Another AI Sorta Acquisition as Character AI Founders Head (Back) to Google"

Today on the AI Daily Brief, character AI becomes the latest company to be sort of acquired by a big tech firm. Before that in the headlines, does this big Wall street dip have Anything to do with AI? The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, follow the Discord link in our show. Notes welcome back to the AI Daily Brief Headlines edition. All the daily AI news you need in around five minutes. It is very likely that by the time you've listened to this, you have heard a bunch of coverage about what's going on in the stock market right now. If you haven't, honestly at this point, I would suggest you turn this off immediately, go outside, enjoy your life, come back in a couple days and we'll talk about markets again then. But assuming that that's not you and that you have been paying attention, there has been a huge dip in markets today. Many stocks, particularly big tech stocks, open down 9, 10, 11% this morning. They the NASDAQ overall at the time of recording is down 4%. And so the question that we're kicking off this Headlines edition of today is whether any part of this actually has to do with AI and this question that people have been facing about a bubble or whether it is all just part of a larger macro story. TLDR While there are some small catalysts, I think that this is a big macro story. But let's look at Nvidia where there is at least some reason to think that part of it is not just what's going on with the macro, but a specific announcement as well, investopedia writes. Nvidia stock plummets on reported delay of its new AI chip amid global sell off, reports started coming out that Nvidia's Blackwell system would be delayed at least three months due to design flaws. Given Nvidia's central place in the AI ecosystem, that will have implications for Amazon, Alphabet, Meta and more. Once again, this report came from the Information and came out at the end of last week. The source was that Nvidia had told Microsoft, who is one of its biggest customers as well as another cloud provider, about the delay. And one of those employees at Microsoft told the information design problems apparently arose unusually late in the production process, leaving Nvidia scrambling to work with tsmc, who is its chip manufacturer, to fix the problem. The information writes. If the upcoming AI chips, the B100, B200 and GB200 are delayed three months or more, it may prevent some customers from operating large clusters of the chips in their data center in the first quarter of 2025 as they had planned. This may have downstream implications for how fast software can evolve. In early trading this morning, Nvidia shares were down almost 8% and have ranged there pretty much all day. Like I said though, even with this specific catalyst and maybe Nvidia would have been off some because of it. Overall, this seems like it's part of a larger story. It's not that people don't think that this AI skepticism might have something to do with this. It's more just that that is very clearly part of a much bigger set of issues, including, according to Dario Perkins, for example, managing director at TS Lombard, global recession risk with central banks perceived to be behind the curve, the unwind of the yen carry trade and risks of a major intensification of Middle Eastern tensions. To say nothing and I'll add this one as many analysts are of the slightly chaotic US presidential campaign still looking at the mag 7 tech stocks, many of whose gains have been driven this year by their AI strategy. Those stocks saw about a 9% drop at the open that had come down significantly throughout the day. Now, even if this is macro, there is still this looming AI question. Over the weekend, for example, the Guardian published a piece called why have the Big seven tech companies been hit by AI Boom doubts? The Guardian pointed out that the shares of those companies had fallen by around 11.8% from last month's peak, and that was even before this. We've talked extensively about whether the market is in a repricing period when it comes to AI and their perception of its return on investment and frankly a lot of this feels to me like just a natural recalibration. Analyst Dan Coatsworth, for example, said expectations have arguably become too high for the so called Magnificent Seven. Their success has made them untouchable in the eyes of investors and when they fall short of greatness, out come the knives. Another analyst Angelo Zeno, said valuations were getting to 20 year highs and we were due for a pullback as well as a pause to digest some of the gains we've seen over the past 18 months. There will be, I think, a lot more to analyze on the macro and stock side of the equation. For now, let's move over to the private markets where an Nvidia challenger, at least that's how the information is describing them, called Groq, not Elon's Grok G R O Q has raised a $640 million Series D at a valuation of 2.8 billion. The fundraising was led by BlackRock one of the things that GROK is notable for is its extremely fast inference. You might have seen, for example, a demo of someone on Twitter X showing just how fast one of these new open source, state of the art models running on Grok can be. Following up on some stories that we've previously covered, you'll remember that AI ad that Google was running at the beginning of the Olympics that raised a ton of consternation. Well, that has officially been pulled, a Google spokesperson said while the ad tested well before airing. Given the feedback, we have decided to phase out the ad of our Olympics rotation. Finally today, a set of five Secretaries of state have written to Elon Musk urging him to fix his GROK for spreading election misinformation. In an open letter, the secretaries from Minnesota, Pennsylvania, Washington, Michigan and New Mexico said, as Secretaries of State whose offices and 37 million constituents were recently impacted by false information provided by your platform, we are calling on you to immediately implement changes to XAI Search Assistant Grok to ensure voters have accurate information in this critical election year. So far as I can tell, Elon has not commented on that yet, at least not on X. But I will certainly be keeping an eye out for it. For now though, that is going to do it for today's AI Daily Brief Headlines edition. Next up, the main episode. Today's episode is brought to you by superintelligent, the platform that helps teams maximize AI. Super is of course the platform that we've been building that pairs fun, fast, practical video tutorials with step by step instructions to get you actually using AI and from there unlocks information about hundreds of use cases that show you how people are actually getting value out of AI right now. Now we have just launched super for Teams. This is a new add on experience that allows teams to share more information about what AI they're using, how it's working, and how to get more value out of it. Whether your company is 25 or 2500, superintelligent is going to be the best platform for unlocking information about how to get the most out of AI right now. If you'd like to learn more about the Super Intelligent Teams offering, go to Besuper AI Partner and send us a note so that our team can get right back to you once again. That's BeSuper AI partner. Today's episode is brought to you by Venice, the leading AI companies store your entire conversation history and attach it to your identity for forever. That's every question you ask, every answer you receive, every image you generate, every thought you share with the machine. It's all being spied on. If you trust all the company's hackers and NSA board members that will ever have access to your AI conversations, then rejoice, for you are well served. For the rest of us, Venice is an alternative. Venice is a powerful AI app for text, image and code generation that respects you as a sovereign individual and believes privacy and free speech are not only human rights, but necessary for civilizational advancement. Private, permissionless, and uncensored, you can try it for free without an account. AI Daily Brief listeners receive a 20% discount on Venice Pro. Visit Venice AI NLW and enter the discount code NLWDailyBrief. That's NLW Daily Brief, all one word welcome back to the AI Daily Brief. One of the companies that there has been a lot of speculation about recently is Character AI. I've said previously on this show that Character AI is for me, the AI startup where I most know that I am not the target audience and so its usage numbers always surprise me. The long and the short of it is that a ton of people spend a ton of time every day interacting with AIs on character AI. Some of the AIs that people interact with are of course, representations of famous people. Others are just new characters that they've invented. You can see when you land on the homepage that they suggest that people do things like practice a new language or plan a trip or practice interviewing or write a story. But whatever people are doing, they are doing a lot of it. People are literally spending hours every day interacting. But of course, in AI land, all of that interaction costs a lot of money. And so one of the big questions is, would Character AI be able to remain independent or would they have to cozy up ultimately with one of the big tech companies? The last we discussed this was about a week ago when there were reports that Elon Musk had been in conversations on acquiring Character AI as part of his XAI startup. Now, Elon denied those reports, leaving the future of the company still in question at the time. As I said, it wasn't that Character AI was at risk of running out of money imminently. It was just that either they were going to have to raise an additional amount of money pretty soon or end up landing with one of the big tech cos. At the very end of last week, the answer became clear as Character AI was the latest company to have the sort of acquisition, but sort of but not exactly that has come to characterize the AI space. Just like the deal between Microsoft and Inflection and a more recent deal between Amazon and Adept. It was announced on Friday that Google would be hiring the Character AI co founders and many of the team members, as well as licensing its model. This was also notable because the founders of Character had previously been Google employees. From the information, Google has agreed to pay a licensing fee to chatbot maker Character AI for its models and will hire its co founders and many of its researchers. Character's leaders told staff on Friday that investors would be bought out at a valuation of about $88 per share. That's about two and a half times the value of shares in character 2023 series A, which value the company at $1 billion. As part of the deal, CEO Noam Shazier and President Daniel DeFreitas are going back to Google after leaving to start character back in November 2021. Now, unlike in the inflection deal where a huge portion of the team seemed to head over to Microsoft, in this case around 30 of the roughly 130 team members will be headed to Google, while the rest will remain with Character AI. Those 30 represent the part of Character's team that are working on model training and voice artificial intelligence. Reports are that Character itself will switch to open source models such as Llama 3.1 rather than in house models. And really, this seems to come down to a couple big things. On the one hand, it is a reflection of the cost of running an AI startup, particularly when it comes to one who's using custom models. Character had also held talks with Meta and xai, which the Information framed as a recognition of the steep cost of training conversational AI models with a relatively small base of paying subscribers face. The other side of this is the big tech acquirer side. There clearly seems to be a goal here of avoiding antitrust scrutiny, but it's not exactly clear that that's going to work. For example, the FTC is currently investigating the Microsoft inflection deal to decide whether it should have actually been considered an acquisition. Response in the AI community has been sort of surprised on a general sense, but not surprised in specific, investor and CMO Kieran Flanagan writes. Google hiring Character AI founding team is pretty wild. Character AI got a bunch of usage, but doesn't seem to have a path to revenue growth if their founding team is willing to jump ship, the information writes. Did Google just pay $2.5 billion to hire character's CEO? The piece reads. Google may have just paid one of the biggest signing bonuses in the history of Silicon Valley to hire two AI researchers who left the search giant three years ago Basically the point of this is that Google is paying a 2.5x premium on the previous valuation of the company, although what's not clear is how much is actually going to these founders. In other words, what package they got. Investors are getting out with the shirts on their backs, but but obviously this kind of sucks for the employees that are not being invited to go over to Google. And overall to the information, it just seems pretty surprising. They wrote it's been obvious for a while that Character would have to sell itself, so the main questions were the buyer and the price they'd be willing to pay. While millions of people apparently used Character's product to talk to all manner of chatbots, it was never going to be easy to turn that usage into revenue. Character generates money by selling subscriptions to its chatbots, but a number of Character's users are young, which complicates things. And second, some people use the app for romantic roleplay, but those types of chatbots are actually prohibited and Character says it removes them. There was also no indication that character had invented LLMs that were on par with those of Google or OpenAI. The company touted its ability to run its chatbot cheaply, but that's not enough to convince us this was anything but an Aqua hire. If nothing else, they write, this deal shows just how important AI talent is to Google. Now another piece that you may have seen me take some issue with on Twitter was this one from 4chan, which was titled Google's hiring of character AI's founders is the latest sign that part of the AI startup world is starting to implode. The piece is by Sharon Goldman, who is far from a cynical AI critic when it comes to journalists. And my issue was this characterization of implosion. Sharon's argument in follow up conversations was that she was referring specifically to the portion of the AI field that is trying to support its own models and finding it very, very costly to do so. As she writes, Character AI is not alone in having a tough time in an industry in which eye popping amounts of fundraising is required in order to survive due to the massive cost of computing power to train the AI models. My issue was the framing of this as an implosion rather than the natural consolidation process that happens a couple years into any new startup movement. I think there is a lot that's interesting here and worthy of discussion. It is certainly another indication that trying to compete with custom models is a very, very expensive proposition and one that cuts off all but the most extremely well resourced companies and even frankly some of Them as well. To the extent that's the argument, I completely agree. But I also think that this idea of implosion suggests that somehow this wasn't the default likely outcome. All of these startups like Inflection and Adept and Character AI now that have had these weird aqui hire type deals have been competing in extremely difficult areas. Inflection was, we can't forget, making a bet on this personal artificial intelligence, some sort of interactive experience which was in and of itself a total bet on a new type of behavior that didn't necessarily have a lot of precedent. In other words, it wasn't clear that there was even room in the market or demand for that particular type of consumer experience. Same with Character AI. Character AI is also betting on a new type of consumer interaction. Now notably, Character AI found that there was some amount of demand for this type of interaction. It just wasn't, at least at that time, or at least for that team, a cost effective business to service that type of demand. Perhaps a more interesting question is whether Character AI really needed to be pursuing its own custom models. Was that actually the way that this type of startup was going to work? Or is there a different model available? Unfortunately, this time we're not going to be able to find out because that talent is being absorbed back into the Borg of Google. But it's not at all clear to me that the differentiator, when it comes to an experience like Character AI is trying to create, is an underlying model difference versus an overall experience design. We will certainly be able to see going forward whether Character AI is able to rebuild something for the people that stick around, how it compares to for example, Meta's attempt to integrate these sort of character based AIs into existing platforms like Instagram. And of course we continue to have this question of whether these deals will just become the norm or whether the government will come down on them as just trying to get around the rules. My biggest concern frankly here is not what it says about hype in the AI cycle so much as what it means for early employees at these companies. It doesn't do a lot to incentivize people to work at startups if they anticipate that basically the exit path is just founders being bought out and going back to big tech anyways. That is the story though for here. It a really interesting one no doubt, but that is going to do it for today's AI Daily brief. Until next time, peace.