Brian McCullough (0:04)

Welcome to the tech meme. Right home for Monday, July 14, 2025. I'm Brian McCullough. Today Google snatched the windsurf acquisition out from under OpenAI in a big new aqua hire. SpaceX is investing in XAI, Chrome OS and Android to merge, but for real this time. Was there a new sort of deep seat moment over the weekend? And example number 74 of how YouTube is now the king of video entertainment? Here's what you missed today in the world of tech foreign to borrow from the football chant, it's happened again. It's happened again. Another big AI accuhire Remember how OpenAI was going to acquire the AI coding startup Windsurf well over the weekend. First, the exclusivity period on OpenAI's $3 billion offer to acquire Windsurf expired and within hours Google swooped in and hired Windsurf CEO Veren Mo, co founder Douglas Chen and some of windsurf's R&D employees to focus on agentic coding efforts at Google owned DeepMind. Now the information was reporting that OpenAI and windsurf talks ended after Windsurf raised concerns over how its tech would fit into OpenAI's agreement to share tech with Microsoft. More on that in a second. Google is paying $2.4 billion for a non exclusive licensing fee that will pay for Windsurf's technology and multiple years of compensation for Windsurf staff coming to the according to a person with direct knowledge, it will not take a stake in the company. Google said the deal will generate a return for at least some of the startup's investors, according to a person familiar with the deal. Employees with vested shares will receive cash, but those without vested equity, many of whom joined in the past year, will not receive a payout, according to a person who spoke to leadership at Windsurf. Meta Platforms also held talks with Windsurf about a deal or acquisition, according to a person familiar with the matter. Other large tech companies have paid huge sums of money as part of similar transactions in which they hire much sought after AI talent without outright acquiring their companies. Last month, Meta invested $14.3 billion in scale AI while also hiring the startup's CEO Alexander Wang and several of its top employees. Last year, Google hired Character AI co founders Noam Shazir and Daniel Defreitas while paying the startup a $2.7 billion licensing fee, which was used to pay out shareholders. And also last year, Microsoft paid a roughly $650 million licens to inflection to hire most of its staff, including its co founders. The unconventional agreements have allowed the big tech companies to quickly hire top AI researchers without going through the lengthy review process of a formal merger and quoting TechCrunch notably, Google is not taking a stake in Windsurf and will not have any control over the company. However, as part of the deal, Google will have a non exclusive license to certain Windsurf technology, meaning the AI coding startup remains free to license its technology to others. And then back to the details about how the deal with OpenAI reportedly blew up, Quoting Bloomberg When Cerf didn't want Microsoft to have access to its intellectual property, a condition that OpenAI was unsuccessful in getting Microsoft's agreement on, people Familiar said that was one of several sticking points in Microsoft and OpenAI's ongoing talks about the AI company's efforts to restructure into a commercial entity. Microsoft's existing agreement with OpenAI says the software giant is entitled to access the startup's technology. OpenAI thought it had an agreement with Windsurf and almost announced the acquisition in early May, according to people familiar with the deal. There had been a signed letter of intent and Windsurf investors were given what's known as waterfall agreements, notifying each party how much money they were expected to make, the people said. End quote. Again, in my 25 years in tech I've never seen a talent hiring spree like this, but also it's been years since there's been a horse race in tech like this. You probably have to go back to the dot com era to see this level of frenzied jockeying for position. Quoting M.G. siegler this is all not a great look for OpenAI. Whereas the reports about the ongoing OpenAI Microsoft negotiations are just that, reports of goings on behind the scenes which they can brush off or quasi deny. This deal is a very outward facing sign of Microsoft's level of control in the relationship. OpenAI had a deal to acquire another startup and Microsoft effectively killed it. At the same time, you have to wonder if part of OpenAI isn't okay with the deal being killed. Beyond the aforementioned issue with Anthropic, Windsurf's main competitor, Cursor seems to have most of the momentum in the market right now and keeps raising money or getting offers to raise money at ever increasing sums. And the fact that Thrive Capital is both a key investor in OpenAI and cursor seems interesting if nothing else. At the same time, one has to wonder if Windsurf isn't okay with this new deal on some level because Cursor keeps setting new price comps for such startups. That said, it's undoubtedly not great news for the company to be losing their co founders here, but at least someone internal their head of business is stepping up to lead the company company going forward, which is still said to be retaining, quote most of its roughly 250 employees per the information. Oh yes, and they're getting paid some percentage of Google's $2.4 billion for their troubles. But back to OpenAI, one has to wonder if they're going to have even more M and a trouble going forward, at least until they're able to renegotiate with Microsoft. If they're able to renegotiate with Microsoft, any deal seems like it could be at risk now and it may just ice a bunch of would be deals such as OpenAI trying to be competitive in whatever deal of the mom moment Meta is trying to do to help them get back into the AI race. Yes, the IO deal with Jony I've went through, but that's apparently only because Microsoft wasn't worried about competing in future AI hardware, which seems short sighted. Perhaps if the space actually takes off Slash works, end quote over the weekend, sources were telling the FT that XAI was preparing a new fundraising round, its third round of fundraising in just two months, targeting a valuation of $170 billion to $200 billion, which would be up 10x from its May 2024 valuation of $18 billion. And then within hours, sources were telling the Journal that SpaceX has agreed to invest $2 billion in XAI as part of XAI's $5 billion equity fundraise announced by Morgan Stanley last month. Quote Elon Musk has repeatedly mobilized his business empire to boost the AI startup, which is racing to catch up with OpenAI. Earlier this year, he merged XAI with X, combining what was a small research lab with a social media platform that helps amplify the reach of its Grok chatbot. The merger valued the new company at $113 billion. Musk has long used SpaceX to support his other businesses. He personally borrowed $20 million from the company to help fund Tesla early in its history and also use SpaceX's equipment to set up his tunneling venture, the Boring Company. More recently, he turned to SpaceX for a $1 billion loan around the time he was acquiring what was then called Twitter, which he paid back shortly after taking it out. SpaceX's investment in Xai may pose risks for Musk's space company. SpaceX's revenue has jumped in recent years, but the company is investing billions to develop a new rocket called Starship. The experimental vehicle is behind schedule and has suffered multiple setbacks this year, including consecutive failures in flight and a large explosion during an engine test last month. SpaceX recently had more than $3 billion in cash on hand, the Journal has reported. End quote yeah, and remember way back when SpaceX merged with that solar company Musk also owned? It's all one conglomerate really, people. It should just be called Musk Inc. Or something. Or as Hojina Kojima said on X, a Musk company has agreed to back the Musk driven company that recently acquired one of Musk's companies, whose product will soon be integrated into products developed by Musk's other company. End Quote TechRadar has an interview with Google President of Android ecosystem Samir Samat, who just sort of casually mentioned that Google is probably going to combine Chrome OS and Android into a single platform, quoting the Verge. Samat, who's responsible for Android's implementation across mobile wearables, XR TV and auto, added that he's, quote, interested in how people are using their laptops these days, suggesting he may be adding a new string to his bow. The comment is the closest thing yet to official confirmation of a change that's been rumored for months. In November 2024, Android Authority reported that Google is, quote, migrating Chrome OS over to Android with the aim of competing with the iPad. That process may have already begun, with Google itself announcing last June that Chrome OS will now be developed on large port of the Android stack. Chromebooks can already run many Android apps. Meanwhile, Android is getting a little closer to Chrome OS this year with new features including a desktop mode, resizable Windows, and improved support for external displays. Google bringing its two operating systems under one roof makes a lot of sense on paper, allowing it to speed up feature development and work on improving functionality on tablets, where Both its current OSS lag behind Apple's iPad OS. Then again, that's been true for a while. A merger of the two platforms was reported 10 years ago in 2015, and the Verge wrote that it makes perfect sense to bring them together two years before that. This is a change that's been a long time coming, but that means it might be a long time still. End quote ever wonder what ChatGPT and Claude are actually doing with your conversations? Have you ever even stopped to think about that? We all know Alexa listens to us and recommends products based on our conversations. Meta retargets us based on our browsing and engagement history. But now in this new AI era, there's a new privacy problem to consider. Think about what else we tell these AI platforms. Our thoughts, our dreams, sensitive questions, business ideas, etc. They take all this information, tie it to your identity, and then sell it to various third parties and governments. ChatGPT literally has the former director of the NSA sitting on their board right now. That's why I've started using Venice AI who is sponsoring today's podcast. Venice AI is a generative AI platform that is private and permissionless. They utilize leading open source AI models to deliver text, code and image generation to your web browser. There's no downloads, no installations of anything. Venice AI doesn't spy on you or censor the AI Messages are encrypted and your conversation history is stored only in your browser. This is a cause I can get behind. If you want to join this too, and you want to use AI without fear of handing over your most intimate thoughts to a corporation or the government, you can get 20% off a pro plan using my link. Venice AI Techmeme and Code Techmeme that is Venice AI Tech Techmeme and Code Techmemee while single AI agents can handle specific tasks, the real power comes when specialized agents collaborate to solve complex problems. But there's a fundamental gap. We have no standardized infrastructure for these agents to discover, communicate with and work alongside each other. That's where Agency agntcy comes in. Agency is an open source collective building the Internet of Agents, a global collaboration layer where AI agents can work together. It will connect systems across vendors and frameworks, solving the biggest problems of discovery, interoperability and scalability for enterprises. With contributors like Cisco, Crewai LangChain and mongodb, Agency is breaking down silos and building the future of interoperable AI. Shape the future of enterprise innovation. Visit agency.org to explore use cases now that's a G N T C Y.org Moonshot AI, the Chinese artificial intelligence startup behind the popular Kimi Chatbot, has released Kimi K2, a 1 trillion parameter mixture of experts architecture model with 32 billion active parameters, which they say outperforms models like GPT 4.1 and Deepseek V3 on key benchmarks. Is this another little deep seek moment or Deep seek echo? Because quoting VentureBeat Kimik2 does not just answer, it acts, the company stated in its announcement blog. With Kimik 2, advanced agentic intelligence is more open and accessible than ever. We can't wait to see what you build. The model's standout feature is its optimization for agentic capabilities the ability to autonomously use tools, write and execute code, complete complex multi step tasks without human intervention. In benchmark tests, Kimi K2 achieves 65.8 accuracy on a challenging software engineering benchmark, outperforming most open source alternatives and matching some proprietary models. Their performance metrics tell a story that should make executives at OpenAI and Anthropic take notice. Kimmy K2 instruct doesn't just compete with the big players it systematically outperforms them on tasks that matter most to enterprise customers. But here's what the benchmarks don't capture. Moonshot is achieving these results with a model that costs a fraction of what incumbents spend on training and inference. While OpenAI burns through hundreds of millions on Compute for incremental improvements, Moonshot appears to have found a more efficient path to the same destination. It's a classic innovator's dilemma playing out in real time. The scrappy outsider isn't just matching the incumbent's performance they're doing it better, faster, and cheaper. And the implications extend beyond mere bragging rights. Enterprise customers have been waiting for AI systems that can actually complete complex workflows autonomously, not just generate impressive demos. Kimmy K2's strength on swe Bench Verified suggests it might finally deliver on that promise. Buried in Moonshot's technical documentation is a detail that could prove more significant than the model benchmark scores their development of the Muon Clip optimizer, which enabled stable training on a trillion parameter model with zero training instability. This isn't just an engineering achievement, it's potentially a paradigm shift. Training Instability has been the hidden tax on large language model development, forcing companies to restart expensive training runs, implement costly safety measures, and accept suboptimal performance to avoid crashes. Moonshot's solution directly addresses exploding attention logits by rescaling weight matrices in query and key projections, essentially solving the problem at its source rather than applying band aids downstream. The economic implications are staggering. If Muon Clip proves generalizable, and Moonshot suggests it is, the technique could dramatically reduce the computational overhead of training large models. In an industry where training costs are measured in tens of millions of dollars, even modest efficiency gains translate to competitive advantages measured in quarters not years. More intriguingly, this represents a fundamental divergence in optimization philosophy. While Western AI labs have largely converged on variations of Atom w, Moonshot's bet on Muon variants suggests they're exploring genuinely different mathematical approaches to the optimization landscape. Sometimes the most important innovations come not from scaling existing techniques but from questioning their foundational assumptions entirely. Moonshot's decision to open source Kimik 2 while simultaneously offering competitively priced API access reveals a sophisticated understanding of market dynamics that goes well beyond altruistic open source principles. At $0.15 per million input tokens for cache hits and $2.50 per million output tokens, Moonshot is pricing aggressively below OpenAI and Anthropic while offering comparable and in some cases superior performance. But the real strategic masterstroke is the dual availability. Enterprises can start with the API for immediate deployment, then migrate to self hosted versions for cost optimization or compliance requirements. This creates a trap for incumbent providers. If they match Moonshot's pricing, they compress their own margins on what has been their most profitable product line. If they don't, they risk customer defection to a model that performs just as well for a fraction of the cost. Meanwhile, Moonshot builds market share and ecosystem adoption through both channels simultaneously. The open source component isn't charity, it's customer acquisition. Every developer who downloads and experiments with Kimme K2 becomes a potential enterprise customer. Every improvement contributed by the community reduces Moonshot's own development costs. It's a flywheel that leverages the global developer community to accelerate innovation while building competitive moats that are nearly impossible for closed source competitors to replicate. End quote finally today, some Netflix analysis According to Nielsen, Netflix's share of the most watched shows on streaming dropped from greater than 80%. That's again share of the most watched shows in 2021 to only around 50% now. Quoting Bloomberg, this isn't all that surprising. Netflix has gone from being one of three major streaming services to one of seven or eight. Rival services have added tens of millions of customers over the last few years. As those services get to scale, their biggest hits can rival Netflix's biggest hits as well. Amazon, Apple, HBO, Max, Hulu and Paramount plus all have shows in the top 10 so far this year. Yet no other service delivers as many hits on a consistent basis as Netflix, because no other service is used by as many people or spends as much money. While Netflix's share of the top shows is slipping, it still accounts for more hits than every other service combined. It also accounts for more viewing than Disney, Hulu and Peacock combined as well. End quote. Yes, but as the New York Times points out, the real action is One battlefield over YouTube and Netflix are increasingly locked in a battle over TV, full stop in the U.S. nielsen says YouTube was 12 1/2% of all TV viewing time in May, above Netflix's 7 1/2 quote. The brutal truth is that YouTube is indeed the biggest competitor of Netflix at this point, says Jason Kalar, the founding chief executive of Hulu and a former chief executive of WarnerMedia. And YouTube's lead keeps getting wider. Two years ago, YouTube's share of TV time was roughly half a percentage point higher than Netflix's. Now it's 5 percentage points. Their strategies for success are very different, but in ways large and small. It's becoming clear that they are now competing head on. Top executives at both companies are beginning to mention each other in public more, sometimes dismissively, and the companies are veering into each other's turf, with Netflix executives showing an increased appetite for signing up creators who otherwise call YouTube their home and trying to explain why their business model would be better for them. Both companies are competing from a position of strength. Netflix's revenue in 2024 reached $39 billion and it has more than 300 million global subscribers, more than any other streaming service. The company is also hugely profitable. Netflix had more than $10 billion in operating income last year. YouTube, which is owned by Google, had revenue of $54 billion last year. The only media company with more was Disney. Moffat Nathanson, a media analyst group, projected that YouTube would eclipse Disney in revenue this year and described it as the new king of all media. The company does not disclose profits, but Moffat Nathanson estimated that YouTube's operating income was just under 8 billion dol in 2024. On average, YouTube has an audience of 7 million viewers watching off a TV set at any given moment during the day, more than Netflix's daily average of 4.7 million, Nielsen said. During primetime hours, however, when the highest concentration of viewers is watching TV, the margins are tighter. An average of 11.1 million Americans are tuned into YouTube on their TV screens at night this year while 10.7 million are watching Netflix, Nielsen said. And cross quote Nothing more for you today. Talk to you tomorrow.