[music] I see a large IPO on the horizon.
You're surrounded by journalists. Hold your position.
[music]
Overnight success.
The place [music] of 3,000 double, right?
There's misinformation. Hard clearing order inbound. [music]
>> Let's just roll.
>> We are surrounded by journalists. Hold your position.
>> Come on. Get up.
Trust the experts here. We are experts.
I see multiple journalists on the horizon. Stand by.
UAV online.
>> Double blaze. [music] Triple blaze. Double kill.
>> [music] >> Five color is wins. Team deathmatch. We are experts. Triple blade. Let's just roll. Right.
>> Hard clearing order inbound.
We are surrounded by journalists. Hold your position. [music]
>> Strike one. >> Strike two. Activate golden retriever mode. [music]
>> Hard clearing order inbound. 5 on the horizon. Founder.
YOU'RE WATCHING TBPN.
>> Today is Friday, January 23rd. >> It's casual Friday. >> Casual Friday cuz I got the SemiAnalysis jacket on. Zoom in on this. Thank you. Normally I'm the casual guy. I thought I'd let Patel.
>> Thank you to Doug. Thank you everyone over at SemiAnalysis. This is fantastic. It's extremely comfortable and uh I look like I'm ready to go to our data center. I feel...
>> I feel like this is built for abling. You throw this on, you're ready to go. Uh see what the hyperscalers are up to. Uh maybe wire up some GPU racks, do whatever you need to do.
>> I I love it. So, thank you to everyone over at SemiAnalysis. Uh we have another we have another gift that came in the mail from a friend of mine, Alex Moore, over at the Arc phone case. Arc Pulse thing. It's got uh it's a phone. It's so heavy. Feel how heavy this is. [laughter] It's a [clears throat] It's a fantastic uh presentation.
>> It has a certificate of authenticity in metal. The Arc Pulse Exclusive Edition Titanium Opal Damascus.
>> The new iPhone has really been testing my case-free lifestyle that I've been running for...
>> Yes. Oh, yeah. You know, super bullish for cases. Like my my phone's only a couple weeks old, maybe months old, and it's just like completely destroyed on all the corners. It's really bad. Um, so born from a fascination with precision engineering and material mastery, the Arc Pulse Titanium Opal Damascus unites advanced metal energy with sculptural intent. Every case is forged from over a hundred individual layers of titanium purged with argon gas bonded at temperatures exceeding 3,000 Celsius. Uh, so thank you to everyone over at Arc with this fantastic case that has been delivered with its own case, series of cases. It's a uh sort of a Russian nesting doll situation. So, do you have news or can I tell everyone about ramp.com? Time is money. Save both. Easy to use corporate cards, bill pay, accounting, [applause] and a whole lot more all in one place.
>> Couldn't have said it better myself.
>> What did you want to chat about?
>> What uh what did you write about today, John?
>> I wanted to dig into the question that we had with Joe Weisenthal yesterday. What does China actually import? Uh obviously, this was off the back of Davos. Vice Premier He Lifeng highlighted China's determination to become the world's market. They say, 'Hey, we we're the world's factory, but we also can buy stuff. Send us your stuff. We'll buy it.' But what are they gonna buy? So, he said boosting domestic demand was now on top of their economic agenda. This is in their their 15th five-year plan, something like that. Uh and we were going back and forth like what are they actually importing? Uh we were saying semiconductors, but you know, obviously there's there's chip deals and and GPU bans. Now, interestingly, uh, China is like the number one trade deficit for China is semiconductors and it has been since 2005. So, um, they wanted they they they they want to buy semiconductors. Obviously, US trade policy oscillates. Um, but semiconductors are the single biggest deficit for China. In 2020, even before the AI boom, China imported 350 billion worth of semiconductors. Uh, which was more than the value of the crude oil it it imported that year. And it's been the largest importer of chips since 2005 and accounts for huge chunks of revenue for Qualcomm. It's over 50% of Qualcomm's revenue I think and over 25% for Intel. So everything that you get from China, if it's an electronics product, it's some gadget, it has a chip in there, that's probably going to be made in the West, you know, fabbed in Taiwan or made in in an Intel fab, sent over to China, assembled and then sent back out.
>> and then sent back out. Exactly. So they they aren't imported there. Uh energy is a big driver of the trade deficit with China. Uh China imports 74% of its oil supply and 42% of its gas supply. Uh soybeans, which you mentioned, and iron ore are also big categories. The iron ore obviously goes into a lot of again the factory, the world's factory.
>> Why can't China make enough soybeans in China?
>> It must be climate related. They love making stuff.
>> Yeah, I don't know. That's a good question. We should we should dig into that. But um um luxury was what you said and you were right. Uh luxury is a clear net import category for China. Estimates are around 100 billion annually. Now it's not the it's not the biggest driver of their trade deficit or trade surplus. Um but it is uh and it's also not critically important in the way chips or oil or soybeans or iron ore are. Um but it is but it is an important story and almost all of Chinese luxury good imports are from European conglomerates. You got to think LVMH and Kering. Uh it's not strategically important really and uh the the foreign luxury market in China. So European conglomerates sending their goods to China. It actually completely fell off a cliff in 2024. Uh imports went down about 20% based on more more international shopping that was happening, travel and a domestic boom. So domestic Chinese luxury brands are on the rise. Laopu Gold is now drawing from the same customer base as Louis Vuitton, Hermes, Cartier, Bulgari and Tiffany. Gucci is closing stores. I think around 18 stores are closing. Uh while the local champion Songmont is seeing significant growth in their handbag business. Uh so China does want to be the world's market but they're also buying domestic a lot. Um the other dynamic that changed recently uh in the east versus west brand dynamic has been the acquisition of successful brands by Chinese companies with more competitive supply chain. So Morris Garages started in the 1920s and became synonymous with affordable British sports cars under the badge MG. Are you familiar with MG?
>> We got to pull up some videos of old MGs to show you. Uh my family actually had one when I was a kid. A really old one. Uh are you familiar with MG?
>> I think my grandparents also had one.
>> Yeah. No, it it was like a cool It was a really really cool like British lux not luxury, not high-end, but uh a true driver's car. So, we'll pull up one of these videos uh while I tell you about the brand. It's changed hands a ton. British Aerospace owned it at one point. So did BMW. Uh but the MG business just collapsed. Uh because Britain had a lot of, you know, sports car manufacturers. They had uh Jaguar and uh Land Rover and Aston Martin and there's a whole bunch of other companies. But the MGs, the old MGs, they still look great. Here's a video of one of them. Jordi, give me your review. Would you pick one of these up? I like the soundtrack, too. It's a good This is a good song for for an old retro video about an MG. These were these were iconic back in the day. Yeah.
>> I like it. It's got a nice silhouette.
>> Yeah. Um certainly certainly a throwback sort of a Porsche Speedster style. I like it. Um but eventually the company failed in 2005 and the brand was sold off to Nanjing Automobile Group which was rolled into SAIC Motor which is China's largest state-owned automobile manufacturer. And so the acquisition it actually worked and uh sales of new MG Motor models have picked back up. Since 2019, MG has been China's most exported car with more than uh 88% of sales happening internationally. So, they bought the brand. They...
>> uh Europe, Canada, basically everywhere except the United States. They sell them in in England. So, the MG brand is still well known in uh in Britain. And uh with the new uh models, they're selling pretty well. We can pull up a video of the newer MGs uh because they have a sports sedan. They actually have I think um like sort of minivan as well, but here's here's a review of the sport sedan. See the quad exhaust.
>> All of the world has never...
>> Look at this. Uh look at this. They have this crazy like tail fin that comes out. They have a lot of like fun gimmicks on here, but there's the MG brand.
>> Still alive and well thanks to China.
>> This is one of the best looking MG cars on the market. At the front, you can see the LED headlights with the striking DRLs, 19in alloys, sleek fastback profile with aerodynamic lines running down this 4.9 m long car and frameless doors. At the back, we have wide hips with flowing LEDs with not two but four exhaust tips. And surprise, they actually work because this Chinese car is not electric. Finally, I know. And after all that sportiness outside, in here you find...
>> Yeah, it does. The the rear end's nice.
>> 375 liters of space with a liftback design. So, it's super easy to load and unload stuff. So, this car is not just sporty. It's also daily friendly.
>> They also have a sports car that we can pull up that [snorts] has some very interesting features. Uh it has some scissor doors. It's electric. Uh there's there's a reveal video that we can play here. Look at this. This is the MG... It's like Sportster, Speedster, something like that. Cyberster. Cyberster. That's what it's called. The Cyberster. Look at those doors. That's kind of It's It kind of ribs. It looks...
>> Tyler. I mean, we've been We've been wanting you to to for your next car go up to kind of like establish dominance in the in the...
>> I don't know if I can the Chinese. Yeah. It just feels...
>> Yeah. Yeah. Yeah. Uh let's play the review of the Cyberster from Forest Auto Reviews. I I really like this creator. He does uh one minute vertical video reviews of cars from start to finish. So essentially compressing like what normally...
>> and this is an electric drop top sports car and...
>> he just goes through everything.
>> The electric thing aside, this thing is gorgeous. You have these awesome headlights. I get these gorgeous body lines. But I think my favorite part has to be the way the door opens. I just come over here, push this button, and it pops right out. So cool. But don't worry because if it's going to hit something, it's hard to get those doors BMW i8. That's the only option in America really at this point. They really are cooking here. They are paddle shifters. Paddle shifters in an electric car. It just changes your drive mode. But you could imagine with an update, you can do the uh like synthetic DCT, digital DCT, which has been very popular in the uh...
>> lot of storage space down there. Storage under here. Tons of storage back there. [music] More storage.
>> He really gives you like a full tour of the car in just 60 seconds.
>> Back here. In the back, you get literal arrows as your tail lights. This can do 323 mi on just one charge. It does 544 horsepower. And because it's all-wheel drive, if I push, you cannot buy these in the US. Do 0 to 60 in 3.1 seconds. And then he rips away. Boom. This is some traction. Back to a safe stop. You get Brembo brakes. And if this were sold in America, it'd be $41,500.
>> That's not bad, right? Like that's kind of a steal. Anyway, quickly, let me tell you about Sentry. Sentry shows developers what's broken and helps them fix it fast. That's why 150,000 organizations use it to keep their apps working. Um, so uh this this trend has been going on with American or you know storied brands, European automobile manufacturers uh selling to Chinese companies. Volvo changed hands from Ford uh to China's Geely Holding Group. Uh in 2010 the deal valued Volvo at 1.8 billion and it also basically worked out. Sales have grown from half a million vehicles in 2015 to over 700,000 vehicles in 2023. And when I told you this, you were like, 'What? Volvo?' And and I Yeah, I couldn't believe it either. Um, Geely also bought Lotus, the maker of the Lotus Elise, the Evija. There's a bunch of other sort of like more affordable sports cars, ultra light uh sort of tossable cars. Uh, and so Lotus is also another British sports car manufacturer to wind up in China. And so the the basic trend is like China excels at uh developing efficient supply chains, streamlined manufacturing operations. Uh and it doesn't stop with cars. They're doing it with TVs now. So on Tuesday, Sony announced that they would be spinning off its home entertainment business, which includes the TV brand. And uh uh to their Chinese rival TCL Electronics Holdings, Sony selling a 51% stake to TCL while the brand will remain Sony. Uh the display technology will be TCL. And so Sony's been falling behind Samsung, TCL, LG, Xiaomi in terms of TV shipments for a while. Uh, but the brand still holds a ton of value. Like Sony TVs were the best uh a decade ago. And still today, people like the the software a little bit more. And the Sony brand, it still has the aura of the the Walkman, the the PlayStation, like it's just beloved electronics. And you see the Sony logo, and it's just way more familiar than...
>> you think Japanese excellence.
>> Yes, you do. And and so uh Sony has been you know falling behind on the manufacturing front. TCL has been basically delivering a much stronger value prop on a price to quality basis and the panel technology is is is uh is loved and and very competitive. Um but there's just more faith in the Sony name. So the Chinese manufacturer is effectively pulling forward heritage and brand legacy by buying the name. Uh and this feels like a trend that will continue for a while. China's fantastic at quickly grinding down manufacturing learning curves, developing high quality products at affordable prices, but creating an iconic brand, it just takes decades and so it's better to buy than build. Like this is the one thing you buy instead of build. Uh and so Western companies after getting beat up on margins and seeing their market share slowly shrink. I think the Sony TV market share was like dropping to less than 1%. Uh while I believe Xiaomi is at like 6%. I didn't even know Xiaomi makes TVs.
>> phones and cars and apparently they make everything. Um it is it is such an interesting dynamic because you have like you know western brands and manufacturers that have just been saying like yes China is out competing us on manufacturing by like 2x we're getting destroyed. We need uh we basically need uh tariffs and things like that to protect uh to protect market share etc. Uh, but the one thing that they'll never have is brands, right? It just takes so long, you know, building up the Volvo brand. It's not just like an advertising budget. It's time and it's their safety record and all these things that go into that. Um, but of course, if you if you're out if you out compete just on manufacturing long enough and then you're and then you get these opportunities to actually buy the the the top brands, eventually you can actually own the full stack.
>> Yeah. Yeah. Uh, so we got to look at one more MG, their hypercar. It's a concept car. Probably will never ship, but uh I want your review on whether or not you would step into this and take it for a spin. This is the MG Hypercar or Supercar. This is from Supercar Blondie. And Supercar Blondie has an uncanny ability to get access to every concept car ever made. I'm constantly seeing Supercar Blondie content about cars that feel like they barely exist or or...
>> single seater like this is like this is like a competitor to the Valkyrie and the or something Valkyrie's...
>> Valkyrie's two seater but yeah it's like a crazy like I I mean this is probably just for show during a big conference something to turn heads and draw eyes but uh pretty pretty remarkable to see this from MG the company that made, you know, these open top roadsters in the 1920s. Anyway,
>> yeah, I mean, it just looks like uh looks pretty close to a Speedtail.
>> It does look like a McLaren Speedtail, which is also the center...
>> Also McLaren Speedtail. Uh inductive charging. Have you seen this?
>> So, there's a pad that when you drive over it, it will charge the battery. And I think that's also like a regular trickle charger.
>> Yeah. So, you don't So, it's still a trickle charger. 8 hours, but you don't need to plug it in.
>> And I think that's also happening with the the new uh Porsche SUV EV. Maybe the Cayenne EV.
>> Ferrari should pick that up. Ferrari loves to put put their trickle charge.
>> Tesla should pick it up. Tesla's the obvious one.
>> No, but but but Ferrari is notorious for being like, 'Oh, that thing you need to use every time you drive your car, we're going to put it buried in the trunk. You're going to need to crawl in.'
>> Yeah. I mean the uh the like the Tesla charging remember the the robotic arm the snake that was gonna do you remember this there Tesla was at one point teasing this might have been like a decade ago but they were teasing basically a a robotic snake like arm that could go and plug in the car like autonomously. So you just pull in and then at some point it would just go and plug in and they never shipped it. Are you're looking at me like I'm crazy.
>> I've never heard of that. I think they should also do the um you know there's the Chinese EV where the battery explodes at the center. [laughter] That thing is crazy. Can we pull that? Yeah. Yeah, we got to pull that out. You know, if you're next, you pull next to your your ops, your enemies, you just hit them in the shins with your battery. Yeah. Yeah. Uh Jordi, I want you to try and pull some of that up. But first, I'm going to tell you about Turbo Puffer. Serverless vector and full text search built from first principles and object storage. Fast, 10x cheaper, and extremely scalable. Uh [laughter] do we have any luck? Uh, John Arnold says, 'I don't know if Chinese manufacturers will ever make money, but uh, I came away not wanting to invest in any manufacturing business in the rest of the world.' And Ian Roundtree says, 'I'm going to keep investing in American manufacturing.' So, people are still still pushing it. Uh, there's also big news. Uh, Anduril is bringing a 1 million square foot union built campus uh, along with 5,500 jobs uh, to the city of Long Beach. The the the office of Mayor Rex Richardson, the Long Beach mayor, uh posted this uh over the next three years, Anduril will be building this. This is the largest private investment in Long Beach history, and it's a major vote of confidence in Long Beach's leadership in aerospace and advanced manufacturing. It's accelerated by 28 and uh grow Long Beach in action. Great work, Lex. Love to see it, says Grunt Daniel. Uh it's fantastic and what a banger. Thousand likes on this. Everyone's happy about the new Anduril campus in Long Beach. Uh anyway,
>> um I have this pulled up on Futurism. China unveils EV that can violently eject its battery.
>> Okay. Send that to the team. I'll tell everyone about public.com investing for those that take it seriously. Stocks, options, bonds, crypto, treasuries, and more with amazing customer service.
>> Okay. If you scroll down this video and we can see this in action and see what kind of situation...
>> Yeah. This is so wild. I remember when this video went viral. It's like [laughter] what what is the point? Oh, it's because if it crashes and it's going to burn, it'll shoot it out. But oh, that just seems like so so dangerous. [laughter] Who? But yeah, at the same time, people people, you know, watch these videos and they and they think like, oh, this is like widespread Chinese best practice when really like this is probably like their Gunpla equivalent, you know, just like a couple dudes. It is interesting that they're like, 'We don't just need to get it 10 ft away. We need to get it 30 ft away.' [laughter]
>> Yeah. Yeah. Because you could just like like gracefully drop it out or something or like or like slide it out sort of slowly with a little pusher, but they have to like shoot it out like a cannon. It's amazing. Anyway, Gemini 3 Pro, Google's most intelligent model yet. Next level uh vibe coding, state-of-the-art reasoning, and deep multimodal understanding. And speaking of Google, I think we got to watch the video about AI industry news. Every two weeks, one of these models is destroying someone else's model. And we we we we found a fantastically funny video on Instagram reels that we will play for you today, folks. So, let's play the re the com...
>> destroyed Pinterest and ChatGPT. [laughter] ChatGPT destroyed Perplexity and Google AI. [laughter] LinkedIn AI just ruined WhatsApp AI. We met Meta AI. IRCTC AI just destroyed Uber AI. Wow.
>> So good. [laughter] What is that last one? IRCTC AI. I've never even...
>> He's just making stuff up.
>> He's just making up acronyms at this point. But it really is so true that like there's just a time-honored clickbait title of like DESTROYED in all caps. I used to do it on my YouTube videos. Always went viral. It's it's it's been proven. I was I was taking credit for for inventing that format on YouTube. Brandon Garell put me in the truth zone and and said that I did not in fact create that. That was created [laughter] long before the BuzzFeed the BuzzFeed era and is is potentially time-honored. Uh anyway,
>> don't worry. I will I will drop I will drop the video link in the in the YouTube.
>> Please do. And uh and while you're doing that, I will uh pull up the linear lineup and tell you who we have on the show today, folks, because we have a great lightning round coming up. Uh we have Tuhin from Base 10, Bryce from Nominal, Max Spero, the AI slop janitor from Pangram, and then Russ closing out with LiveKit. Some fantastic uh funding rounds coming up. Linear of course is the system for modern software development. 70% of enterprise workspaces are using Linear with agents and you should be too. So...
>> that's right. What's going on with TikTok?
>> TikTok is it it's sort of the reverse honestly. It's very interesting. Uh you have this dynamic of like the American companies or the western companies uh MG and Sony selling to Chinese because they have great manufacturing and then TikTok you know we want America wants uh control over the algorithm over the data security and so you have a Chinese company that's now going to be operating in the US and the Wall Street Journal has a piece here on it. First I will tell you about Lambda. Lambda is the super intelligence cloud building AI supercomputers for training and inference that scale from one GPU to hundreds of thousands. So the agreement was negotiated to comply with a 2024 law requiring the company to do a deal to address US national security concerns. Of course uh it was it was a total like will they won't they for all of 2024 with TikTok and then also in 2025 we were constantly tracking like when will this happen? It took years and years and years. Like the actual TikTok should divest discourse is now like three or four years old. Lots of people have been working on this for a long time, but Rome wasn't built in a day and TikTok was not divested in a day. Uh but TikTok officially established a joint venture that would allow it to keep operating in the United States. The company said Thursday, resolving a years-long fight to address Washington's national security concerns. Under the terms of the deal negotiated by the Trump administration, the popular video sharing app will be operated by a new US entity controlled by investors seen as friendly to the US. Its data management and algorithm training on American users will be overseen by Oracle, the cloud computing giant that has safeguarded its data for the US for years and has close ties to the Trump administration. The deal was negotiated to comply with a law passed in 2024. President Trump delayed the implementation of the law a year ago after starting his second term to keep TikTok operating in the United States. He signed a series of executive orders to extend the deadline for completing a deal until it was met Thursday. Uh Trump said in a social media post, 'I'm so happy to have helped in saving TikTok.' He thanked Chinese leader Xi Jinping for working with us and ultimately approving the deal with a capital D. [laughter] Uh he could have gone the other way but didn't and is appreciated for his decision. Trump and TikTok's investors and allies uh pushed the deal through despite lingering concerns among lawmakers and security hawks that China could still influence the new entity through TikTok's parent ByteDance, which...
>> uh it makes sense that this took so long. It sounds really simple just divest but in reality you have to effectively rebuild an entire app, right?
>> Yeah. It's got to be it's got to be it's going to be like a new app figuring out the logistics of that.
>> Uh there's also just a ton of hair on the deal, right? There's a huge revenue share that's going back to ByteDance in perpetuity that uh is obviously reflected in the valuation.
>> Um and so and then the other the other factor is like if you have a US TikTok now and then you have normal TikTok. If I'm on the US TikTok, am I sharing content to international TikTok? If I'm on the international TikTok, is it sharing my content back into the US app? Yeah. How does that all work out? Yeah. And so there's a lot of stuff that still needs to be handled. You know, I I would say like most of the social media app, pretty much every social media app that you use, if you just cut off all international content creators, the experience on the app would get worse. Yeah. So, there needs to be like some content still flowing back and forth. But figuring out how exactly that works is um uh still in process.
>> Yeah. Uh, TikTok CEO Shou Zi Chew said in an internal note to employees, 'The majority American-owned joint venture will operate under defined safeguards that protect national security through comprehensive data protections, algorithm security, content moderation, and software assurances for US users.' Uh there was a the the whole TikTok debate really erupted when uh TikTok CEO Shou was in front of Congress and was not really acting like he was the boss basically. That was the main criticism. It felt like he didn't have full control over the entity because of course it's a subsidiary of ByteDance and um and some of the US lawmakers were pressing him on how much control he actually has. Did they send the right person to the congressional hearing? Uh but it seems like he's held on and he's done well. He's also Singaporean, by the way. Uh so, you know, easier to, you know, be a part of the new US uh TikTok organization um and and allay any concerns of Chinese control, which is uh certainly at the heart of the uh of the administration's goals here. So, who's in the deal? You got Oracle, you got private equity firm Silver Lake, you got Abu Dhabi based MGX. They will each own 15% of the new entity. While TikTok investors will own about 30%. Other notable investors include JD Vance's former firm Revolution and tech executive Michael Dell's family investment office. Dell's getting in the deal. Uh Vance spent a brief stint at the firm founded by AOL co-founder Steve Case during his time as a venture investor which preceded his 2022 Senate campaign. I didn't realize that uh JD had been at uh at Revolution. I I know he was uh he had a separate venture firm. Um but I didn't know that he was hanging out with Steve Case, the former co-founder of AOL. Uh Vance has said previously that the deal values the new entity at about 14 billion. A lot of people thought that that was really really low given uh TikTok's immense growth. But there is another side of this which is that TikTok I don't believe was ever monetizing or as profitable as its competitors YouTube and and Instagram and uh Mark Zuckerberg and uh and the and the Google team were not like exactly slow to move and launch competitive products and so a lot of the you see this continuously where uh Snapchat comes out, Stories is on a tear and you're looking at the Snapchat growth curve and you're like, 'This is going to kill uh this is going to kill Facebook. It's going to be the next Facebook.' And Mark Zuckerberg needs to acquire it. He puts in an offer, gets declined. Um and it looks like it's over. But then the Instagram team moved really quickly. They launched Stories and that effectively, you can see in the chart, Stories launches on Instagram and people stop moving on to Snapchat. And so a lot of people went back who had who were on Snapchat, they went back to Instagram and Instagram continued to grow and Snapchat it didn't flatline but it like definitely put a dent in their growth. And I think the same thing is true for TikTok. Like TikTok a lot of people who just want that format vertical video endless scrolling slop. Uh they can get that now. We we got we got American-made troughs all over the place everywhere the eyes can see. [laughter] So, uh, the investors are paying the US government a multi-billion dollar fee for arranging the deal, a concept Trump previously called a tremendous fee. Plus, interesting. TikTok said it had 200 million users in the US, up from its 2024 estimate of about 170 million users. So, decent growth. Uh, but I don't know. I mean, 200 million in the US is a lot. Like, that's pretty much everyone.
>> Yeah. I guess the question is will they ever get the rest of the US in the in the way that YouTube you could assume has like effectively...
>> you think they have like 300 mil.
>> Yeah, I would assume that that uh...
>> I don't know. There's a lot of Philistines out there. There's a lot of Luddites who are just like nah I want I want DVDs. I want VHS tapes. Anyway, Trump touted his popularity on TikTok earlier Thursday, posting on Truth Social that his posts on the platform TikTok get more engagement than posts on TikTok competitor Instagram, which is owned by Meta Platforms.
>> TikTok Trump said that TikTok helped him win.
>> I mean, that's always uh for what it's worth, that's always been the case, right? Like TikTok has always had allegations that they were botting. Yes. Like as a platform effectively. People would go on there and they'd be like getting a tremendous amount of followers, tremendous amount of just engagement broadly, far more than uh than they would have on Instagram with the same content.
>> And you know, some people would say, 'Oh, that's cuz the TikTok algorithm is so good.'
>> But uh uh and there certainly are a bunch of very real people on TikTok. I don't I don't mean to say that it's all botting, but uh the experience of using TikTok, a lot of creators will just naturally have six times as many TikTok followers as they do Instagram followers.
>> Yeah. Yeah. It does seem like the algorithm is set up to uh like serve you more content in a 10-minute session because you're more quickly scrolling. And then there's also accounting issues like uh on some platforms if you just scroll past something for even one second that counts as a view. On other platforms it might take three seconds. On other views other you know platforms it might be 10 seconds a minute you might need to watch the full thing. So there's always like accounting abnormalities there. Anyway, if you want to stream on TikTok because it's Americano now, you want to bring, you know, some Oracle focused content to TikTok on live stream focused content, get on Restream, one live stream at 30 plus destinations. [applause] If you want to multistream, go to Restream.com.
>> Uh, this might hit you like a sack of bricks. Yep. But, uh, breaking news, JD Vance is younger than Future.
>> That is a crazy stat. 22. Future is is straight up an elder now.
>> Oh, yeah. He's an elder, I suppose. Yeah, you're right.
>> He's uh not quite an OG, but Oh, no. It It goes It goes anc then OG then elder.
>> Uh Gen Z are arriving to college unable to even read a sentence, says Fortune, reposted by Uncle Wales.
>> See if you can read this. See if you can [laughter] read this.
>> See if you can read this.
>> Try and read this. Wow. He can't. He can't. He can't do it.
>> I I I just I need like some Subway Surfers or something.
>> I'm going to need some Subway Surfers [clears throat] to read this next sentence.
>> Anyway, Gusto, the unified platform for payroll, benefits, and HR built to evolve with modern small and medium-sized businesses. Get on Gusto. Um, we have we have some drama that's continuing to unfold in the in the payroll space. The Justice Department opens a criminal probe into Silicon Valley spy allegations. Subpoenas seek information on allegations. The deal valued around $17 billion recruited a spy inside a rival company. This is a huge story last year and then it's been pretty quiet and both companies have just kind of been chug chugging along. Um...
>> spending a lot on legal.
>> Yeah. So, the Justice Department opened a criminal investigation. Uh grand jury subpoenas were sent out in recent weeks by Craig Misakian, the US attorney for the Northern District of California. The documents show they seek information related to a spying operation that Deel allegedly ran inside a rival company. Rippling, an Ireland-based Rippling employee, in case you've forgotten, Keith O'Brien alleged in an affidavit filed in April that Deel CEO Alex Tangrija recruited him and gave him instructions for what information to take from Rippling. O'Brien alleged that other executives were involved in the spying plot, including his father, who is Deel's executive chairman and chief strategy officer. A spokeswoman for Deel said the company isn't aware of a criminal investigation, but is willing to cooperate with authorities. That's an interesting statement. The the Wall Street Journal is reporting this and and and you find out from a Wall Street Journal reporter. That's that's wild. Uh so the company has previously said, 'We deny any legal wrongdoing and look forward to asserting our counterclaim.' So Deel says we didn't do anything wrong. Of course, the affidavit from Keith O'Brien is is extremely dramatic. There's that whole like uh you know, oh send that watch to London. We're going to be like uh James Bond. Like it's a it's a really dramatic read and uh you can go back and watch some old TBPN episodes if you want to hear our full take on that.
>> Yeah, it's interesting. Uh they're not they're they're...
>> Yeah. I wonder I wonder uh what what information is actually available kind of publicly on this because like obviously in a situation like this you have both sides that are kind of like feeding the media their side of the story. Uh but I mean the facts of the facts here like the Justice Department did open this uh because because the the Wall Street Journal confirmed that the Justice Department opened the investigation. So um anyway,
>> I think a lot of people weren't expecting this to last actually uh have any type of like criminal component.
>> Yeah. I think a lot of people were just expecting like a lot of drama, a lot of you know and they would eventually just settle.
>> Exactly. Settle. And like the worst case was like, okay, maybe it's like a like a, you know, like a Series D size payment for Rippling, but they're both going to just keep grinding. Um, but it seems like it's going to the courts. So, uh, we will see. Maybe there will be, uh, maybe there will be, you know, courtroom sketches of these folks.
>> Uh, anyway, uh, Nick, the chief communications officer at OpenAI, uh, has some breaking news from CNBC. Uh, OpenAI chair Brett Taylor says AI is quote probably a bubble, expects a correction in the coming years. And Nick is he his camera roll is truly just filled to the brim with images of Sam Altman in various like like you know...
>> the guy has probably the largest collection of Sam Altman...
>> Rare Altman shirt. He has every photo that's ever been taken of Sam. Uh and and Nick just is looo...
>> Yeah. But uh I don't know this could be true like you know who knows I wonder what what more of the context was that...
>> yeah I mean a Brett Taylor quote taken out of context turned into a headline [laughter] and then and then and then it's like nowhere of course would Brett Taylor be like oh yeah I think I think LLM usage is going to fall off a cliff.
>> Yeah. Yeah. There's variety of different companies and kind of subcategories that could be overheated. But yeah, you could take uh you could uh you could easily uh dig into that. And he's like there's a bubble in like seed-stage valuations like I can't get an angel check in at less than 50 pre like you know the bubble can take a lot of different shapes and stuff. Who knows what he was actually commenting on. Anyway, CrowdStrike, your business is AI, their business is securing it. CrowdStrike secures AI and stops breaches. Uh, another clip from Alex Kantrowitz from the Big Technology podcast. He's taking shots at OpenAI according to Euchan Jin.
>> Think actions think actions speak louder than words. Going back to the original conversation we were having with, you know, Sam and others claiming AGI is around the corner. Um, why would you bother with ads then? So, that is I think a reasonable question to ask. Um but I I I think uh look from our point of view we have no plans at the moment to uh to do ads if if you're talking about the Gemini app right uh specifically I think uh we are going obviously we're going to watch very carefully what uh you know the outcome of what uh ChatGPT are saying they're going to do um I think it has to be handled very carefully but think actions speak louder...
>> he's post AGI the clankers need to advertise to each other. You got to do ads for the clankers. Ads...
>> I mean it I mean the thing here is like OpenAI has to not only like figure out how to actually implement ads within the product, but they have to build out all the advertising like infrastructure, the platform infrastructure in order for people to run campaigns successfully and at scale. So Google already has all of that. It is far e like it it is what do you think it's like a hundred times easier for Google to turn on ads in Gemini, right? They already have all the customer relationship. Literally any company that advertises online is already working with Google. And so it's really just like you can just flip the switch. And so I could easily, you know, don't have any inside knowledge, but I could easily see Google just like having it all basically ready to go and it literally just being like, okay, there we can launch this whenever we want.
>> There there is an interesting steelman here. Let me try and do it. So uh you need the helmet.
>> I I I might need the helmet, but I think I'm good for now. So, so Sam and others. Well, first let me tell you about advertising. Vibe.co where DTC brands, B2B startups, and AI companies advertise on streaming TV, pick channels, target audiences, and measure sales just like on Meta. So, uh, the steelman on Sam Altman claiming that AGI is around the corner, but we still got to do ads. Well, around the corner, even if that means 6 months, like if you have an incredible capital expenditure to get over that hump to create AGI, it's only 6 months away, but you got to do that last data center and you got to raise that last 50 billion and all of the investors.
>> Yeah, it's easy it's easy to say like we don't want to do ads when you have hundreds of billions of dollars in existing ad revenue coming in funding everything that you're doing. So anyways, I I think this is...
>> I mean we can also debate whether or not ads will exist post AGI. I would argue that they would. But uh what do you think Tyler?
>> There was that headline about how OpenAI maybe they're thinking about like taking a percentage of the of the share of like the discoveries. I don't know how they would actually do that or if it's like they're being serious at all. But that's like much more AGI pill right than ads.
>> Yeah. Yeah. Oh, that Yeah, that's a good point. So they're doing both. They're doing both. Uh anyway, New York Stock Exchange. Want to change the world? Raise capital at the New York Stock Exchange. Just just do it.
>> Uh apparently Vimeo uh almost everyone at Vimeo was laid off. Uh including the entire video team. I would assume that most most of their team so unfortunate. I have some fond memories of being a teenager when Vimeo was like really...
>> Vimeo was so differentiated. It was where you there was like a level of quality that it wasn't just...
>> I would go there to watch short films. In my case, it was like surf like really it was like if somebody was putting a ton of effort into making like a surf movie or you know some snowboarding movie, it would be on Vimeo.
>> Yeah. And they wanted higher quality footage. You could get higher bit rates 4K HD. Like they they were really on the front tier of that. Yeah, I remember watching a lot of uh ski videos there. There were some great ones and also just like it was a place where people would put their their short films. They had the Vimeo uh like you know awards with the little wreath laurels that you could rest on. Um but uh eventually, you know, YouTube just went everywhere all places. The ads went away for the people that care about that because they had premium subscriptions and uh ultimately the quality on YouTube just got better and better and better and now you can get 4K really high bit rate on on YouTube and it's indistinguishable from Vimeo. So fewer and fewer uh fewer and fewer um ways to make a case for you know consumer adoption essentially like I I I I certainly would not recommend anyone who's uh making video content uh distribute on on Vimeo. I I put everything on on YouTube because YouTube has a massive audience and it's discoverability.
>> Yeah, discoverability is just so high. Harry has the meme for the moment. He says but now we shall both surely drown said the frog. Lol, said the scorpion.
>> I don't think I don't I I don't think this is exactly right. Like, if if you buy uh if you're a private equity firm and you buy an asset like this and you and you lay off the team, like you're likely there are still people that are...
going to be trying to extract the value from that asset. It just might be a team on a different, in a different section of the firm or an outsourcing agency or something else. I would be surprised if this means that Vimeo is shutting down. And I would imagine that they're trying to continue whatever business relationships they have and whatever subscriber base they have.
Yeah. Um, anyway, Banjo says, "Vimeo was cool, but if you know about Vimeo, it probably means you're old."
And uh, that's true. That's certainly true.
Yeah, I know what Vimeo is.
What's the last thing you watched on Vimeo?
I have no idea. Yeah, but no, but I remember. Yeah, it was like, like artsy films would be on there or something.
Yeah, it was good. Uh, anyway, um, app loving profitable advertising made easy with axon.ai. Get access to over 1 billion daily active users and grow your business today. Um, what is, what is all this? So, uh, Matthew Zaitelland says, "My basic model of modern artificial intelligence labs is that they are like emotionally intense graduate school programs but with unlimited money." Matt Levine.
Matt Levine is writing about this. And Mary says, "My friend is dating a guy who works at one of the frontier research labs in SF, and his sole job is to think and come up with a big idea every six months."
So his work day consists primarily of long walks spent pondering. That's very Womanitis coded.
That's amazing. I mean, that seems like honestly in the age of research in the LLM world, like this is probably a good use of time as opposed to there are plenty of people that will, that will lock in and, and you know, improve, make minor improvements to the consumer application or the consumer side of the business, but um, yeah, in a world where you have, you know, infinite minds, uh, when you're a manager of infinite minds, what, uh, Satya was quoting...
Um, yeah, you're probably just walking, pondering, wondering, thinking about what, uh, what the next set of, uh, tasks to assign to your, to your infinite minds.
Well, Epic Games and Google apparently have a secret $800 million Unreal Engine and services deal, and the quote in their scoop quote, sorry, I'm blowing this confidentiality. What a crazy quote to give to a Verge reporter. Uh, but we're going to dig in. Uh, Addie Robertson and Sean Hollister have the story. A judge is questioning whether Epic Games and Google are settling their long-running antitrust fight partly because of a previously unannounced partnership involving the Unreal Engine, Fortnite, and Android. In a hearing in San Francisco today, the court revealed that Epic and Google have struck a new deal that apparently includes joint product development, joint marketing commitment, and joint partnerships. California Judge James Donato expressed concerns that the agreement, which he indicated would involve Epic quote helping Google market Android and Google quote, uh, Google's newly quote apps using Epic's core technology, could have led Epic to soften its demands for changes to the overall Android ecosystem. Donato allowed Epic and Google to keep most of the details of the plan under wraps. But during the hearing, he quizzed witnesses, including Epic CEO Tim Sweeney and economics expert Doug Bernheim on how it might impact settlement talks, revealing some hints in the process. Quote, "You're going to be helping Google market Android, and they're going to be helping you market Fortnite. That deal doesn't exist today, right?" Donato asked Bernheim, who answered in the affirmative. He also described it as quote a new business between Epic and Google. Sweeney's testimony cracked the mystery a little further. He referred to the agreement as relating to the metaverse, a term Sweeney has used to refer to Epic's game Fortnite. Epic technology is used by many companies in the space Google is operating in to train their products. So, the ability for Google to use Unreal, the Unreal Engine more... Wholesome. Sorry, I'm blowing this confidentiality. Sweeney said, oh, that's a great line. Um, yeah. I mean, we've seen a number, what was it? What is the DeepMind model that plays the games for you that I'm in the market for because I don't have any time to play video games. So, I need an AI agent to do it for me. Uh, they have a number. So, they have Genie 3 which generates basically a 3D world that you can walk around in.
Sim 2 will play the game for you. And both of those, uh, should in theory benefit from, uh, using Unreal as the backbone.
Elon's probably pretty excited about that as well, right?
Because he likes to play on accounts that are ranked, but he all the time.
Yeah. So, instead of needing to hire someone, you could, you could bot your account. People already do that with, uh, like aimbots and stuff. Uh, usually you get banned, but, uh, makes a lot of sense in terms of training Google's generative AI models. Use Unreal, create a bunch of virtual worlds, walk around in those, interact in those, use them as reinforcement learning environments for agents, and then bake that out into a bunch of different models that can be used for a bunch of different things. I'm, I'm, I'm super excited for all of this. Um, but, uh, it's, it's just funny the way it's coming out. Anyway, if you need a reinforcement learning environment, you got to go to Labelbox. Reinforcement learning environments, voice robotics, eval, and expert human data. Labelbox is the data factory behind the world's leading AI teams. So, Coinbase established an independent advisory board on quantum computing and blockchain. Quantum computing, if built at scale, have the potential to reshape entire industries. We were talking to Cathie Wood about this a little bit yesterday. From finance to healthcare to material science and national security. For blockchain, the stakes are especially high. The cryptographic foundation that secures digital assets today could be challenged by the advances of the coming years. At Coinbase, security is our highest priority and preparing for future threats, even those many years away, is crucial for our industry. Most people, I think, handicap the quantum thing at like 2030, 2032, 2040. I think Cathie Wood's estimate was 2040 for serious quantum adoption. Even if you get on a Moore's law type, uh, adoption curve. Um, but...
Yeah, Nick, Nick Carter's been super vocal on X about, about quantum just being like the biggest, uh, headwind to like further Bitcoin price appreciation.
Yeah. I mean, there's also this question, maybe I'm just like very, I'm like super out of the loop, but like if you can crack like whatever the Bitcoin or any like blockchain with quantum, like surely you can also just get into like, you know, banks or like tons of like actual like things that are like probably way more important than, than crypto. Like it's like if we have a world where you have like these incredibly powerful quantum computers, like is crypto actually the most important thing that you need to be worrying about or is it like the nuclear codes that are...
So surely those are stored.
Yeah, you're probably right with the nuclear codes, but, uh, in terms of just like fiat stored in a bank, there are a series of backups that are offline. There's even paper backups at certain points, and so, uh, like the financial system, although they probably don't want to talk about it, does have the ability to like effectively roll back. So if there was like some massive quantum hack and like all the Morgan Stanley or JP Morgan accounts or Goldman Sachs accounts were like drained, uh, they would just be like, let's revert, but you can't do that on a blockchain necessarily. So, uh, but there is still the dynamic of like what happens if you crack Bitcoin? Like how much are you stealing? How many new bitcoins you...
Yeah. I mean, really, if you like steal any Bitcoin, Bitcoin just goes to zero, probably the point.
But that's obviously still a threat. Even if you just have some malicious actor who's just like, I'm going to take out a short position on Bitcoin, tank it, and profit from that. That, that could potentially, uh, be disruptive, certainly disruptive to Coinbase business. So, good to see them taking it seriously. Anyway, MongoDB, choose a database built for flexibility and scale with best-in-class embedding models and rerankers. MongoDB has what you need to build what's next. Uh, let's go to Scott Bessent. He was complaining about the food in Davos and he did an interview on Real America's Voice, I guess, with Jack. Uh, and...
So, so again, if you're looking at Davos as though it's a bunch of kick streamers...
Yes, that is the correct...
Has effectively been just clip farming this whole time.
Okay, let's watch the Scott. They don't like carnivores here in Davos with the, with the food. You're, you're not a huge fan of the food here, I heard, by the way. You know, I, I, no, and no, I, you know, I, I, I don't know if you remember a couple years ago, one of the recommendations from the, the Davos elite was we could all be eating bugs and insects.
Yeah. I tell you, after, after a couple of days of the food here, I, I may switch to, to bugs and insects.
You take the insect food up.
Unless that's what it is.
Or, or I'm going to bring snacks. They did. Do you remember they were really, people really pushing this? They were saying, "Oh, the protein content is so high."
And I think humanity rejected it. Uh, well, you know, you know the story of Magic Spoon.
Were they going to put cricket? No. Originally...
They did a previous company called Exo.
It was an American insect food company. Uh, they made some protein bars using cricket flour from pulverized house crickets. Uh, house crickets. I guess...
That sounds like you're taking the pets.
They did a Kickstarter in 2016. They raised $55,000, which was more than their $20,000 target. Uh, they raised, uh, $5.6 million in Series A funding. I wonder from who. There's an article in the journal. There's an article in Business Insider. Nas invested.
Legendary rapper Nas invested in a company that makes protein bars out of crickets.
Okay, so on the Exo Protein site now, all their bars are sold out. No, a couple you can still get banana bread cricket breakfast.
Company was acquired in 2018. Aspire Food Group.
You can get salted caramel.
And you can just get one pound of cricket powder.
They should just get into selling live crickets. Just, just deliver them. Get your enemies' addresses and just deliver them. Uh, no, but, uh, so I, I, I think...
I think they would, they, I mean, the team certainly found a niche, a weird like, you know, breakthrough. You're going to get a lot of earned media with a, with a bizarre ingredient like that, but the ultimate traction, I think people just have an aversion to bugs and, and they, it just conjures, you know, images of bugs crawling around like it.
Yeah. And so they sold the company, but then they started Magic Spoon, the not cricket-based...
Which just immediately ripped.
It ripped. Fantastic business.
And I'm sure you've seen a ton of, uh, you know, podcast ads about Magic Spoon and everyone's familiar with that company. Uh, so, you know, they, they, if you're in crickets, if you're in bugs, pivot to cereal, I guess. Uh, and that's what Bessent's saying. He's saying he wants Pop-Tarts. Pop-Tarts basically Magic Spoon. Magic Spoon might even make Pop-Tarts. I don't know. Uh, Anthropic published a new constitution for Claude. The constitution is a detailed description of our vision for Claude's behavior and values. It is primarily written for Claude and used directly in our training process.
Was it written by Claude? Did Claude write its own constitution?
That's a good question. You'd think yes.
Rahul over at Ramp said basically TLDR is I love you, you love me. We're a happy family with a great big hug and a kiss from me to you, quote. You say, "You love me, too."
You really know the Barney song.
I didn't just sing it as a kid. I studied it.
Um, uh, more importantly, yes, BuzzBallz is incredibly important. Thank you. Is selling a $35,000 diamond engagement ring shaped like its drinks. The ring will be auctioned on eBay starting February 3rd. Uh, so I don't, how is it, how is it actually going to be? Is 35k like the opening, opening bid or, or maybe it's, it's $35,000 worth of diamonds. It cost them $35,000 to make it and they're going to put it on eBay. Starting price 100 bucks and then it'll get bid up, I guess. I don't know. Yeah. Um, proceeds from the sale will benefit a heart-related charitable cause.
It's kind of odd for an alcohol brand to be like, we're, we're...
I don't know. That's cool. I mean, if you're, if, if anyone out in the audience is wanting to propose and like doesn't really know how to, is struggling to figure out like the right kind of like moment or way to propose.
This is like a BuzzBallz out of the box.
That's the woman in your life.
If she catches it, she has to marry you.
Uh, is this rage bait? Is this good marketing? Do you think this is good for Buzz strategically? What do you think?
Uh, I don't know if it's, I mean, I, I, I find plenty of moments to talk about BuzzBallz already. Okay. So unnecessary. I mean, it is, it is gimmicky, but the whole brand is sort of gimmicky to begin with, and so it doesn't feel out of context. It doesn't feel like they're trying to be a serious company and then they did something wacky on the side and I'm confused. This like feels within the, the, within the BuzzBallz lineage, within the BuzzBallz, uh, brand world very much so. Uh, anyway. Whoa. If you're heading to Cisco AI Summit, be sure to bring some BuzzBallz because on February 3rd, Cisco is bringing together leaders from NVIDIA, OpenAI, AWS, and more to discuss the future of the AI economy.
The whole thing will be live. We will be there for...
Do you know who came up with the idea for Buzz?
Absolutely not. I have no idea.
A public high school teacher.
What? Uh, Merrilee Kick, a former public high school teacher, has become one of America's richest self-made women after selling her ready-to-drink cocktail business BuzzBallz for at least 500 million.
What started as a side hustle has now transformed into one of the biggest brands of the ready-to-drink cocktail industry. Kick founded BuzzBallz in 2009. Inspired by a simple idea while grading papers by her pool. I thought I shouldn't have this glass container out here. I should have a plastic pool-safe type of cocktail. From this spark of inspiration, BuzzBallz was born. Fun, high-proof cocktails served in colorful plastic spherical containers. BuzzBallz quickly gained popularity and is now sold at supermarkets, liquor stores, and convenience stores.
She says, "I've been living the American dream. We've built a legacy. We've become a contender in a space where women never went." Brutal.
Says the brand has grown intentionally, distributing to 29 countries with an estimated annual revenue of $500 million.
This is an incredible story.
In April of 2024, the drinks firm, uh, Sazerac acquired BuzzBallz in an all-cash deal, 500 million. Though Kick suggests the figure is much higher.
Kick is an absolute dog. It has also cemented her place among America's richest self-made women, with Forbes estimating her net worth at 400 million after taxes. So yeah, of course...
She's going to 10x that. Get involved. I'm sure she's cooking on her next idea.
Run it up.
Uh, Kick's journey is remarkable because she never raised money from investors. She bootstrapped her business on the show.
She used a small inheritance, maxed out.
Where is the Founders podcast episode about her? And took out a loan from a local community bank to get started. I scraped and scrambled. She says, "I took every bit of every penny I could find and poured it into the business."
Amazing story.
Her unique company started making a profit in the second year with 1 million in sales and 100k in profit. Not bad. By 2014, the brand was expanding quickly and the drinks were sold at major retailers. By 2019, annual sales were over 100 million.
Get her on the Davidson Center.
Wow. She, I mean, she's elite. A key to BuzzBallz success was owning its supply chain.
She vertically integrated on day one like...
Like most, like DTC founders in, in, in California, like can't figure out how to get off a co-packer.
She moved production of the patented plastic spheres and the spirits used in BuzzBallz in-house to ensure the brand's quality and reliability. Despite many investor offers over the years, Kick held on to her company until she found the right partner. I wanted somebody that was going to come in and have big guns.
She explained Sazerac, which owns over 400 brands, brought the resources and the expertise to scale BuzzBallz.
The Kick and her family are still part of the business. I didn't sell because I didn't like what I was doing or wanted to leave. She explains, I sold for the exponential growth and because it's selfish to hold it back. It really has legs. From a teacher grading papers by the pool to a multi-millionaire, Merrilee Kick's story shows the power of a good idea and the determination to make it happen.
That is an incredible story.
We got to get her on the show.
I love it.
Uh, let me tell you about Console. Console builds AI agents that automates 70% of IT, HR, and finance support, giving employees instant resolution for access requests and password resets. And before we move on to our, uh, to our lightning round, let's debate Kylie Robinson's latest post. She says, "Do you think if the AI companies IPO this year we'll get insert frontier lab arenas like the Crypto.com Arena? What do you think? Should there be a Gemini Arena, an Anthropic Arena, an OpenAI Arena? Should they do this? Is this good for the brand? They're already buying, uh, Super Bowl ads, so it doesn't seem out of the question. What do you think the probability is?" Yeah, I feel like, uh, these type of, uh, advertising on like an arena like that, you typically see this with like somewhat commodity-like products that are sure that are in large markets, but are, uh, you know, in the case of like Crypto.com, they're just compet-, like the products Crypto.com is very interchangeable with Coinbase, right? Give them Bitcoin or whatever digital assets you want. And so they're really just competing on brand entirely. And like the way that you're seeing the market evolve, like I don't, I don't know that...
Yeah, they're competing on technology, like how does the model actually answer? What are the benchmarks? They're competing on product, like how does the app work? How, what, what are the mechanics? What can you do? Can you generate images in there or not? Can you generate video? How good is the video? Uh, does it, does do deep research quickly? Like there's a million different things. How's the voice mode? Like they're, they're definitely playing like a product race right now. There is a marketing race.
Yeah. I guess the question is at the same time, was there ever, was there any of like the OG search companies? Did they ever get out and did Yahoo ever have a...
I don't know. It's interesting. But, uh, I do know that, uh, it's interesting that she mentions the Crypto.com Arena here in Los Angeles, the former Staples Center, because Crypto.com is not a public company. So there's nothing that says that you can't buy an arena before you go public. Uh, who knows, maybe we will have a Vanta Arena by the end of the day. Automate compliance and security. Vanta is the leading AI trust management platform and I'd love to head to the Vanta Arena no matter where it lands. Um, anyway, we have our Lambda lightning round starting now with Baseten. We have Tuhin, the co-founder and CEO of Baseten in the StreamYard waiting room. Let's bring him in to the TV Ultra. How are you doing?
Thanks for having me again.
You've been busy. Give us the news.
Yeah, look, we've had a crazy year. Um, you know, we're 10x over last year.
Um, 10x alert. Hit the gong for the 10x.
All right, back, back to you.
Last time I was on here, I was telling you about our Series D. We've raised our Series D now. We've raised $300 million at a $5 billion valuation, uh, led by IVP and CapitalG with participation from NVIDIA and a few others. Um, look, um, it's just been, I just to remind you guys where, like, what we do.
Um, AI infrastructure platform. We focus on production-grade inference serving the fastest-growing companies in the world. So Abridge, OpenEvidence, Cursor, Notion, Clay, Gamma, Writer, you, you know, you look at that, that crop of companies and how they're growing and, you know, we're playing a supporting role in making sure that they can run their models as fast and reliable as possible. Um, and you just on the back of that, um, you know, we realized how big this could potentially be.
Take it. Yeah. Take us through how quickly enterprises are ramping up running their own models. Like if we go back to 2022, uh, Google had PaLM internally. They hadn't really launched Gemini. OpenAI had GPT-3 and then they launched ChatGPT. Uh, but Microsoft had, you know, they were vending GPT-4 through, uh, through Bing and stuff like that, but big companies were not really running their own large language models. Is that roughly correct? But now it feels like we're in the hundreds of companies. But out of the Fortune 500, like how many, how many companies are actually running their own models at this point in time?
Yeah, look, I, I, I think it's still relatively small but growing, growing rapidly. I think, you know, if you go and look at like when we go and speak to enterprise customers, it's still incredibly early. But I think what they are doing is that they're looking at, you know, these other companies that, you know, they, you know, you and I would probably agree that are the future Fortune 500 kind of eating their lunch and if they don't like figure out how to move at that pace, I think they'll just be left behind. And so like what we, we focus really, really heavily on, you know, those really fast-growing companies because we, we think not only are they going to be the future Fortune 500 but the Fortune 500 are going to look at them as like how do we bring these models up to speed as quickly as possible.
Yeah. And so I mean that makes perfect sense in the case of, uh, of a company like Cursor. Obviously they're, they have a ton of training data. They have a need for a custom model. Uh, are we years away from someone like a Coca-Cola or, you know, developing their own or how does that, how does that play out?
Yeah, I, I, I think look, there's two things that are happening, right, which is like if you go look at OpenAI, Anthropic adoption at these, at these companies, it is, there is definitely, it is definitely there. Um, once you've proven out these use cases internally, then you just have to think about better, faster, more specialized, more secure, running on-prem, all these sets of problems. And so, you know, that is coming. Like is it like six months away, is it 12 months away, is it 18 months away, who knows. Um, I think the other trend that is pushing, um, the enterprises online is RL, which is like how, how do you take these open-source-based models, um, and RL them and get them to be as good if not better than, um, these frontier models for very specific tasks. And, you know, that's a big focus for us as well. We recently made an acquisition around that, um, but we do think that'll be a big driver of enterprise adoption of these models.
Walk me through the pushback from a large enterprise, uh, that says, yeah, it's great, I, I can train a custom model on my business, but I looked at the smaller models for a variety of benchmarks. And it feels like the big models just sort of are better at the small models' tasks as well. Like I should just use the latest and greatest big model for everything because the frontier is better at the specialized tasks as well. It just one-shots things.
Yeah, totally. Well, I, I think the, um, and I, I think honestly a lot of the times they'll be valid there. You'd be like, "Yeah, like I, I, I don't disagree with you." I think what, what you will see though is that as the frontier is getting better, open source is already, um, you know, if you go look at the open-source models that are out there today, like the GLMs, the Qwens, and, um, and the DeepSeeks, it's not like they are 18 months behind.
You know, some would argue that they're, you know, within a quarter and for certain tasks are better. The complexity actually comes around honestly from enterprises. Not so much like can I get a better outcome. It's, hey, you know, these large, um, lab companies have massive inference teams.
Um, do I have the skill set to be able to run the service at scale within my premise with all my enterprise requirements?
That's ideally where we can help in and be that partner to them to bring them online.
Okay. So, walk me through, uh, what it looks like when a Fortune 500 company says, "Okay, we're getting off of just a generic model. We want to do something that we have control over the inference. We're working with Baseten. We're doing, uh, you know, we're doing some custom RL. There's going to be some proprietary data in there. Who's involved? Are they hiring you or a consultant to set up the RL environment? Who's doing the training?" Walk through all that.
Yeah, I think just engineers and infrastructure engineers. I think that, that is like one of the realities right now is that I don't, engineers who don't know how to grapple with AI just, you know, they don't, I don't know if they exist in the future to be honest. Um, but, um, that being said, like for a lot of engineers, it's like, hey, how can we make it as easy as possible for you to take this custom model, deploy it, scale it up either, you know, in our cloud or within your own cloud. And so that is, you know, the job of the software. We have a, a really amazing forward-deployed team that is very happy, very happy to help, um, if necessary, but our goal is also just, you know, to enable them to do these things. And, and that is happening. I think it's early though, especially the enterprise.
From, uh, from your view, do you think we need more Neo labs, like based on the conversations that you have with customers?
Okay, look, more models are great. I think more people developing models are amazing. Look, I, I'm making a bet, uh, with my career, with this company on, on the long tail of models and that there are existing models outside of, you know, two amazing companies. Um, and so I think more people developing models just gives consumers and developers more choices down the line. And like, look, do we need more of the same? Probably not. But like, you know, all, all these companies have their own takes, um, on, on, on the problem.
Uh, help me understand your current thinking around model routers, routing inference across, uh, big heavy-duty models that might be very expensive, like using all parts of the parade of frontier. Uh, what are you seeing on, uh, on the inside? On, uh, are, are companies trying to handle parceling out the workloads themselves or do they want that to be something that's off the shelf that you sort of provide? What are some of the trade-offs that people are considering these days?
Yeah, look, I, I think it's a spectrum. I think, you know, the, the most advanced companies are 100% doing that. They are breaking up, they're breaking up their, their tasks almost like model by model and then figuring out how to route. I think less sophisticated folks, um, or people earlier in their journey, yeah, probably a better way to put it, um, are relying on other people to do that. And, you know, the model, the model companies are doing this themselves. I, I think, um, the place where I struggle a bit is like, you know, some of these routing things are just, you know, routing for failover, which is, hey, model just stopped working.
But like we, we believe that like reliability like should, should be table stakes for serving models. And, you know, ideally you don't have to build across what happens when X model provider falls over, Y model provider falls over, and have that, and that is a different routing problem. Routing based on capability is happening. It is happening outside the model. It's happening inside the model. And as the long-tail model comes on, come online for specific model tasks, I think there's more and more going to be the case that that is being handled by an inference company or inference platform like Baseten as well.
Yeah. I mean, that's certainly happened in the database market. Like you have a big Postgres installation, you don't have a MySQL installation there as like a backup, but you might have a Redis caching layer in front of it, right? Yeah. Um, so, uh, what is, what, help me understand that at the hardware layer? Uh, I know you're, you have some insight into NVIDIA's strategy here. I'm very interested in, uh, how you or other partners, other, uh, folks in inference might be thinking about the future of NVIDIA's like big powerful racks versus their more legacy chips that might be depreciating, but you could still run a great model on it versus some of the more exciting stuff that's happening with Groq in the future.
Yeah. Yeah. Yeah. Look, I, I don't have a ton of strategy. They're very good. I, I could hypothesize. Okay. You know, look, they're amazing partners to us. I, I think we, we are chip-agnostic and we think, look, every, every, um, task and it will have different requirements from a latency perspective, from a cost perspective.
Even what type of model runs on it. And so, yeah, so like we use A100s, we use A100s, but we also use H200s and GB200s. And as, as you know, we get these, you know, new, these new types of chips or accelerators, we'll be working on them as well. Like if you think about the NVIDIA Dynamo stuff, you know, what are they solving for? It's like, you know, it's kind of breaking out prefill and decode. You're using GPUs for prefill, you know, the compute-bound problems. You can saturate the GPU. You can do batching and have really good throughput, but then you have the LPUs or like the Groq chips or the decode, which are memory-bound. You're not doing a lot of new math per token. Why this is hard is that that's a pretty complex, um, orchestration problem between, of handling, um, workloads that are doing stuff on GPUs and on LPUs. I think NVIDIA's, you know, obviously built amazing software around this, um, to break out prefill and decode. Dynamo is the name of the software that, you know, we work pretty heavily with. But like I think that would just be a new type of chip that the chip providers, um, provide. And, you know, I think, um, one thing I don't think you can understate is how much, how much, how powerful, um, NVIDIA's supply chain is, CUDA is, CUDA is, and, and their, their ability to, um, you know, be dynamic with architectures changing. And I think, yeah, um, you know, we, we have very, very much bought into that ecosystem and what that enables, especially for inference and customers.
Can you give us your take or any insight on how companies at the application layer are kind of wrestling with, uh, the, like laziness specifically? Like everyone at this point has experienced, you know, product, like, uh, asking, you know, basically, uh, you know, wanting a product to be able to do something that, you know, it can do and then, and then, uh, you know, running into this like laziness element. What, uh...
What, what do you mean by laziness? I, I don't quite get that.
Uh, I would say like, you know, thinking about, uh, let's, let's pick the, you know, biggest, uh, not to pick on anyone, but like the biggest consumer AI app, ChatGPT, right? Like everyone has experienced asking ChatGPT to do something that, you know, it can do and like sometimes it will just kind of like circle around or, or like not really get to the thing that you want it to do, um, even if you're paying for it on like the Max plan or something like that. And so, and, and so I would, like, I would assume that every company at the application layer that gets to the point where they're caring about margins is like starting to get to the point where they're like, okay, like you don't always, because sometimes the model can be lazy and you're like, great, like...
You got me, you got me what I needed and I can just move on. Yet other times it's like wildly frustrating because it's like, you know, asking if, like, it's like asking an employee to do something at five o'clock and they're like, "Oh, like, you know..."
I, I mean, isn't that like, to, to me that's like what defines a great application layer company, right? Like the...
Yeah. The ones that aren't good are just wrapping around, you know, GPT and you, or, or whatever model. And, you know, you as a user are left to figure, like, deal with that laziness. I think the best application layer companies, you know, be...
It's like serving the right amount of intelligence for the task and, and setting up the harness and how to use it for the problem that your user is asking you to solve for them or you were trying to solve. Like, you know, I, if Open, if, if a company like OpenEvidence was just, you know, doing something the lazy way, that would not only be really, really expensive, but they wouldn't do a really good job of solving that problem for their users and it wouldn't be a great business in the long term. Like, you know, what makes these companies so good is how they use the models, how do they use multiple models, and how they architect that to, you know, the thing the user is asking them. And then, then layering in how do we do this efficiently so we can make a business out of it. And like, you know, to me there is going to be a great divide in solving that laziness problem. I think it's a good way to put it actually.
You said two things that I'm sort of having trouble reconciling. You said that, uh, you're chip-agnostic, but then you also praised the, the, the dominance of the CUDA ecosystem. Uh, is it, is it getting easier to implement LLMs in a way that run on multiple chipsets? We've obviously seen, uh, really strong performance from DeepMind on the TPU stack and Anthropic's talking about TPU now. Um, but for a long time the narrative was like, oh, it's going to be really cumbersome to re-platform off of, off the CUDA ecosystem. How are you thinking about, uh, yeah, just multi-chip architectures these days?
Yeah, I, I mean, look, I, I think it is getting easier, but it's not, it is not easy, you know.
Welcome to entrepreneurship.
But, you know, like, you know, it's a, and like, look, just sitting on top of like, you know, that's why like NVIDIA is such a great partner for us. It's like sitting on them doing the stuff. You know, we were, I mean, did you guys play FIFA or anything like that when you were growing up with Madden? Like...
I was more of Counter-Strike, Call of Duty.
Yeah.
It's pretty amazing that when we are doing stuff for GPUs today, we're downloading from this website that looks exactly the same as the NVIDIA driver page.
2000. But that is also the amazingness of it, of like how rich that is and how powerful it is. And I think that is, you know, yes, it is getting easier and like, like a diversification is great everywhere downstream, but also like NVIDIA is just amazing at what they do. And, and, and, and, and being able to run a model on AMD does not necessarily mean that it will inference at a lower total cost per token or vice versa might be a little higher on certain models, blah, blah, blah. Uh, and so, yeah, there's all these different trade-offs.
But that's why companies come to Baseten, correct?
Totally. And that's why the open-source ecosystem is so important, right? That's why, you know, like, you know, you, you mentioned databases earlier, like I think similar to databases, at the, in the fullness of time, open source has the fastest runtimes. And so I think that will, like with NVIDIA, you will see that, which is like, you know, do you know who's going to be really good at running stuff on NVIDIA chips? It's going to be NVIDIA. Who's going to be really good at running stuff on AMD chips? It's going to be AMD. And so working with these providers of chips, I think, to get the best runtimes from them, um, is very important. And I think like this cross-compilation stuff, while important, you know, I, I'm a little skeptical, I'd say. It's just like, you know, like it's first principles for me to think that we would be better at running, um, software on NVIDIA chips than NVIDIA or, or better...
Yeah, at the lowest level obviously there's all the orchestration software stuff that we are building that we think is very important, but we're also very, very invested in those ecosystems.
Well, I believe in you. I think you can do it. But no, I, I take your point, of course. Uh, well, congratulations on all the progress. Thank you so much for taking the time to hop on the stream and, uh, have a great rest of your day.
Sure, you'll be back on any day now.
Hopefully not. Um, thanks for having me and, um, appreciate you guys.
We'll talk to you soon. Goodbye.
Phantom, fund your wallet without exchanges or middlemen and spend with the Phantom Card. Up next, we have Bryce.
Bryce from Nominal. I believe Bryce is the brother of the previous guest from Nominal, Cameron McCord.
Got it. Okay. Well, you know, you build a business with someone for a while, people just assume you're brothers, but, uh, anyway,
Second spouse is what I like to say, but...
