OpenAI Snaps Up OpenClaw, Yahoo’s AI Search Engine ‘Scout,’ Inside Anduril’s $60B+ Valuation Talks

0:12Welcome everyone to the information's TI TV. My name is Akash Pasritch. It is Tuesday, February 17th. It is an exciting start to the week. OpenAI has snapped up OpenClaw's founder Peter Steinberger. We'll talk to one of our AI reporters who scooped news of those talks happening late last week before the announcement went live. Shares of Pinterest are trying to rebound this week after earnings triggered a 20 % single-day plunge late last week. My colleague Anne Guillen has in-depth reporting on what's happening behind the scenes, including a series of code reds at the digital scrapbooking platform.

0:51I'll also be joined by Yahoo CEO Jim Lanzone to discuss the company's own turnaround strategy and how AI factors into its playbook. And finally, the information has exclusive reporting that Anduril is in talks to raise billions of dollars at a$60 billion valuation, roughly double its last private funding round in June. We will speak with one of the reporters behind the scoop about what this funding would mean for the company's expansion plans. It's going to be a fun show, so let's get right on into it. In the latest development in the Open Claw saga, Peter Steinberger, the founder of the project, is joining OpenAI.

1:30The announcement over the weekend followed exclusive reporting from the information that OpenAI was in advanced talks to hire Steinberger. To break it all down, I want to bring on Rocket Drew, one of the reporters behind that story. Rocket, welcome back to the show. It's great to have you here. Thanks, Sakash. Great to be here. So we've started each of the last three weeks with big news about OpenClaw and what a full circle in some ways this story has come. Although I don't know if you can call it full circle. It's a development that I don't know if I expected. Did you see this coming? You know, I did see this coming, at least as of last week.

2:07We knew that Steinberger was in talks with some of the AI companies. He also had VCs banging down his door. It was clear that OpenClaw was destined to become something greater than the sort of open source side project that it's been so far, but we didn't know where he was going to end up. So what have both groups said about what Steinberger is going to be doing at OpenAI? Well, OpenAI has made clear that they want him to work on personal AI agents, which is what OpenClaw is. I mean, that's clearly been a key area of success for him so far, and OpenAI is looking to ride that wave a little bit. For Steinberger's part, he's made clear that he's dedicated to open source.

2:50He really cares that OpenClaw remains open source in the sense that anyone can download it, they can tinker with it, developers can modify it. That's core to his vision of what OpenClaw is. So to accommodate that, OpenAI is going to set up a foundation that will house the OpenClaw project. We're still waiting to learn more about what the details of that will involve, but it's sort of on brand for OpenAI to create a kind of bespoke nonprofit structure in order to accomplish a very particular mission. Right. Bespoke and nonprofit and then maybe somewhat regretting the way that they structured it.

3:28Yeah, yeah. And then I'm sure Elon Musk will have something to say about it. Right, right. But so it sounds like this is very much, this is sort of an acquihire in terms of the talent sense. I mean, they're not getting any part of OpenClaw itself that's remaining an entity on its own. But they're saying to Peter, hey, we want you to do what you did, but do it for us. Help us build out our agents. What exactly is OpenAI's agent product portfolio right now as it is? Yeah, they do have many agents, but they're different from OpenClaw in a few notable ways. So maybe their best known agent is Codex, which is a coding agent, which means it's specialized for writing code.

4:08It's meant to replace the job of a software engineer. In contrast, OpenClaw, as we've talked about, is sort of a general purpose framework for creating whatever kinds of agents you want that can do anything on your computer. So sure, they can write code, but also they can check your email and your calendar. They can send Slack messages. They can do things online. You can text with them from your favorite messaging app, from Telegram, whatever. It's much more general purpose than just coding. OpenAI also has Deep Research, which prepares research reports, and a shopping agent, which is sort of a specialized version of Deep Research.

4:42And they also have ChatGBT agent mode. But agent mode is really focused on completing sort of shorter self-contained tasks on a browser, like book me a flight or fill out this form for me. In contrast, OpenClaw is very general purpose. And a distinction that we've made here before on the show is that OpenClaw is not itself an agent. It's a framework for creating agents. In contrast, ChatGPT agent is pretty narrowly scoped. So I would say on the one hand, that's something that Steinberger offers is this direction of more powerful general purpose personal agents. But also to your point, I mean, clearly Steinberger is doing something right.

5:22He had this overnight viral success with OpenClaw. It took off like crazy with developers. And presumably OpenAI is interested in tapping into some of Steinberger's talent. And he clearly has a finger on the pulse of what developers are interested in, in terms of agents. I'm also trying to put this news in context with what we've seen at Anthropic, we saw Claude Cowork really taking off and dominating headlines a couple weeks ago. This was around the same time that OpenClaw was coming into the conversation. Do you see this as a clear, you know, OpenAI's answer to Claude Cowork in some ways? Yeah, yeah, I do.

6:03I think they will be sort of natural competitors. I think Claude Cowork will have a greater in with enterprises because it offers more security. It is kind of more narrowly scoped. Like you usually tell Claude Cowork what particular folder you want it to work in and what particular sort of close-ended task you want it to carry out. And then when the task is done, it's done. It's unlikely to be sort of running in the background indefinitely the way you could set up an open claw agent. So I think Cowork will have sort of a natural advantage there, but they're accomplishing different things. They're reaching for different audiences.

6:37I think it's interesting that Anthropic is not where Steinberger ended up. And you could think there are different reasons for that. I mean, one is maybe just naturally Anthropic is more inclined in this direction of enterprises, so they're less interested. Two, Anthropic has kind of an interesting relationship to open source historically, and you could think that maybe they wouldn't be interested in developing open source technology like this. And three is that Anthropic and Steinberger got off on kind of an interesting footing because anthropic asked him to change the name of claude bot you'll recall open claw was originally called claude bot and anthropic saw that as sort of stepping on their intellectual property with the the claude name so right maybe that foreclosed some some avenues there but it is kind of interesting you know trying to think about what open ai hopes to create with peter steinberger now on their team i mean i'm trying to think about if open ai would ever try to replicate something as wild as OpenClaw?

7:38Because you and I have talked on the show previously about, look, the tool is great and you can do a lot with it. It also means that there are security concerns that could come with it. And there are modes that really unlock a lot of possibility and also a lot of risk. And so I guess the question I have for you, Rocket, is do you see open ai really trying to replicate open claw in you know in everything that it was within its own company or is this something where it will be a lot more guarded and maybe resemble a little bit something more towards the clod co-work enterprise solution you know i think we'll probably start with something that does have more guard rails it has more protections against your agent getting hacked or prompt injected when it interacts with random users and random pages on the internet.

8:28But I think gradually we'll take steps in the direction of the sort of freedom and openness that OpenClaw brings. It is going to take a while to get there, but Steinberger has said his goal is to get OpenClaw to the point where even his mom could download it and install. What we've talked about so far is that OpenClaw is difficult to set up. It takes a technical user to do that. And it especially takes some technical expertise to set the permissions just right and implement the sort of security precautions that will ensure that OpenClaw doesn't accidentally share your bank account information.

9:03But if Steinberger is saying his goal is to get it to the point where even his mom could download it and use something like this, I think that's the direction we're going to be seeing from OpenAI. It just may take some baby steps to get there. What has been the AI community's reaction to this? Have people been celebrating this or have they kind of been saying that this is just the big getting bigger? Yeah, yeah. You know, reactions have been mixed. I think some people are pessimistic about the future of OpenClaw now that one of the sort of large players in the space has grabbed onto it. People celebrated it so much because it was this open source side project.

9:38I think people should be reassured that Steinberger and OpenAI have doubled down on their commitments to keeping it open source. but there is a sense that it's going to spoil some of the fun and the weird, wacky freedom that it has unlocked. On the other hand, I think people should realize that, you know, Steinberger has been very vocal, that he didn't have the personal time to deal with all of the requests that people were sending in for OpenClaw, for debugging all of the issues that were coming up and adding the security precautions that people were asking for. I mean, he was just overburdened.

10:09He needed a bigger team in order to grow OpenClaw into a really usable product. and he was also burning cash, spending money on developing it. I think he had said$10 ,000 to$20 ,000 a month. And OpenAI stepped up and started footing some of that bill. But I think people should also recognize that this is maybe the path that will bring OpenClaw to the masses or develop it into a fuller-fledged product. And if you read his blog post, I mean, he basically says, he said, look, I didn't really want to build another company. He says, I've done that drill before. And so one question I had as I was seeing this news play out was, well, how big could OpenClaw have become in terms of a business, right?

10:52It's an open source project. Could this have been another multibillion dollar business? And basically what he lays out in the blog post is, I don't really care what it was going to become. I didn't want to do the work. I wanted to build the tools, which is what OpenAI is going to let me do. It does lead me to the last question, though, which is, I wonder how much OpenAI is paying him because that number is not out there. But I think we do know that he seems to have beginning offers from other labs as well. And so I do think it's a pretty hefty paycheck. I bet it is. Yeah. You know, I would just say don't trust the numbers you see on Twitter.

11:28There's a lot of numbers floating around on the Internet that seem like mostly speculation. but it's true. And, you know, he said he also had an offer from Meta, which has been throwing around some pretty big numbers in terms of compensation for the people that it's acquiring. So, you know, it's possible Steinberger was able to play them off against each other and get them to raise their offers. It's difficult to say, but surely he's being compensated pretty well. Yeah. If he had gone and made a startup himself, since it's open source, it's hard to say what would have happened in terms of revenue.

11:57But again, he had VCs banging down his door. I'm sure he would have had no trouble raising a lot of money, especially in the current ecosystem for raising funds for buzzy AI startups. Well, I'm excited to see how this all plays out, Rocket. I want to thank you for coming on. That is Rocket Drew, our AI and robotics reporter here at The Information. Thanks. Pinterest shares are looking to rebound today after the company reported disappointing quarterly results last week that sent shares falling 20 % in a single day. At the center of that story is revenue growth that is decelerating. Our e-commerce reporter Anne Guillen today published some in-depth reporting about what's been going on behind the scenes.

12:37The company declared its own code red well before this stock drop, it turns out. I want to bring on Anne to tell us more about what she has learned. Anne, welcome back to the show. It's great to have you here. Hi, Akash. What's going on at Pinterest? So you mentioned the code red, and in my story today, I get into what that has looked like. Internally, that's mainly aimed at addressing some of the slowdown in their ad revenue growth that they've been dealing with over the past couple of months. Also, my story, I get into Pinterest has really been trying to revamp a lot of their sales and ad teams.

13:17They just brought in Lee Brown, who was at DoorDash. He's also been an ad executive at Spotify. So Pinterest is really, I think, feeling the pressure from investors now, you know, with their earnings report last week, but also they've certainly been feeling the pressure internally for the past couple of months and really trying to both kind of turn around the ad business in the short term, as well as deal with this longer term, more kind of existential question of what AI is going to do to their business. So you talked a lot about all of the problems facing the company. And just to run people through the gamut here, one of the problems is that users don't check the app as often compared to other social apps.

14:06You talked about the recommendation algorithms, the fact that they haven't been able to target and measure how well the ads work. And then there's obviously the competition aspect for Meta and Google. I wonder if you think about all these problems and from your reporting, what is the single biggest problem facing the company that Bill Reddy is most concerned about right now? Is it AI or is it really just the fundamental platform? I mean, the immediate challenge and the immediate changes they're going to have to show to investors is turning around the ad business. And so it's interesting that you kind of describe them as a social media app because I think Pinterest is trying to kind of sell the narrative, you know, to investors in particular that they shouldn't really be compared as directly to other social media apps.

14:57They're really trying to position themselves, especially in the age of AI, as a search app. And so they're really pointing to all of the, you know, it's mainly visual platforms. So they're trying to lean on kind of their expertise there and all of the search data they have about what people are searching for on their platform. and really use that to, you know, say to advertisers that it's a good place to advertise because people are often looking on Pinterest for, you know, things like clothing or beauty or furniture. And so those are all really strong categories for advertisers. So I think kind of the main challenge will be for Pinterest to kind of show to both to advertisers and to investors that they really should be thought of as a search platform and not in the same way as like an Instagram or Facebook and really kind of make that transition into more of a search app versus more traditional social media.

16:02Well, and I think one thing that I just realized actually as I've been thinking about this conversation is that when I talk about Pinterest, Sometimes I refer to it as a digital scrapbooking platform. Sometimes I refer to it as a social media platform. And so I think this sort of identity crisis that you're describing, where advertisers don't really know what the one thing is that we should be going to Pinterest for, I mean, that's sort of a problem everyone has right now is we're not really sure what the thing is and what it specializes in. I want to go back to the code reds that you talked about.

16:37What do we know about where those stand right now and if those have yielded any results or changes within the company? Yeah. So I spoke to Pinterest's chief technology officer, and he told me that the Code Red is still going on. There are a couple of kind of key metrics that the company is working towards in order to get out of the Code Red and declare that it's over. But a lot of the Code Red projects have been aimed at boosting some of those key metrics, like I mentioned, in particular in the ad business. And so what Matt, their CTO, told me yesterday was that already some of the changes to their ad recommendations and modeling systems that they've put into place, those have already translated to a better return on spending for their advertisers.

17:26And Pinterest has been able to kind of reallocate some of their GPUs so that they can, you know, kind of test these changes very quickly with all of their advertisers. So they've been able to kind of come up with some of those changes and then also get them live for all their advertisers. So they clearly, you know, are feeling the need to work quickly and then also get those changes live for advertisers. So I would say, you know, the Code Red's still going on, but I think Pinterest is feeling good for now, at least about the work they've been doing so far. And where do you think this story ultimately plays out from here?

18:06You had a prediction late last year, maybe early this year. Your 2026 prediction was that OpenAI should buy Pinterest. And we obviously know that that went somewhat viral and people seem to latch on to this idea that maybe it was happening. We should say it was just a prediction. It was a prediction. Yes. But people really latched on to it, probably because maybe they saw some sense in there. Do a lot of people think that Pinterest is going to be acquired? What are you hearing? I think that's kind of the open question. I think there's no doubt there is a ton of underlying value in Pinterest's platform, both in their user base.

18:49You know, it's typically, it's dominated by Gen Z and millennial women who are basically the most valuable categories for advertisers. So they have, you know, a huge user base that's very valuable. And also just all of the data that they have. They have tons, you know, they have billions of images and posts on their site at this point. The company said last week they have 80 billion searches per month. So that's, you know, that's definitely legit search volume. I think just the question for them, at least in the near term, will be can they kind of fix the monetization and fix the ad business? And so I think if they are able to fix the ad business, at least in the short term, that will keep Wall Street happy.

19:37But I think they're also contending with, you know, the questions that a lot of tech and software companies are thinking about right now is how does AI kind of impact their long term future? And I mean, you see companies like Google trying to make their own Pinterest copycats. And so I think it just remains to be seen if anything will kind of take hold with users the same way that Pinterest has. And then, yeah, if the business, if the ad business isn't able to recover, then I think the expectation is they would think more seriously about a potential sale. Great. Well, Anne, I want to thank you for coming on.

20:14That is Anne Guillen, our e-commerce reporter here at The Information. Our next guest is leading one of the most historic brands in Silicon Valley. Jim Lanzone is CEO of Yahoo, the company he joined in 2021 after it was acquired by Apollo. Lanzone has been leading the business through a turnaround plan that has included introducing its own new AI tools. He himself has had quite the career previously leading Tinder, Ask Jeeves, CBS Interactive, and founding a few companies himself. I want to bring on Jim to chat with us about where he plans to take Yahoo from here. Jim, welcome to TI-TV. It's great to have you here.

20:52Gosh, thanks for having me. Great to be here. So Yahoo is a name and it's a brand that certainly people know very well. I think given the long history of the company, people might be less familiar with what the current business is. And it's got so many different properties in it between fantasy and mail and news. So just give us the primer here. And when you think of Yahoo the business today, what are the different segments of the business and how do you structure it? It's funny. You know, having been a part of this turnaround now for a few years, I actually think people are very familiar with it.

21:23It's just that they're familiar with the brand that they use. So Yahoo Finance, Yahoo Sports, Yahoo News, Yahoo Mail, these are all number one or number two in their categories. and every any given month you know we are uh 250 million users a month in the u.s 700 million globally so for our users and there are a lot of them you know they're very familiar with those brands obviously in silicon valley the larger yahoo story you know which is now 30 years old uh has a lot of baggage with it that needs explaining um i think that kind of adheres to the maybe the global brand of it um but uh but no yahoo is uh you know especially now a few years into our turnaround.

22:00It's a very healthy, thriving business. And we've now kind of evolved to the point where we're starting to take shots on goal on the product side. So walk us through the business today, the structure of it. I mean, how do you think about it? Do you think about, you know, content as one bucket? Do you think about the mail and, you know, the advertising business as another bucket? What are the different sort of pillars? Look, I'd say the CEO of Yahoo used to be put on stage kind of for sport and asked to define whether it's a media company or a tech company, the answer is it's a product company, right?

22:34When you're using fantasy football, that's a product. Stock charts, that's a product. Yahoo Mail, Yahoo Search, where we're still the third largest search engine in the US. So like any major consumer internet company, you know, we come at it from a consumer first basis. You know, these things collectively then add up into one of the best treasure troves of data on the planet. We use that not only to provide better user experiences, but also to provide insanely well-targeted advertising that converts at an A-plus level compared to most major internet companies out there. And that's still the spine of the book for us from a monetization point of view.

23:12So it is all of those things collectively you know, together. But each one of them operates independently, internally. And from that standpoint, you know, we have a general manager on charge of each business. So someone who's an expert at mail, an expert at sports, at finance, et cetera. And we let them run and compete in their categories. And, you know, by that token, I would say we're very similar to other conglomerate internet companies, right? Neil is the CEO of YouTube. You know, you have Adam running Instagram, And we have our own, you know, smaller, in some cases, version of that. But they're, you know, so Yahoo really collectively comes together with the top level brand.

23:49But we are, you know, as a company, we are strongest as a bunch of independent products, you know, competing in their categories. Walk me through the turnaround plan that you have executed on over the past five years. What were the initial problems that you identified? and then one by one walk us through the past five years and what the major milestones you've hit are. And then the second half of this is I want to talk to you about the AI era and what you guys are launching there. But just walk us through the last five years and what your playbook was. Really, it's a little over four years. We got a hold of this asset at the end of 2021.

24:28Very interesting time to take over a company. It was the middle of COVID. People were not able to see each other. We kind of had to get through some of that. the first couple years of a turnaround are foundational where you're identifying you know the opportunities uh both to improve things but also like what you might be able to do with these assets on the upside um so we did that transformation work and that involved everything from reorganizing the entire company um bringing outstanding talent you know from from you know uh my you know past days uh as well as you know their networks and and we really have an incredibly talented people running these properties now.

25:03And then we also had to do things like completely redo the business model. The underlying revenue engine of this company was very different and really had to be replaced. So in early 2023, we swapped out a lot of the past 10 to 15 years of ad tech that was supporting the revenue and replaced it with a whole new revenue engine. So you really had to go through all of those steps, which then left you in 2024 with the ability to start to rebuild all of you know our consumer facing products and every pixel on yahoo was was rebuilt um between you know 2024 and early 25 uh leading to uh you know some of the things we'll talk about in a minute on the ai side that we've done um but really you know those are all the steps so i would say for a turnaround is that yahoo had great bones and even after all this time it had a huge amount of traffic which is the hardest thing to get it had a very well-known brand which even though it had been been through the ringer is also very hard to build and on top of those two things you can you know really do things you can optimize the team you can you know figure out the right ways to make money and you can start to build better products uh and you know it's funny with some of the things that we've done people have said i wouldn't have expected that from yahoo and and you know my first instinct is well this is definitely not the same yahoo you know these this is entirely a new team of people who are not here for the periods of time that they're they're talking about um we're just people who came in saw an opportunity with these assets to do something special and yeah if we're successful at that uh then you know i think the turnaround is a great story but i promise you that every day that we go to work we're focused on how do we create you know great user experiences and deliver for our advertising partners So now you've firmly entered the AI era and you've come to market with Scout, the Yahoo-powered AI chatbot.

27:02And you've said that it leans on a lot of the data that you have in the background through all your Yahoo properties. What to you is the differentiating factor here with Scout as you compare it to all the other chatbots that are built from search engines? Well, first I'd say there aren't very many. So I'd say just by entering the category. The big ones, the big ones. Yeah, yeah. The big ones. I agree. How do you compete? It's interesting that there, as you really look around the category, you know, you have the major LLMs, which would include Google. And then it's really perplexity in us is the only like really scaled search engines of this nature.

27:40As we looked at that category, you know, coming out of our transformation period and into the ability to take shots on goal and still being the third largest search engine in the U.S., you know, We had to make a choice. Yahoo has been outsourcing search to Bing since 2009. So it's a very long time since Yahoo really tackled search. But as we looked at the AI category, it really turned out to be an opportunity to get back into it ourselves and that the best people to build the next generation search engine for Yahoo was Yahoo. And the reason is because we are the ones who have the first party relationship with all these users.