EP 566: Google acquihires Windsurf, OpenAI and Perplexity go for browsers and more AI News That Matters

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business and everyday life. Just about every big tech company was in the AI news this week. Some for good reasons, some for bad, and some news that was just kind of straight up unbelievable. So we had Google Aqua Hire hiring windserve after OpenAI's acquisition deal failed. OpenAI and Perplexity are going straight for agentic browsers and Meta Literally is paying someone $200 million from Apple to come join their super intelligence team. These numbers each week, I'm not making them up. It's getting wild. So if it's hard for you to keep up with what's happening in the world of AI and how it impacts your company or your career, don't worry, that's what I'm here for. We do this every single week, doing the AI news that matters on Monday. And there was a lot of news that definitely mattered this week. So let's get into it. What's going on, y'? All? My name is Jordan Wilson and welcome to Everyday AI. This is your daily livestream podcast and free daily newsletter helping everyday business leaders like you and me not just keep up with what's happening in AI, but how we can use this information to get ahead, to grow our companies and our careers. This is unedited, unscripted, like states the realistic and artificial intelligence. And each Monday, from our Monday through Friday live stream, we recap the AI news. Because there's always a ton going on and it's hard to keep up what's real, what's not, what's fluff, and that's what we do. We just give it to you straight. So there's a lot to get into. For the AI news for the week of July 14th. Let's start at the top. OpenAI is reportedly preparing to release an AI driven browser within weeks, directly challenging Google Chrome's market dominance, according to a report from Reuters. So the new browser could potentially reach ChatGPT's 500 million active users, threatening a key pillar of Google's ad business, which relies heavily on Chrome's user data. So OpenAI's browser will reportedly feature an integrated chat interface, keeping some user interactions within a chatgpt like environment, rather than directing users to external websites. That's pretty big. Yeah, just might not be visiting websites very much in the future. So the browser is designed to enable AI agents to perform tasks like booking reservations or filling out forms on behalf of users. Streamlining online activities. So OpenAI's move is part of a broader strategy to weave its AI services into users everyday personal and work lives, increasing its access to valuable web behavior data. The browser will be built on Chromium, Google's open source browser code, which also underpins Microsoft Edge, Opera and also Perplexity's Comet browser. So if you want to know more, we actually did just have an episode on Friday, episode 565 with Crystal Hu, the reporter from Reuters who broke the story. So if you want to know more on this pretty big story, make sure to go check out Friday's episode, episode 565. I'll say this, here's what it means. This is pretty big, right? And we've seen it now from Perplexity and from Chat GPT, you know, two of the bigger players in the game. So I would assume we might see something similar from Claude as well from Anthropic. Here's what this ultimately boils down to. I think probably the future of work is, is the final frontier, at least for now. If we're using physical computers, might be the browser, right? You could say, oh, ultimately it's going to be, you know, voice and we're going to have, you know, pop up displays in our, you know, glasses or whatever, which I'm sure. But for the majority of people still doing knowledge work sitting in front of a desk for the foreseeable future, I think it's ultimately about the browser, right? There's a reason why Google paid apple more than $20 billion to be the default search engine, because when you control a user's first action when they go to the web, you control a lot. It's not just the data, but it's also the ability to monetize off that data. So pretty exciting times I guess, right? Because the other thing is I think with these computer using agents that have been rolling out, they've probably have left a lot to be desired, right? Even OpenAI's operator, it's greatly improved since they upgraded it to the O3 model that was running it versus the previous G GPT4O. However, it's still just kind of using computer vision and kind of, you know, clicking around based on a screenshot of what's happening on the screen. When you go to the browser level and you're running this AI locally in your browser, it's much more intuitive, right? It's not that same, you know, kind of taking a screenshot of something. Anthropic was one of the first with their computer using Agent and then we got operator from OpenAI, we got project Mariner from Google, But I think ultimately what's going to happen is it's going to be a browser war, not necessarily an AI chatbot war. So pretty, pretty big news there from OpenAI. And speaking of Perplexity also, well, they've already launched it. We got reporting on OpenAI's AgentIC browser, but Perplexity has launched Comet, their agentic browser. Funny enough, I actually have it running literally right now as I'm doing this live stream and I'm getting ready to send another command. So for me, the past like week or two it's really been this big shift toward the agentic browser and Perplexity kind of leading the way here. You know, we, if you remember back to our 2025 AI predictions and roadmap series, I said at the time, this is back in January, I said Perplexity, if they want to survive and thrive, they're going to have to hard pivot. And we've seen that because everyone else kind competed on their grounds of being the answers engine with chat, GPT's deep research and Gemini Google, Gemini's Deep Research. Pretty impressive pivot so far from Perplexity. I'm actually personally glad they shied away from hardware, which is where they were reportedly going to focus on. And now instead they have a pretty good iOS app for the iPhone that has better AI capabilities than Siri. Not a diverse set yet, but better. And now we have this agentic browser again based on Google's Chrome. So Chromium. So Perplexity's Comet is the all new AI powered browser designed to change how people find and use information online. So it replaces cluttered tabs with endless links with a single conversational interface where users interact directly with an AI assistant for both browsing and organizational tasks. So instead of typing formal searches using Boolean operators, right, which is what I do all the time, users can simply ask comments, questions or explain tasks in natural language. And the AI handles the workflow and keeps the context throughout the session. So right now it blends together different models, Right. So it uses OpenAI's GPT models and Anthropic Squad models, aiming for a more robust and versatile answer than traditional search engines. So right now, early access to Comets has already been rolling out to Perplexity's new Max subscribers who pay $200 a month. But I'm on the pro plan, the $20 a month and I did just get access to it a couple of days ago. So if you do have a paid account, check your email. It might have got buried, that's all. You just get one email. If you log in, you're not going to see it because you obviously are taken to a download page and then you can launch the browser. All right, so let's go. Next. Next AI new store. Yeah, starting off with the browsers, y'. All. Next Crack 4 released benchmarks. Impressive reception so far, not very good. So Elon Musk's AI company Xai has released Grok 4, its latest flagship AI model, and introduced a new. Yeah, now they're just trying to beat everyone and everything because now they have the most expensive plan per month at $300 called Super Grok Heavy. We're getting weird with the names. So Grok 4. So, yeah, there's two different variations. There's the normal paid plan and then the ultra. You know, not to use Google's terminology, but the Ultra Premium plan with Super Grok Heavy at $300 a month versus the Pro plan for Grok 4 at $30 a month. So Xai claims that Grok 4 scored a 25% on the challenging Humanities last exam, which outperforms just about everyone by a pretty big margin, including Google Gemini's 2.5 Pro and OpenAI's 03. So with tools enabled, Grok 4 Heavy reached 44% on that test, nearly doubling Gemini's 26.9%. So the $300 super Grock Heavy. Super Grock Heavy just sounds weird. It gives users early access to Grok4 heavy and upcoming features, making it the most expensive AI subscription among major AI providers. So SAI plans to release an AI coding model in August, a multimodal agent in September, and a video generation model in October with Grok4 available via the API for developers. However, it doesn't show its thinking or reasoning, at least right now, in the responses. So X AI leaders focused on Grok4's technical advancements during the launch, but avoided discussing the recent content moderation incident. Yeah, let's discuss it because that is our next news story. Yeah, a lot of bad things have happened since Xai has released Grok4. So yes, you can use Grok4 like you can chatgpt, but also there is a Grok4 response bot. So that response bot that goes on Twitter and you can ention it and ask it anything and it'll give you a response has gotten into a lot of trouble because of some of the responses. But that might not be the worst news that XAI received received this past week. That's because it's been discovered that Grok4 aligns its answers with Elon Musk's views on controversial topics, according to TechCrunch and user reports. So users posted screenshots showing Grok4 referencing Musk's recent post on X, formerly Twitter, to shape its responses on sensitive issues like the Israel, Palestine conflict, US Immigration, and abortion. So the AI explicitly stated in its reasoning process that alignment with Elon Musk's view is considered and often first searched Musk's opinion before answering any contentious questions from users. So as an example, when asked about US immigration, Grok 4 included a section on Musk's stance highlighting his support for quote unquote reformed selective legal immigration on less controversial topics, though Grok4 did not reference Musk's views, suggesting this alignment is specific and intentional to be deployed on divisive subjects. So Elon Musk and Xai, like we said, launched Grok 4 in the live stream a couple of days ago and it's been kind of a weird rollout ever since yet. Because not only that, but the Grok chatbot was going off and saying some very inappropriate things that I don't even feel I can be repeating here on the show. But we've been sharing about them in the newsletter, if you want to know. Not good. Not good rollout reception so far for Grok4, as well as some third party benchmarks have shown that Grok4 is actually trailing behind even companies, not state of the art models. Right? So in certain benchmarks Grok4 is behind like GPT4.1, which is not necessarily a, you know, state of, you know, it's not Even one of OpenAI's top three model tiers technically. So yeah, we'll see. As more and more, you know, now that the API is out, we will be getting more and more third party benchmarks to see how Grok 4 actually stacks up. But you know what I'd love to hear from our live stream audience and our podcast audience. You know, let me know. Always put the information to sign up for the newsletter in my LinkedIn in the show notes. Do you want to hear more on grok? I don't know. I'm. I'm iffy on it. Right. It's controversial. I think a lot of people, if you're really, really into AI, you care a lot about it, but I don't know if the majority of our audience does, so let me know if you do or not. I personally would never advise a company to use Grok. Right. And that's something I think that SAI probably knows, and they're trying to invest more resources to win enterprise business. But when you have these snafus like we've had multiple of them with Grok's new chatbot, it's, I'd say, hard for a company to consider this. Right? I mean, the fact that it is by default that GROK is intentionally looking up Elon Musk's point of view to answer contentious questions. It's extremely concerning. That's extremely concerning. Yes. You know, it's been shown that just about every large language model has some sort of bias, Right? But generally that just might reflect humanity's bias towards something. Right? Because more or less large language models are a reflection of society, and it's extremely dangerous for a large language model to reflect a single individual. So will biases and stereotypes ever be out of large language models? Probably not. Is it extremely concerning that it's reflecting one single individual's thoughts? Yes, it is. And that's not a good look. So hopefully the XAI team cleans that up. All right, our next piece of AI news. Huge partnership. So the American Federation of Teachers AFT is the second largest teachers union in the US Will open a national Academy for AI instruction this fall at its New York City headquarters. So this is backed by the major players in the game. $23 million in funding between Microso, OpenAI and Anthropic. So the academy aims to train 400,000 educators in the U.S. that's about 1 in 10 U.S. teachers on using AI tools safely and ethically by 2030. So Microsoft is contributing 12.5 million over five years. OpenAI is providing 8 million in funding and 2 million in technical resources. And Anthropic is adding $500,000 for the first year. So the initiative follows similar high tech training models from other unions such as the United Brotherhood of Carpenters, which have partnered with industry for workforce upskilling. So the academy will offer hands on workshops for teachers, focus on practical classroom uses of AI, including lesson planning and administrative tasks. So the move comes as schools and universities nationwide increase AI adoption after burying their head in the sand for two years and screwing college graduates over. But now you at least have some institutions being like, yeah, maybe we should pay attention to this AI thing. As an example, California State University recently made ChatGPT available to more than 450,000 students. So that's the different state universities in California and Miami Dade county public schools began rolling out Google's Gemini AI to more than 100,000 high schoolers. So according to the New York Times About 200 New York City teachers previewed the training, experimenting with Microsoft Copilot and Con Amigo, an AI tool for schools. But some experts warn that tech companies could use these educational partnerships to build brand loyalty among students, potentially turning them into long term customers. I would assume that's what's going on, right? It's the other side of philanthropy, right? Yeah. Of course the big players are going to want to get their large language models into as many students and teachers hands as possible, Right. I don't think there's anything bad necessarily with that, especially when there's multiple players involved. This was just one player involved. I'd be like, all right, that's probably not a good look, right? If the future educators are only using one model, probably not good, but it makes sense, right. For these big tech AI companies to be investing money into this. Not just to maybe try to build a little bit of brand loyalty and familiarity, but to also help. Right. Students are coming out of specifically higher education universities without any job skills, without the basic requirements for jobs. Right. Which is incredibly bad. So actually had a recent episode on that, if you want to go back and listen, episode 562, how will new grads get a job with AI? So yeah, if you care about that, make sure to go check that out. All right. Another piece of interesting. I'll say interesting, to say the least. News. Microsoft has announced a $4 billion global AI training initiative called Microsoft Elevate, aiming to train 20 million people in AI skills over the next five years. But the announcement comes at the same time. Interesting timing that Microsoft has laid off more than 15,000 employees worldwide since May, including more than 3,000 in Washington state, highlighting the tension between investing in new technology and training and job cuts. So the Elevate program will include a new Elevate Academy and partnerships with schools, colleges and nonprofits, with Washington State receiving the highest per capita funding in at least 1 million indirect grants. Microsoft president Brad Smith stated the goal is to democratize access to AI, emphasizing the importance of teaching not just how to use it, but also when to use it and when to stop App. The company is focusing on building trust in AI by providing education and integrating privacy and security features directly into its products. So Smith compared AI guardrails to car safety features, noting that while technology can provide tools, individuals must also take responsibility for safe and ethical use. Also, he did acknowledge the human cost of layoffs, stating that while decisions are difficult to lay off employees, they are made to enable future growth and that Microsoft strives to offer strong severance Packages support for affected employees. All right, Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on Genai. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you can partner with us to do. Just like some of the biggest companies in the world do. Go to your everydayai.com partner to get in contact with our team. Or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on gen. We have a new number, y'. All. Well, not a new number, but a company reaching new heights because Nvidia became the first public company ever to close a stock market day with a $4 trillion valuation. So Nvidia became the first company to close a trading day with a $4 trillion market cap. So shares are up 21% for the year and over the past 12 months, despite earlier concerns about competition from Chinese AI firms and shifting industry trends. So US President Donald Trump used Nvidia's milestone to call for lower interest rates, pointing to the company's success as evidence the Fed should act. Fears, though, that AI companies would move away from Nvidia's top tier chips have not materialized, obviously, and demand for their powerful GPU chips for both AI training and deployment remains strong. So yeah, if you don't follow the whole chip game, essentially, probably any AI that you use, any AI tool is currently powered or was likely recently powered by Nvidia's GPUs. So despite US bans on chip sales to China costing the company billions, Nvidia's stock keeps climbing, signaling that high performance AI power hardware is still critical for tech progress and business growth. Who would have thought Nvidia would ever become a $4 trillion market cap company? Me, as I sip on my coffee. Oh wait, I told you like two and a half years ago, didn't I? I did. Back when hardly anyone had heard of Nvidia unless you were a Big AI dork. I said Nvidia at the time, which they were not Even a top 10 company in the US at the time. I said that their most important company probably in the history of the world and you'll see, here they are because they just achieved the highest market cap of a company ever. So if you want to talk about not just the future of work, right, the future of medicines, the future of scientific discovery, the future of, I mean technically humanity, right in our day to day interactions with the physical world, they're all being driven by Nvidia's chips. So even if you don't think that they're a big deal, still, they are. All right, we have two more quick AI news stories. So first, Google has hired Windsurf CEO Varun Mahan and co founder Douglas Chen for its Google DeepMind division, paying $2.4 billion for non exclusive licensing rights to Windserve's AI coding technology. So this move follows the collapse of OpenAI's reported 3 billion DOL dollar acquisition attempt, which fell apart due, according to reports, due to Microsoft's refusal to limit its access to Windsurf's intellectual property. So Google's deal allows Windsurf, the company, to keep operating independently under interim CEO Jeff Wang, with no equity stake taken by Google. So essentially OpenAI was trying to acquire, acquire Windsurf, the AI coding platform. I'd say it's probably one of the top, top five, right along with GitHub, copilot, cursor. I mean there's tons of them out there now. It's like every day there's a new, you know, I hate to say like vibe coding platforms because there's so much more than that. But Windsurf has definitely been one of the leaders and OpenAI was reportedly in the almost final stages of a $3 billion acquisition of the entire company which fell apart because reportedly it was something about passing the IP rights or withholding them from Microsoft, which Microsoft part of their long term agreement reportedly with OpenAI was access to all of OpenAI's IP. So that includes when a company gets acquired by OpenAI, which apparently was one of the reasons according to reports that those deals fell apart and Google very quickly be swept in. So they did not though interestingly enough they did not acquire the company. This was more of an aqua hire. So they get some exclusive licensing opportunities, but they will probably not fall under jurisdiction by the Department of Justice or other federal regulators here in the US for anti competitive tactics, which all the big tech companies have fallen under recently. So this follows a very similar path that happened with Microsoft as they aqua hired inflection AI. I think that was last year. So yeah, we'll see. I don't think there will be anything, any regulatory concerns with this one, but I mean, wow. I mean that's a pretty big move here from Google and the DeepMind team to add to its roster as the tech poaching continues. And speaking of, of tech poaching, that is our last AI news story that matters because yes, we've talked about this now for three or four straight weeks because it keeps getting nuttier than a squirrel on keto. Because Meta is reportedly offering more than $200 million in compensation to Rumang Pang, Apple's top AI executive overseeing artificial intelligence models. At Apple's. At as at Apple as part of its push to build an artificial general intelligence division called Meta Superintelligence Labs. So the reported pay package, which I believe was first reported by Bloomberg, is the largest in tech history for a non CEO according to early reports. So we'll see if that is ultimately the truth. But this is wild built $200 million package. So the reported pay package is among the largest in the industry. And right now it's reported to say that Apple did not attempt to even match Meta's offer, signaling a possible shift in how aggressively each company is willing to compete for AI talent. And right now apparently no one can compete with Meta because their pay packages, it's like monopoly money. No, it's, I mean there's not even that much monopoly money. So these Meta pay packages are often structured though with performance based incentives. Incentives and require several years of service, meaning the full amount may not be realized until certain conditions are met. So this move highlights Meta's commitment to advancing their AGI and ASI kind of investments. So that could one day match or surpass human intelligence. So according to reports, Meta shares have risen nearly 25% this year while Apple's stock has dropped more than 15%, reflecting investor concerns over Apple's slower progress in AI. Yeah, you could say it's slower progress in AI, but I would say they don't really have any progress in AI, aside from probably hiring a lot of lawyers to handle all of the class action lawsuits against Apple for not delivering on basic AI that it marketed, that it had and it actually didn't. Oh, that's not a burn, that's just facts. All right, let's go over our new section and I think you guys like this, right? The, the what's new and what's next? The Little bullet points. So maybe things here, smaller things that maybe didn't make our top AI news stories or what's around the corner. So first Grok is expected to launch connectors for Google products, Slack and Notion. Yeah, everyone's doing the connectors and integrations. I feel really bad for companies that invested millions of dollars three years ago to build their own kind of bespoke rag system. So Claude is rolling out a limited preview of thinking to free users. So before Claude, users couldn't get that thinking. Now they can. Google deepthink is maybe dropping in the coming weeks. So essentially a way that you can force on the front end of Google Gemini to think more, to have Google Gemini think more. An MCP directory may be coming soon to Claude, making it easier to find more trusted. MCPs are the model Context protocol, which allows Claude and other AI chatbots, any company that has adopted the MCP protocol to essentially talk to each other. Google has released photo to video for its powerful VO3 video model inside of Google Gemini. Another Gemini one that I'm super excited about. Google's Canvas mode, which is one of the most powerful AI tools out there and it doesn't get any love now works with their gems, which is their kind of version of GPTs or a custom smaller version of Google Gemini. Speaking of Gemini Gemini 3, we saw a Gemini 3.0 reference spotted in Google's CLI code. So maybe we'll be seeing another Google model sometime soon. ChatGPT may get the ability to upload videos. According to a leak on Twitter, ChatGPT's study together mode was spotted in some accounts, but then it got taken away. So maybe that has something to do with ChatGPT's recent educational partnerships, but essentially allowing collaborative studying or collaborative use of the ChatGPT models. AWS is launching an ancient marketplace reportedly on tomorrow. On Tuesday, Grok may be rolling out to Teslas in the coming weeks, according to elon Musk. And OpenAI has delayed its open weights model. So yeah, there's been a lot of talk on OpenAI saying, hey, we're going to release this open model, but its CEO Sam Altman essentially said, hey, it's not ready, we need to do more so safety testing, so we'll see when it's released. But I'll tell you this, that is going to set off a lot of things. The combination of when OpenAI does release their open model, which was reported to be more of an O3 mini type of model, and whatever meta is cooking up next, presumably also open, unless they decide to go closed source, which would be a surprise to everyone. The future of AI development is extremely bright right when you look at what Meta has assembled with their msl, their Meta superintelligence team, and also when and if Google or sorry, when and if OpenAI does release an O3 mini level open weights model, that's going to shake up the entire industry but also the business world. All right, that is a wrap. Let's very quickly go over the main news stories. So OpenAI is set to launch an AI powered web browser according to to reports Perplexity already has and just started rolling it out to normal paid users with their Comet Browser. Xai launched Grok 4 and a very expensive $300 Super Grok Heavy plan which also has a multi agentic mode. Grok 4 also got in a ton of hot water both from its responses and as well as aligning its answers intentionally with Elon Musk Major Teachers Union. The EFT has partnered with Microsoft, OpenAI and Anthropic to train as many as 400,000 educators on using large language models. Microsoft has launched a $4 billion AI training initiative called Microsoft Elevates, even though it has laid off recently about 15,000 employees. Nvidia became the first public company ever to close a trading day at $4 trillion. Continuing their AI market dominance, Google Aqua hired some of Windsurf's AI leaders and str a deal for some exclusive rights to their technology after the Open AI acquisition of Windsurf reportedly collapsed. And Meta has a $200 million man on staff after reportedly they lured away Apple's top AI talent with a total package worth $200 million. That is wild. All right. What's wild is the rest of shows for this week, so maybe you heard something there and you're like, what's that going to look like? We're probably going to be doing a hot take Tuesday tomorrow, I believe, probably on Agentic AI and Agentic browsers. That's what I put it out in the newsletter. That's what you all wanted to hear. And we'll probably do Putting AI to Work Wednesday something for Agentic browsers as well. So we're going to get super Agentic this week, so make sure to join us. And if you haven't already, please go to your everydayai.com Sign up for the free daily newsletter. So thank you for tuning in. We'll see you back tomorrow and every day for more Everyday AI. Thank you y'. All. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit your everydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.