In short
Podcast Summary: The Information's TITV - Episode on China-US TikTok Negotiations, AI Bubble, Enterprise AI Search Engine
Podcast Title: The Information's TITV Episode Title: China-US TikTok Negotiations, AI Bubble, Enterprise AI Search Engine Date: September 15, 2025 Host: Akash Pasricha Guests: Jing Yang, Cathy Perloff, Ann Gehan, Deedy Das, Richard Socher
---
Episode Overview In this episode of The Information’s TITV, host Akash Pasricha discusses critical topics including the ongoing US-China trade negotiations involving TikTok and NVIDIA, Perplexity's advertising strategies, the valuation of AI startups, and You.com's shift towards enterprise AI search solutions.
---
Key Discussions
- China-US Trade Negotiations
- Context: The third round of trade talks between the US and China commenced in Spain, focusing on TikTok and NVIDIA.
- Key Insights from Jing Yang:
- President Trump hinted at positive developments regarding TikTok on Truth Social.
- The inclusion of TikTok in negotiations is unprecedented and reflects China's willingness to negotiate.
- The Chinese government views TikTok as a pawn in the broader trade negotiations.
- Antitrust investigations against NVIDIA by China are also prominent in ongoing discussions.
- Perplexity's Advertising and E-Commerce Efforts
- Discussion with Cathy Perloff and Ann Gehan:
- Perplexity's advertising initiatives are slow to develop, with limited advertiser participation and a cautious approach.
- E-commerce features, such as "Buy With Pro," are limited, focusing on a narrower range of products.
- The company seems to be experimenting rather than fully committing to advertising and e-commerce strategies despite evident demand.
- AI Valuations in Venture Capital
- Guest: Deedy Das from Menlo Ventures:
- Discussed the volatility of AI valuations and the potential bubble in the industry.
- Emphasized that high valuations may not necessarily reflect long-term viability.
- The venture capital environment is competitive, leading to inflated valuations for promising startups.
- You.com's Pivot to Enterprise Search
- Guest: Richard Socher, CEO of You.com:
- You.com has shifted focus from competing with Google in consumer search to providing enterprise search solutions.
- The emphasis is now on delivering accurate search capabilities for industries where accuracy is critical, such as publishing and legal.
- Socher discussed challenges related to accessing data within larger companies but noted strong motivation from clients to facilitate integration.
---
Key Takeaways
- Negotiation Dynamics: The inclusion of TikTok and NVIDIA in US-China discussions highlights the evolving geopolitical landscape and the complexities of tech diplomacy.
- Advertising Challenges: Perplexity's cautious approach to advertising reflects the broader struggles faced by new platforms in competing with established players like Google.
- AI Market Sentiments: Investors remain optimistic about certain AI companies but are wary of potential overvaluation in the rapidly changing tech environment.
- Enterprise Focus: You.com's strategic pivot towards enterprise search underscores the growing demand for specialized search solutions that prioritize accuracy and integration with existing organizational data.
---
Conclusion This episode of TITV provided a detailed overview of crucial developments in tech and trade, highlighting the interconnections between global negotiations, venture capital dynamics, and the future of search technology. The discussions showcase the multifaceted nature of the current tech landscape and emphasize the need for adaptability in strategies across various sectors.
For more insights, viewers are encouraged to tune in daily at 10 am PT / 1 pm ET.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Transcript
Automatic transcript. May contain errors.0:13Welcome everyone to the Informations TI TV. My name is Akash Pasricha. It is Monday September 15th. Hope you had a fantastic weekend. We have got a great show lined up for you today. We are talking to our advertising and e-commerce reporters about perplexity and the status of its efforts in both of those spaces. We've also got Menlo Ventures' newest partner, Didi Das, coming on the show. We're going to get him to explain some of his hottest takes on AI and on venture capital. We've then got the CEO of You.com coming on the show. They just raised a ton of money earlier this month. But I want to start with the big story that is developing right now as we speak in Spain.
0:52The U.S. and China began a series of meetings in Spain over the weekend to negotiate key elements of their trade relationship. And it's not just tariffs that are on the table. TikTok and NVIDIA have also become central to these negotiations. And I want to bring on our Asia Bureau Chief, Jing Yang, to help us make sense of all this. Jing, hello. It's nice to see you. Hi, Akash. So we should say that we are taping this. It's 9.30 in the morning Eastern time because it's 9.30 p.m. where you are. And so when people see this, a couple hours may have passed. We have to say that because it is really a fast-moving story.
1:31We want to let you go to sleep at some point. But, Jing, I want you to sort of walk us through the backstory. And specifically, let's start with the president's post on Truth Social, which is sort of the latest news that we have to this story. Yeah. So, as we know, the third round of trade talks between the U.S. and China began on Sunday, yesterday in Spain. And then about 24 hours later, we have President Trump posted on his favorite platform, Truth to Social, in which he dropped a very, very strong hint about a deal either has been or is going to be reached over TikTok. And he said, you know, a deal was also reached on a certain, quote, company that young people in our country very much wanted to save.
2:24They will be very happy. And he also said that he will be speaking to Chinese President Xi on Friday and that the relationship between the two countries remains a very strong one. Okay, so we have that post and there will be many posts like this, I'm sure, over the coming hours and days. But take us back in time a little bit. What is the backstory to these meetings and why are these meetings happening at all in Spain for that matter? Hmm. Well, Spain is just a location that where, you know, government officials from both countries are traveling to at the same time. We've seen that, you know, we've seen that this is third round, as I said earlier, and we've seen before the previous round of trade talks also happened in sort of a more neutral location, you know, Geneva and London previously.
3:21So it has been actually sort of a 72-hour roller coaster, I would say, for those who are watching this, not just on TikTok, but also on whether US and China can reach a so-called, a so-called, you know, grand bargain on trade and tech. So we got the very first strong hint from the Chinese side on Friday, where the Chinese Ministry of Commerce actually issued a very rare statement specifically mentions that TikTok will be included in the agenda, the Spain talks, which never happened, which didn't happen in the previous runs of talks. If some of you recall, I remember a few months ago, U.S. Treasury Secretary Scott Benson, actually, in talking to CNBC, when asked about whether TikTok was part of the trade talks, He said yes and no.
4:18So this was a surprise that TikTok was going to be included in this. Yeah, it is a huge surprise that Beijing actually issued a statement seemingly out of nowhere just for the purpose of telling the world that we're going to talk about TikTok. And the tone, the undertone of that statement also was quite positive to me. A lot of the sources I talked to over the weekend also sort of echoed my sentiment and say that if Beijing actually remarked on this, that means that there is a strong willingness from China's side to actually reach a deal. And obviously, however, what happened after that was also on Friday, U.S.
5:02Commerce Department decided to add dozens of Chinese ship firms to a trade blacklist. While none of these companies is big, because to be honest, all the big companies, all the big chip companies in China have already been added like years ago. But that was sort of seen as, at least by the Chinese side, as Washington trying to up the ante, trying to show a tough stance ahead of the talks. So in retaliation, China actually over the weekend announced, you know, subjecting some U.S.-made logic chips. By the way, these are like, you know, chips made of very mature nodes, sophisticated technology that China can produce as well.
5:50But subject these chips to anti-dumping in the Chinese market. Then on Monday in Asia time, so when you guys are sleeping, the antitrust regulator in China also announced that they are going to continue to investigate a case they opened against NVIDIA back in December, which related to NVIDIA's acquisition of an Israeli startup in 2020. We reported back in December that that case, opening of that antitrust program video was China's sort of preparing to line up its cards in case that Trump comes into office and launch an all loud trade war. And then it exactly that is exactly what has happened.
6:38And this is all over the past 72 hours alone. Yeah. OK. OK. So now we understand why you would like to get to sleep. but it's been a busy couple of days for you. Okay, so we've had all of this movement now in the last 72 hours. The question I want to understand from you is as it relates to TikTok, we can get into the chip firms in just a second, but with respect to TikTok, who has the leverage here? How much does the Chinese government care about TikTok now? You know, we have these delays. You know, President Trump has been sort of delaying the ban for months and quarters, and I don't know how long it's been now.
7:17So who do you think has the leverage here and does China really care about TikTok now? I would just say this, right? If you would just look at how all the internet platforms in China were subject to this brutal crackdown in China back in 2020 to 2022. Remember, you know, Alibaba and all these companies got big fines and everything. That actually showed the preference of China's all-important leader, Xi Jinping. He does not like internet platforms. He doesn't seem to think these represent real innovation or represent innovation that can actually help China move forward. That sort of is an understanding.
8:10He obviously never said anything about TikTok or even Biden. Not that I know of or been publicly documented. But in the grand scheme of things, China has been seeing TikTok as a pawn in negotiation with the U.S. I think, you know, the Chinese diplomats and officials, they are very smart. they can see that, you know, TikTok helped Trump win the election, right? And then Trump has given TikTok a lot of credit for that. And then China also has installed back in 2020, China installed this restriction on, you know, algorithm from being exported. That means that, you know, there's no deal. You know, the TikTok cannot be saved.
9:00The TikTok that we know of cannot be saved without China's approval. Right. On the other hand, it seems like all the signposts suggest that Beijing would be willing to treat TikTok and TikTok's algorithm in exchange for something that Beijing truly wants, be it, you know, semiconductor, you know, some kind of relaxation on, you know, semiconductors or something else. Right, right. Last question for you. What should we be watching with respect to the NVIDIA angle with this story as it develops? Yeah, so right now, as we reported exclusively several weeks ago, Chinese government has come out very strongly against Chinese companies from using NVIDIA chips, H20 or any other, you know, alleging there is national security risks to China.
9:56So I think, and then given the recent development of NVIDIA's antitrust probe over NVIDIA is being carried forward, I think for NVIDIA, the angle to watch is whether NVIDIA will then become sort of another political football that sort of being subject to the unfortunate... Which we've seen. Yeah, it's becoming... Right, right. Yeah, well, with TikTok. Right. Okay. Well, Jing, like we said, I mean, it's a fast-moving story, and I suspect we might ask you to come on again this time tomorrow. So if you can get some sleep tonight and stay up late for us tomorrow, maybe we'll have you back on. Jing, thank you for coming on.
10:42That is Jing Yang, our Asia Bureau Chief at The Information. Okay. Well, Perplexity has been one of the most closely watched platforms as it relates to advertising and e-commerce products. And over the weekend, my colleagues, Catherine Perloff and Anne Guillen, published a story about the status of those efforts. And I want to bring them both on to tell us more about what they found. Catherine and Anne, it's great to have you. Welcome back to the show. Hi. Hi, Kosh. Okay, Catherine, let's start with you. What did we find with respect to Perplexity's advertising products? Well, you know, they launched, they said they're going to have an advertising business in November of last year, same time they said the, you know, experimenting with e-commerce.
11:26And, you know, since then, they really, they haven't accepted that many advertisers into the program. So a lot of advertisers are really eager to join. You know, we've talked about how a lot of advertisers are frustrated with Google Ads or isn't a lot of other places to do search advertising. But they haven't really let that many in. To the extent that they have an ad product, it's pretty limited. It's sort of a sponsored question, a suggested question that will come up under your question. And there isn't that much, you know, in terms of like, you know, the customizability or making your brand prominent.
12:05There aren't that many options. And, you know, perplexity seems to not be that, you know, they've called it an experiment. They're not letting that many advertisers in. And the guy leading the ad product, this guy Taz Patel, left last month. They have a... Why do we know why he left at all? We don't exactly know why he left. That is a question. Oh, okay. And they also, they had this program where they pay publishers. They work with publishers who they cite to some extent. They used to pay the publishers via ad revenue. Now they've changed that to pay them via some subscription revenue. So, you know, it's sort of like they're taking a very kind of putting a toe in the water, but it's been almost a year and they're not really embracing it fully, despite the fact that advertisers, I think, want to try it out.
12:57Okay. So they're tiptoeing around advertising. And what about e-commerce? Have they gone deeper into that? The e-commerce efforts have been a little more robust than the advertising experiment so far. So about the same time they launched advertising last year, Perplexity also rolled out this shopping feature called Buy With Pro. So it's limited to they're paying pro subscribers. but if you are a pro subscriber and you are searching for something, like you're searching for a new pair of running shoes for marathon training, Perplexity will give you special e-commerce focused search results. And with this feature, you can buy certain products with one click right in the Perplexity chat window without having to go to the retailer's website and, you know, fill in your name, your shipping address, your payment information, all that stuff.
14:01So it's supposed to be a very kind of seamless, easy experience. But what we reported in our story is that that experience is pretty limited so far. Perplexity says that they've limited it to certain product categories and certain merchants. so that kind of one-click-to-buy-anything-online experience isn't quite there yet. Right, right. Well, one of the things I want to talk to both of you about is it does seem that perplexity is one of the more forward-looking search interfaces, I guess, that is actually at least they're tiptoeing into these waters. I mean, I haven't seen that much in the way of Anthropic or ChatGPT, you know moving into advertising uh or into e-commerce i mean katherine let's start with you credit where credit is due at least they're trying right or is it just noise i don't know what to make of it really yeah i mean i think it's interesting because i think any media company you know that we've seen throughout history like eventually needs to make money from advertising or often does.
15:16I mean, you know, Netflix said they would never do ads. They did ads. And OpenAI and Perplexity and all these companies have a lot of costs, and they have a lot of users that aren't paying for it. So ads is sort of one, and commerce is one way to sort of make up the difference and still get money for people who aren't paying for your product. Right. So I think that, like, you know, it is interesting that Perplexity has tried, and they're known to be a bit noisy, right? They bid for TikTok, they bid for Chrome, but, you know, they're like, they're honestly, it feels like they're PR geniuses in some ways.
15:54Yeah. But I mean, you have to give them credit. This, they actually, like, you know, OpenAI has announced some shopping, or maybe it's just been reported, but there's been some shopping noise around OpenAI. but the product isn't really rolled out. They haven't done anything around ads. So at least perplexity has tried. But I think I do wonder, I think like how long can they, can these companies not do, you know, ad in commerce efforts when their costs are so high and they are consumer facing products? Right. And what about for vendors and retailers? I mean, how eager are they to get their name onto these platforms and really to start giving them money?
16:30I mean, I imagine you have retailers saying, hey, we don't know what the future of search looks like. We want you to take our money. Did you talk to any of these vendors? Yeah, I mean, it's really interesting. Like Catherine was saying, a lot of retailers and advertisers are really desperate for a new platform that will offer them the same kind of scale, the same amount of users they're able to reach as a Google or a Meta who are really kind of the incumbents that control both the advertising and the commerce industries because they really go hand in hand. So I think a lot of retailers and advertisers are really excited about the potential for a new platform.
17:15It reminds me a little bit of what we saw with the launch of TikTok Shop, where there's this new, exciting app or platform that hundreds of millions of people are using. That's a lot of eyeballs to get your products in front of. And by the way, how is TikTok Shop even going? Just give us the quick refresher. I mean, we haven't talked about it too much on the show. Is it a thriving marketplace right now or has it kind of died down? I mean, TikTok has fully integrated those commerce features into their products. So, I mean, when you're scrolling on TikTok now, you'll see an ad for something on TikTok shop.
17:50You'll see a live stream. So I think that that's something that retailers are hopeful could happen with AI chatbots and AI search engines that the commerce and the ads piece will eventually just kind of naturally fold in the way it has with other consumer apps. Right, right. Catherine, last question for you. Talk to us about what we know in terms of revenue that this is generating for Perplexity. Is this significant right now or is it really just a drop in the bucket? Well, our colleague Sri reported that in the fourth quarter of last year, they only made$20 ,000 from ads, which is, I mean, like, I think we can all agree, no matter what your scale is, that's not a lot of money.
18:38So, you know, that, you know, that's, and then they also said in March at a conference, Taz Patel, who used to lead the app business, but he left, said that they only had around a dozen advertisers. So, you know, I mean, the scale was pretty, it's pretty small for a company that is. But they still, they still, I mean, they're being picky, like we said in the story. I mean, they're being picky. They only seem to want the blue chip advertisers right now. So, you know, it's 20 ,000. I don't know that it's going to get a whole lot bigger based on how picky they're being, but we'll see. So thank you so much to the two of you for coming on.
19:14That is Catherine Perloff, who covers advertising, and Anne Guillen, who covers e-commerce for the information. Okay. Well, you might know our next guest from his outsized following on X, but when he's not posting there, Didi Das is also investing at Menlo Ventures. In fact, he is the firm's newest partner, a promotion that he got this week at the firm. I want to bring on Deedee to talk all about what he's been focused on lately. Deedee, welcome to the show. It's great to have you. Well, thank you for having me, Akash. It's great to be here. So I was looking at your ex, as I normally do, because you've got something like 200 ,000 followers, I think, on there.
19:49And I saw that you were working this weekend. You were coding this weekend. Yeah, I mean, this is not an abnormal weekend. I think every weekend I'm working or I'm coding on something. it's you know what were you doing quickly just give us the 30 seconds what were you coding this weekend this one was completely for fun there's this company there's a startup called listen labs that put out a challenge called the berghain challenge which is uh what we call a stochastic knapsack problem you have to pick people to enter the club and it has to meet that seems like an elite club man a stochastic knapsack problem okay keep going keep going well berghain is this popular club in Berlin that some people might know about.
20:31Okay. It's for not letting most of their guests in. So it's based on Berghain. It makes sense. The founders are European. Right. But you get these people in with certain attributes and you want to let them in and meet certain constraints while rejecting the fewest number of people. It's a public leaderboard, really fun challenge, really interesting way to hire people. And I thought, hey, why not? Why not give this a go and brush up on my coding skills? And it was really fun. And how'd you do? Well, not that well. I only spent about a day on it. And I think the leaderboard has about 1 ,500 or so people and I'm somewhere around 250 right now.
21:07All right. Well, fine. I mean, 250 in one way. For somebody who doesn't do it as often, I mean, I think 250 is for a day's work. Seems pretty good. But anyway, we'll have to come back to that. Look, I want to talk a little bit about what you're investing in at Menlo. And, you know, the question that we've, I think everyone's trying to rack their heads around right now is to what extent we are in a bubble as it relates to AI valuations. We had a great op-ed last week from Pranod Khosla, and he was talking about the extent to which AI valuations are absolutely bonkers right now. And so I just wanted to get your take on it.
21:43I mean, what do you think? Are valuations in a bubble? Look, yeah, I think one interesting thing I learned in my sort of year and a half in venture is a venture doesn't have any downward pricing pressure on startups, right? So it's not like the VCs are sitting here and saying, hey, let's value everything at a bazillion dollars. No, I mean, it's like these startups have potential. They go out to the market and they say, hey, we want to raise a round. And the reality is, because there are so many people with money on the table, we see, in many cases, similar opportunities and then bid each other up to a really high valuation.
22:22And, you know, the way I justify myself around it is, bubble or not, as an investor, our job is to invest in the best companies. And so for every 100 companies, many of which may be overvalued, you know, our belief is that, you know, five or 10 of them are going to have enduring value. And you might come in at a valuation which seems somewhat unreasonable. And many will grow into it. And as far as Menlo goes, you know, we're not doing so many of the billion dollar seed rounds and egregious rounds like that. But even then, I mean, 30 billion dollar valuations for really just one or two people that have a track record.
23:04It happens. I mean, they talk about the single person unicorn and I guess you almost already have it without even having a business sometimes. But, you know, some of these companies are exceptional. I do think they're going to have enduring value. Right, right. Okay. So, you know, I wanted to talk a little bit about an initiative that you are involved in at Menlo, which is the Anthology Fund. This is the fund that the way I understand it, you're sort of jointly investing in startups alongside Anthropic. One of the questions I wanted to ask you was not necessarily related to the fund itself, but when I was covering crypto a couple of years ago at The Information, right?
23:38We had all these different blockchains and developers kind of developed loyalty to one chain or another, right? You had the Ethereum crowd, the Solana crowd, you know? And I ask you this because you're involved with the Etology Fund. Anthropic is a popular tool that people use. You've also got OpenAI. You've got, you know, even companies like you.com coming up on the show later on. You know, my question is, are you seeing developers sort of develop loyalty to, you know, one company's models versus the other in the same way that they had loyalty to blockchains? Or is this kind of a separate game that they're playing?
24:18I think, well, there's one thing I want to clarify about the Anthology Fund, which is a lot of people view, you know, when they hear Fund with Anthropic, they in their head sometimes hear, oh, you mean I have to use Anthropic or Claude to be a part of the fund? And, you know, we've explicitly actually designed the fund for that not to be the case. I mean, a lot of the companies that come into Anthology are not dedicated to using Anthropic or not only on Anthropic, and sometimes they don't even use Claude, and they're just there to further the mission that Anthropic espouses. Though, with that disclaimer, I will say, you know, we actually did a great enterprise report on the team about this, and the results are somewhat intuitive in terms of retention on particular LLMs.
25:01What we found is for developers who are hobbyists or not running large companies, there is a lot more mobility between the LLMs you use. You'll often see, hey, new model comes out. Let's go try something else. Let's go try to get the cost down, see what the performance is here. But when it comes to scaled organizations, they actually have quite a bit of retention on at least the provider level on the companies, the LLMs that they use. And a part of that reason is because when you are using, say, 10 ,000 tokens a minute or some sort of real volume, it isn't trivial for you to go overnight and say, hey, let's move all of my traffic over to this new model.
25:45It might be simpler from a tech perspective. It's just a different API call. But from a feasibility perspective, getting the compute, it isn't that common. And so we see people actually stick to their providers for the most part in the long run. Right. I want to ask you, I want to shift topics a little bit. You used to work at Glean, which is a company that we have covered a lot of the information. And one of the topics that we've covered is this idea of the corporate data wars, right? The idea that you've got data inside these large legacy tech companies now that newer startups have to access.
26:18And in some cases, there are some walls being put up around, you know, access to that data. And this is a dynamic space. It changes by the month, you know, sometimes by the week. But how big of an issue do you think that's going to be for enterprise AI companies and enterprise search companies? Well, you know, I think the information was the first to publish this, but there was a bit of friction, particularly when it came to the Slack API. with Glean because Glean, as with many other startups, need to crawl the data from other SaaS applications to work well. How do I think it'll involve? I actually am very optimistic of the Glean case because if you think about the dynamics involved here, I don't think that the business of at least Glean, and I could say from various other agent companies or companies that involve SaaS integrations, they don't actually take away from the core business of any of these SaaS products.
27:23I mean, if you go back to why do Slack and any of these companies, Dropbox, Slack, email, why do they have APIs? Why do they let people crawl all of their data? The answer is because you get tremendous lock into your ecosystem. Right, right. It just bolsters your own product in a way. Absolutely. And so I think the friction is really when those companies also want to get into the domain of being that data layer across the company. And then they feel like, I think this is a competitive lever that they can use. But why do you think they feel so threatened then? I think it's because of all of these valuations and the companies that are being built in this space.
28:03If I'm sitting here, say I'm Slack or I'm Salesforce, and I say, there's these great companies that are building products on top of my data. At some scale, I wake up and say, hey, shouldn't I be the one building the products on top of my data? What am I giving something away for free? And, you know, I obviously take the stance that the data belongs to the companies and the customers that you sell it to. And these products are additive value. They almost never have I seen an incumbent come up with a product that is actually good. So, you know, it's really, I think that I'm like, okay, that is awesome.
28:41So, you know, that's my stance on the matter. Great. Okay. Well, look, we're at time, so I've got to let you go. But very quickly, I know that you are passionate about all things search, and we've got the CEO of you.com coming on the show right after you. In fact, he's waiting in the wings so he can hear you saying this. What question do you have for our next guest, the CEO of you.com? What do you want me to ask him? I think my I love you.com. Richard is awesome. I think the question I have is how does search, which is fundamentally not a retentive product, build retention over a long period of time and get people to come back in this era of technology?
29:23Great. Okay. Well, you'll have to keep watching because we're going to ask that question to him. Didi, thank you so much for coming on the show. That is Didi Das. He is the newest partner at Menlo Ventures. Okay. Well, like I said, our next guest is the CEO of You.com. You.com initially set out to take over Google search. The company has now been expanding its product portfolio to also sell its tools to businesses. And that focus has landed it a valuation of$1.5 billion earlier this year as it closed $100 million in new funding. I want to bring on Richard, the CEO of You.com. Richard, thanks for being here.
Read the full transcript
30:00It's great to have you. Thanks for having me. How are you doing? I'm good. So look, you were watching, I mean, Didi Das was on the show just now. And the question that he had for you, which I think is a good place to start, is how do you build retention in a search product? Yeah, I think the answer here depends very much on whether you're in consumer or an enterprise search. I think in consumer search, we have seen Google be quite sticky for quite a lot of years. And on the enterprise side, I would argue it's actually even stickier because mostly LMs will talk about what they get from a search engine.
30:37So that search infrastructure layer is not getting as much attention, I think, as it should. And that is one of the reasons we're now focused so much on even the search API infrastructure layer that informs the LMs to give the right answers. And once you're used to searching over your company internal data, as well as your, like the general web data and your LMs merge that together, yeah, it actually is quite sticky. Our net dollar attention is extremely high. So, you know, we wrote this great story about you.com two years ago in our weekend magazine. It was a big read about how you.com was taking on Google search.
31:14And since then, like you said, the company has expanded into enterprise search products. Help me understand the strategy here. I mean, is this a pivot away from competing with Google search? Have you sort of steered the company towards enterprises? Or what exactly is the thinking here on expanding the purview of the company so quickly? Yeah, the main focus is indeed on enterprise search now. What we found is we're the most accurate search and answer engine in the world based on many marks, both internal and external published papers and so on. We don't hallucinate as much. We have the most accurate answers, but really the folks that care most about those accurate answers are actually in the enterprise.
31:58Like if your job depends on it, if you're in a hedge fund, if you're a journalist, if you work for the NIH, you work for publishers like the Telegraph and other companies or Harvey and Legal AI, the accuracy of the answers really, really matters. And so we found this that the search layer is extremely powerful for all of these other companies. And so what we're now doing is actually provide the underlying capabilities we have built to other organizations. And when we're doing that, we can now, we've been able to scale massively the revenue. We do over a billion API calls a month now. I think that's bigger than any of the other startups in the space, both on the API side and even the consumer space, because our customers are now scaling.
32:46and we have many of them like DuckDuckGo also, that are relying on those APIs. So is this basically a product that competes directly with Glean? It does on the enterprise search side. We also provide web search APIs, our index to companies like DuckDuckGo and Winsurf that want to have general web search, but we also do company internal search and then we give those answers and then we build end-to-end solutions for our enterprise customers. Right. So, you know, I think you were watching the stream here. We were talking about the corporate data wars, I guess, for these enterprise companies and some of the challenges that faces for up-and-coming startups to access data at companies like Salesforce.
33:34And, you know, we had the CEO of Atlassian on the show last week, and that was another company that we had written about. You know, he sort of had a bit of a different perspective on it. But my question for you is, as it relates to the challenges to accessing data inside these larger tech companies, how much of an issue is that for you right now? It's not a big issue for us. I can see how it can become an issue for folks that, you know, wanted to use Slack in ways that, like, you weren't supposed to use it and you weren't using official APIs and things like that. But a lot of our customers are very motivated to get us access to the data.
34:13In some cases, we have to white glove them to get access, mostly not for legal reasons, but it's just like non, it's not hard work, but it is, you know. White glove meaning like you have to like really spend time to nurture these relationships to allow for these deals to happen? Yes, but also just, you know, when you get data from lots of disparate parts of the company and you try to bring all of that together, that's just additional work. And sometimes that work is more custom per customer. But for the most part, we work with companies that are already in the cloud and have like an S3 bucket on AWS and things like that.
34:51And then we can storage over, you know, decades of the archives, for instance, with DPA, which is a large press agency. So I do want to ask you about this. So this pivot now that, you know, you've now moved towards enterprise search and there are tons of companies still out there who are saying we are going to be the next Google search. I mean, there's still sort of seemingly, I guess, where you.com seems to have been sort of in 2023, which is that we're going to take over that playing field. And it seems like you guys have come to the conclusion that, hey, this is not where we're going to stand out.
35:25My question for you is, A, why did you make the decision? What was it about that challenge that you deemed too significant? And B, what do you think about companies who still say that they can do that? Yeah, I think in general, consumer spaces end up in monopolies or duopoly type of situations. What we found is that on the consumer side, there are there is Google. And then if there is a second player, duopoly type of situation instead of a monopoly type of situation, it's probably ChatGPT, right? They have billions and billions of players. And you didn't want to compete with that. And that is that is a tough space, but more important than in what was the problems that were in consumer, it's the opportunities that were in enterprise.
36:11When you have companies saying, we love your answers, but we want those answers on our own website and our own rules, you just, and that's where the revenue is, you follow the revenue as a startup. How much revenue are you guys generating on the enterprise side now? Not sharing it, but it's been growing significantly. And compared to a lot of other folks in our space, the multiple is actually quite reasonable on our valuation. And one question I did want to ask you before I let you go is your view on browsers. You know, we saw Atlassian buying the browser company, and this was a conversation I had with them about their rationale there.
36:49Still a little bit confusing for some people, right? But, you know, people are sort of saying browsers will be the next frontier for the search wars in some way. Do you have any interest in getting into the browser space? Are you looking to develop a browser at all? What's your view on that? Browsers are indeed a very interesting piece of technology. In many ways, it was like search four or five years ago, very underestimated, and a lot of people paid attention. But if you think about the higher levels of abstraction, it used to be that we cared about the operating system, and now that doesn't really matter because we do almost everything in the browser.
37:28Zoom is one of the few apps that I still have that isn't just all in my browser and even hang out mostly in the browser. So you can get it to pop out. It's like a lugger. It's like, oh my God, we did something. That's right. And so if you can control the browser, it's an extremely powerful place to be. But it's also a place where you have to have an insane amount of trust with users. And a lot of these companies do like they had companies that say, oh, let us just look at everything you do in your browser. It's just like there's so many, even if you're the CEO, there's so many, you know, company internal things that you would see that, you know, like employee salaries and all of that, it's hard to build that trust with a lot of people to let them use your AI browser to then actually do the interesting bit, which is agentic browsing.
38:14So it sounds like this is not a space you want to get into, then. It's exciting, but you have to think about, you know, we're focused right now on enterprise search and API capabilities, composable infrastructure, and so on, that we believe actually has more staying power than the LM itself, because you just can't open source, for instance, a search index for the web. And so it would be a distraction for us. But it is an interesting space, especially when you believe, and I do believe that, which is more and more, we'll have more and more AI agents surfing the web rather than people surfing the web.
38:48And most companies will fight that because it questions advertising revenue. And so you can circumvent some of that if you just pretend to be the user inside that browser that they're using. Right. Great. Well, Richard, thank you for coming on the show. Like I said, you.com is a company that we have been following for a couple of years. And so it's fascinating to see the progression that the company has gone through from consumer to enterprise. And I thought maybe browser would be the next step. But hey, you're telling us you're focused on enterprise. So we are looking forward to seeing how that evolves and shapes up.
39:23That was Richard Socher, the CEO of you.com. Okay, well, that does it for today's show. A reminder that we are on this stream Monday through Friday at 10 a.m. Pacific, 1 p.m. Eastern. I want to thank Amazon Web Services, who is our presenting sponsor for this production. And I want to thank you for tuning in. We really do appreciate your viewership. I am already excited for our next show tomorrow. And so until then, bye-bye for now.
From the publisher
The Information’s Jing Yang talks with TITV Host Akash Pasricha about the latest developments in US-China trade negotiations in Spain, where TikTok and NVIDIA are now central to the talks. We also talk with reporters Cathy Perloff & Ann Gehan about Perplexity's cautious approach to advertising and e-commerce, and we get into AI valuations with Menlo Ventures' newest partner Deedy Das. Lastly, we talk with You.com's CEO Richard Socher about his company's pivot to AI enterprise search.
Articles discussed on this episode:
https://www.theinformation.com/articles/search-has-its-goliath-could-richard-socher-be-its-david
TITV airs on YouTube, X and LinkedIn at 10AM PT / 1PM ET. Or check us out wherever you get your podcasts.
Subscribe to:
- The Information on YouTube: https://www.youtube.com/@theinformation4080/?sub_confirmation=1
- The Information: https://www.theinformation.com/subscribe_h
Sign up for the AI Agenda newsletter: https://www.theinformation.com/features/ai-agenda
