when we were testing the voice mode in ChatGPT, I think the world was just blown away. We were blown away about what the models could do. And then I think 12 months later, we're like, wait, this sounds robotic. And you wouldn't have said that about, you wouldn't have said that about chachimi voice when it first came out it did not sound robotic it sounded the opposite of robotic right but our tastes and our uh desires as as just needy humans of ever-evolving taste we we put the bar higher and now you have models out there like sesame that are just really doing well in terms of the conversational bit but even that is like kind of our detection of the uncanny valley is getting better and better every day great well peter i

26:26Laura Mandaro:I want to thank you for coming on. That is Peter Deng, general partner at Felicis here on TI TV. Startups hoping to go public this year will have to navigate a number of headwinds, and yet there are still a bunch of companies that are marching towards their public stock market debuts. The Information Publish exclusive reporting on who is where in the process. I want to bring on our managing editor, Laura Mandera, to discuss that story. Laura, welcome to the show. It's great to have you here.

26:53Akash Pasricha:Hi, Akash. Happy Friday.

26:55Laura Mandaro:Happy Friday. Okay. So Valida, our colleague, put out this story with a lot of names of companies that are marching towards their respective IPO, hopeful IPOs. Some of the companies are maybe taking a step back. But what were the most interesting companies that really caught your eye as you were editing the story?

27:19Akash Pasricha:Well, it's kind of a grab bag, right? We had Huddle, which is a video technology that is sold to sports teams so that parents and coaches can sort of look at their student athletes' performance. We did a very nice feature several weeks ago about the rise of this, the demand for this kind of technology. So, I mean, this could be pretty much as far afield as you could get from the kind of enterprise software startups or, excuse me, enterprise software companies that have been so hit in the public markets. And I think that speaks a lot to where there is perceived maybe space for companies that are not going to be, you know, sort of swept up in this worry about AI.

28:13Laura Mandaro:So there's space, but then there's the SpaceX factor of all this, which is that we've got the big three, SpaceX, OpenAI, and Anthropic that we know are marching towards their IPOs. I mean, talk a little bit about the prospect of them sucking all the air out of the room and what that could mean for these smaller companies.

28:32Akash Pasricha:Yeah, I mean, it's there's, you know, one investor told me yesterday or earlier this week, I mean, there's only, you know, some of these mutual funds, you can only buy so many stocks, you got to make room for some, maybe you're going to sell some stocks, maybe you're going to buy some new offerings. So I think there is just this anticipation of tens of billions of dollars in new shares being floated. Now, OpenAI and Anthropik could easily not go this year. I think the earliest would be the end of the year. But SpaceX itself would be a record size, and that seems to be headed towards a summer IPO.

29:15Akash Pasricha:So there's that. And, you know, there's a kind of minutiae tactics of when you actually hold the IPO. And anybody who's covered this space a little bit knows, I mean, there are certain periods when you can't do it, right? You can't do it right around election and, you know, the holiday period. And so these windows get narrower and narrower. And if you don't want to do it right around a mega IPO, that just takes out one more, which is pretty different. And if you think about like 21, 22, we just had this like, you know, just week after week of really heavy IPO issuance. And it does not look like that's going to be this year.

29:53Laura Mandaro:And is there any evidence that the performance of recent IPOs or the IPOs of the past couple months, Corey Weinberg, our deputy bureau chief of financing, he had a great story out, I think, a couple months ago, just outlining all the performances, and many of them have not done well. that's not scaring any of these any of these companies i think it is um i think i think it's exactly the case i mean just a couple months ago um you know we had a lot much longer list of ipo

30:26Akash Pasricha:contenders and obviously they could go public at some point but i think any of the enterprise software companies that and there are many many of them i mean we only listed i think five companies not all of whom are startup, all of which are startups, that are, you know, taking steps. And that doesn't mean they're going to have a public offering this year, but they're doing those sort of concrete actions that's set up for an IPO. They're talking to investors. They're, you know, hiring bankers. But there's many, many more that are old enough and maybe even have the revenue to go public. But if they're in this enterprise software sector, I think it's very hard to go.

31:09Akash Pasricha:You know, and then there's the, we didn't talk about the category that are seen as beneficiaries. So Cerebris is the best example. And probably I should have started with that. I mean, they are an AI chip designer. They have very big contracts with some of the model makers. And I think they would hope to be one of the ones that is seen as on the rising swell of this AI tide rather than the victim.

31:38Laura Mandaro:And then you also have the private equity firms. I mean, we mentioned Anaplan and Genesis as two names that are very much looking for a window here. And, I mean, you've got all these private equity firms that are hoping that the SaaS apocalypse will give them some kind of – maybe there will be a gap or something that they can just take their window and run with it.

31:58Akash Pasricha:Yeah. I mean, the last couple of months, people have been telling us that everybody doesn't really know where the bottom is and how it's going to settle out. And so that is not – and I think that's both for – that's very much about the public stock market and how, you know, whipsaw it's been by every, you know, product released from Anthropics seemingly. But also just the rate – the pace of development and whether, you know, even if we're talking about M &A, can these enterprise software companies, they're going to buy something that makes them seem more relevant. Are they buying the right thing, right?

32:35Akash Pasricha:Like, is AI agents going to solve this, you know, this perception that they don't have enough AI to be competitive with entropic or open AI? And, you know, maybe entropic or open AI are going to, you know, develop something else that makes them seem uncompetitive. So there's just this turmoil. And until that resolves a little bit, I think it would be very hard to go out as an enterprise play. but you know the clarity may come i mean if if we see anthropic for instance go public i mean we've reported on a lot of their financials but there's a lot that hasn't been reported that would give public market investors a much better sense of um you know the competitive nature of these companies

33:22Laura Mandaro:and if it's if it's not doing as well as people think you know on all different metrics then that could certainly help the software companies looking for a window.

33:33Akash Pasricha:Akash, I know you want to know what the net retention rate is for Anthropic. For Anthropic.

33:39Laura Mandaro:Well, that's why we have our reporters working on it, right?

33:44Akash Pasricha:Exactly.

33:45Laura Mandaro:Net retention, gross retention. We want all the magic numbers. And so with that, I will leave you to get back to your Friday, Laura. Thank you for coming on. That is Laura Mandero, our managing editor here at The Information. Meta's newest AI model has been facing some setbacks, according to a New York Times report, which also said that Meta was considering licensing Google's Gemini model in the short term for its AI features. That was the latest in a string of big tech headlines this week that also spanned NVIDIA, Oracle, and data centers. To unpack it all on this week's Editor's Cut, I want to bring on our co-executive editor, Martin Pierce.

34:23Laura Mandaro:Martin, welcome back to the show. Hey, Akash, how are you? Happy Friday the 13th.

34:28Peter Deng:I am missing my partner in the editor's cut. It's just you. It's really, really quite sad. It's just us.

34:37Laura Mandaro:It's just us. We're here. It's Friday. Nobody is paying attention.

34:42Peter Deng:I can say anything, right?

34:44Laura Mandaro:You can say anything. And you can say anything you want about avocado, the codenamed model. I like avocados.

34:51Peter Deng:In fact, I got a sandwich today that has avocado on it. But yes, the New York Times story last night was quite an important story, I think, because it really captures important development that we are beginning to see in the AI world, which is that, you know, you've got five companies right now that are developing the leading AI models. That's Google, Anthropic, OpenAI, Meta, and XAI. But it does look like Meta and XAI are both having problems. We've got this time story on better not being happy with the latest version of the model. Elon Musk has been fairly transparent in his tweets that he's unhappy with the staffing at XAI and he keeps firing people.

35:44Peter Deng:and he's just tweeted yesterday, I think that XAI was not built right, which makes you think that things are not going well there. And there's always been this issue of, is there enough talent in the world to support five leading companies just in the US? Because China obviously has their own. And I think what we might be beginning to see is that the beginning of the end for Meta and XAI, they may have to pivot. Maybe they will focus on models which are not as good as the best, but maybe that's not so bad.

36:28Laura Mandaro:Well, and so the Google part of this was kind of interesting too, the prospect that Meta could turn to Google, maybe in the short term, and this is a huge win for Gemini. They have Apple.

36:40Peter Deng:Well, that's something I should say, though, And, gosh, that is something we have reported several times in the past, that the two companies have talked about it. The last time we reported it was last fall. And Meta said at the time, they sort of played it down. It's, oh, we'll just look at other models to benchmark our own. But the fact that we have reported it, they have discussed it a few times, and The Times is now reporting it, it makes you think that it's obviously something that they are thinking about all the time.

37:10Laura Mandaro:Okay. Okay. So maybe it's not as surprising then, and that's a good point. I didn't realize that I'd been talking about this for so long, but I mean, it does kind of get back to this debate here. We had a guest on a couple of segments before you talking about consumer AI and his thesis was that it won't really matter which model is better. It's about the taste of who's building the product. and I'm sort of thinking about meta and I'm thinking about, okay, well, even if they don't have the best model and they just use Gemini, I mean, maybe there is hope though, because they have showed time and time again that they can build the applications that hook people.

37:52Peter Deng:I guess. I mean, they've shown they can build applications, yes, that hook people, that are really great for you, that really are addictive, that force you to spend, to rot your brain by watching these short videos for hours and hours, so much so that they're facing now, you know, all this litigation. So I'm not sure about Meta. I mean, I think Google has demonstrated that they can develop products that actually people use. OpenAI, that's still a work in progress. Anthropic maybe as well. Meta, who knows?

38:30Laura Mandaro:Okay. What did you make of the other big headlines this week we had?

38:34Peter Deng:I should say that Meta has definitely proven that they can use AI to improve their advertising, which is a really important thing. Their advertising revenue continues to grow at like 25%. It's a giant business and I think a large part of that is that they are using AI to really refine the targeting and the delivery. And that's not something that we should dismiss. So maybe they're not going to be so great at consumer products, but they are going to be very good at the advertising piece of it.

39:11Laura Mandaro:Right. But the advertising is because they can hook people. I mean, that's the whole reason.

39:15Peter Deng:I'm not sure that being great at addicting people to a toxic product is the thing that you want to be really known for. But that's just me.

39:27Laura Mandaro:Right, right. Although, and then, but if you can hook them, then the advertisers, then they get what they want. But the advertisers are also hooked. Right. What are the other big headlines you were watching this week? So we had Oracle earnings, which came up strong. We had NVIDIA. I think you made the point it was their fourth big deal in a month, I think, that they've invested in one of their - Let's cut to the chase here.

39:53Peter Deng:There's a lot of smallish headlines this week, but the only other big news really is the war because I think that's what people are not really paying attention to is how, as this war continues, the uncertainty that it will create for the market, and that is just going to make it harder for a company, even like SpaceX, to go public, which is really important because Musk needs to go public to raise the money that he wants, that he needs to fund XAI. He's already going to have issues now that he's having problems with XAI. But if this war goes on and really it looks like it could just drag on, then all bets are off as it relates to the IPO market.

40:39Laura Mandaro:And other than the IPO market, in terms of AI as a global story, I mean, you have the different regions that people are building data centers. You know, you certainly have the energy part of this equation. I mean, what are the other risks here to the tech story that you're watching?

41:00Peter Deng:Other risks to the – I mean, obviously, the other big one is the question of whether or not the financing markets will be there for all the companies that are raising money. I mean, this week we saw Amazon raise, I think, 40 odd billion, maybe 50. Yeah. Salesforce. 25. Meta has been raising money. I mean, Alphabet as well. That's the big question is, I mean, those companies have got, you know, easy access to the market. But then there are all the other companies that are raising money to build data centers and the like. And so that, I guess, and that does relate, though, to the broader macroeconomic conditions and the war.

41:50Peter Deng:So everything is connected.

41:52Laura Mandaro:Right. And so last question for you. I mean, as a reporter then, you know, as a – if you were a reporter – and it's been around for decades i know better than no i'm saying every editor was once a reporter

42:10Peter Deng:so you know that's why it's not true actually but i was a reporter but okay well you were

42:15Laura Mandaro:you were okay you know this is this is this is like this is like the uh the david letterman and obama thing obama asks letterman a question and letterman says okay let let me let me tell you how this works i ask the questions and you answer them okay so uh as as a uh uh grizzled veteran running a newsroom what are the reporting questions for you that you want to know in the weeks ahead watching the story we've got gtc next week with nvidia they're going to come out with these chips we're watching the data center build out jensen wang he appears in public he gives

42:51Peter Deng:these speeches every month or so there won't be i mean that we've all heard that the really big big stories are going to be the progress that OpenAI, Anthropic, Meta, XAI, the progress that they all make, and whether or not they can all keep up. And I think right now it looks like Google is ahead, Anthropic is close to them, OpenAI is also in there, the other two I'm not so sure. And I think that's really the biggest story because if those, for instance, if XAI has to pull back, that has an impact on SpaceX. Now he's combined them. It just has enormous implications for Musk and people's sort of views of him.

43:38Peter Deng:Meta, they've spent a fortune, an absolute fortune on this. And if this doesn't work, that's also a problem for them. So to me, that's the biggest story that there is.

43:50Laura Mandaro:Great. Well, Martin, I want to thank you for coming on. That is Martin Peers, our co-executive editor. here at The Information. 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. I want to thank you all for tuning in. We really do appreciate your viewership. Make sure to subscribe to The Information on YouTube and follow us on X, Instagram, TikTok, and check us out wherever you get your podcasts. I am already excited for our next show on Monday. Have a great weekend. Bye-bye for now.