Ep 719: Google Gemini 3.1 tops charts, Claude Sonnet 4.6 impresses, New OpenAI leaks reveal their massive AI hardware plans and more

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.

This week was a busy one in AI news. I mean, not only did we get two huge releases from Google and Anthropic, but that was just the tip of the AI iceberg. I mean, we got word on what OpenAI is going to do on the hardware end and I think one of the things that they're working on is going to surprise a lot of people. We got a handful of pretty big non LLM releases that we'll talk about, like Google's Pimelli photo shoot. But what most people are actually talking about, the fact that two AI leaders did not join hands at a global AI summit. So don't worry, the. The imaginary or maybe real AI beef is still alive and well. All right, there was a lot to cover this week. If you miss anything, don't worry, we're going to get into it. Let's go. Welcome to Everyday AI. My name is Jordan Wilson and if you're new here, well, this thing's for you. Everyday AI is a daily live stream, podcast and free daily newsletter helping everyday business leaders like you and me not just keep up with what's happening in the world of AI, but how to make sense of it. Cut through the bs. We tell you, here's what matters and that's exactly what we do on our Monday show each week. The AI news that matters. So if you are new here, yeah, we do this thing Monday through Friday, but for the most part on Monday we focus on the AI news, telling you what happened and what to expect. But we do this thing Tuesday through Friday as well. So if you haven't already, please make sure to go to your everyday AI.com Sign up for the free daily newsletter where we're going to be recapping not just what we cover on today's show, but all the other daily AI news that matters. And while you're on our website, make sure if you haven't already, two numbers. I keep saying you got to know them. 7 12, 7 13. Those are the two episodes that you are going to want to listen to if you haven't already. That is our 2026 AI prediction in roadmap series. So make sure you go check those out. All right, let's get into the AI news that matters. I'm excited for it. Live stream audience, good to see you. Looks like the YouTube is finally working. Big bogey face. Good to see you. Joe joining us from Fort Lauderdale, Brian joining us from Minnesota, Jose from Santiago. Everyone else, thank you for joining us. Yeah, it's unedited, unscripted. We do this thing. All right, first piece of news. Could Anthropic surpass OpenAI when it comes to revenue? Well, according to at least one recent report, it could happen. So according to the information in EPIC AI, Anthropic's annualized revenue has been growing at an impressive rate of 10 times per year since it hit $1 billion, significantly faster than OpenAI's rate of 3.4 times per year. 8 and reports show that if these growth trends continue, Anthropic could actually surpass OpenAI in annualized revenue by the middle or end of this year. However, recent data has shown that that 10x growth rate from Anthropic has slowed and leveled off a little bit to 7x since July, indicating the initial rapid expansion that Anthropic was experiencing kind of from 2023 to 2024 to 2024, 2025 might be slowing down just a little bit. OpenAI's growth rate is also expected to slow, with projections for 2026 showing a 2.2 times increase in revenue and anthropic forecasting growth for 4 times or less for the rest of this year. So despite these slower projected growth rates, according to reports, analysts still expect a crossover point at some point when Anthropic's annualized revenue might exceed OpenAI's. Like I said, might happen in late 2026 or early 2027, according to reports. So this. Com as recent News showed that OpenAI is now telling investors it is targeting roughly $600 billion in total compute spending by 2030, a narrower and lower figure than the 1.4 trillion infrastructure commitment previously cited by CEO Sam Al. So OpenAI projects more than $280 billion in total revenue by 2030, with consumer and enterprise businesses each contributing roughly half, sealing an effort to more tightly align spending plans with expected commercial returns. OpenAI reported $13 billion in revenue for 2025, but burned about $8 billion last year, according to reports, and improved on its own internal targets. So I'll say this, there's going to be a lot of talk about revenue. Some reports are going to be accurate, some aren't. I don't know about this one, but if you haven't seen the headlines, you know, expect this to spill over to social media soon. And you know people are going to be talking about, oh my gosh, Anthropic is, is it's Taking over, right? One thing to realize, especially if you listen to this show every day, if you work in and around AI every day, especially if you are terminally online like a lot of people are, we live in a bubble, right? When I see reports like this, I'm like, probably not, right? If I'm being honest. I'll maybe do a dedicated show on this at some point, right? If you want to hear a, you know, who's winning the AI race, like for real, for real, go ahead, drop me, drop me a comment. You know, I don't know, say, say race. I don't know if you guys want me to talk about those kind of things, but you know, if you're listening on the podcast or you know, know, go, go leave a Spotify comment, or if you're on the live stream, just say race. I'll see if anyone actually wants to hear me blab on about this. But a couple important things to note here. If you are just in this AI bubble, people thinks, think it's a very tight race. It's absolutely not, right? You say, oh, like you say AI, everyone thinks Chad, gbt. No, like, for the most part, no one has heard of Anthropic, right? A lot of studies out there, you know, you look at and you're like, oh my gosh, this says, you know, Anthropic has taken over the enterprise. Oh, guess what? That's. That study was done by Anthropic's largest investor. So you really have to understand what's real and what's fake. Is Anthropic crushing it on the revenue side? I think. Absolutely. I think, and I've said this since probably 2024, I think anthropic has prioritized revenue per user where OpenAI has could care less, right? At least according to what they're doing and the moves they're making. I think all OpenAI cares about is number one, owning the enterprise and number two, owning the consumer market. And they're at least undoubtedly doing one of two of those things, maybe both of them, right? Because they're going to be hitting a billion weekly active users pretty soon. And on the ap, right, so much of Anthropic's revenue comes from the API side and those things are just a, a one click away sometimes from potentially losing billions of dollars. And like I talked about it on the show last week that no one else is talking about. I don't know why Anthropic revenue comes from the API side, right? They don't have millions of business customers or hundreds of millions of users like OpenAI does. Like, I think that Google will be catching up pretty soon. Anthropic has dominated from developers, software engineers on the API side. Their API usage is slipping according to at least openrouter. So for the last time last week, I don't know, maybe I'm the only one talking about this because I'm the only one dorky enough to check. It's the first time in a year that they haven't had a top three model on Open Router, which tracks API usage at least through Open Router, which is the most popular one in the world. So I don't know, could it be this, you know, the, the Claude bot open claw fiasco that caused them to tank? Maybe the super bowl ad left a bad taste in people's mouth, I don't know. But at least when I see reports like this. Could it be true? Maybe. Doesn't matter. I don't think so in the long run because I do think at any point OpenAI can, you know, turn on five new streams of billion dollar business revenue. And I don't think Anthropic can. All right, speaking of this beef that's going on. All right, a couple people want race. Marie said it, George said it. All right, we'll see. Speaking of the beef. Beef. Well, the reported or alleged or real beef, depending on how you look at it, between OpenAI and Anthropic was on full display for the global scene to see. Yeah. So live stream audience, you see it with a picture on the screen. I'm sure if you're listening on the podcast, you've seen this by now. Right. But there's a little bit of a tension between OpenAI and Anthropic that was on stage at the AI Impact Summit in India. So at the India AI Impact Summit, OpenAI CEO Sam Altman and Anthropic CEO Dario Mata shared a kind of awkward on stage moment where they kind of refused to join hands. Right at the end, you know, all the big, literally all the big AI CEOs went on stage with Indian Prime Minister Narendra Modi. Everyone raised their hand, right. It was this, you know, Kumbaya moment. And there you have CEO Sam Altman and Dario standing next to each other and they didn't link hands. Right. So there's, I don't know, about a dozen or so, you know, AI leaders and dignitaries on stage. Everyone joins hands, you know, and Sam and Dario don't. So, Modi. Yeah, there's, there's been so much talk about this I almost say, like, why talk about this? But I think I have to because technically it was probably one of the most talked about stories in AI. So here's what happened. So Modi lifted Altman's hand for the photo, but Altman and Amodi kind of just instead raised their fist. Right? It looked like Sam looked over at Dario, kind of lifted his hand. Right. I watched the video replay a couple of times. So probably intentional. Right. But, you know, a lot of people are saying, oh, it's Sam's fault. Oh, it's Dario's fault. I don't know. But the tension follows Anthropic's recent Super bowl ad campaign criticizing OpenAI's decision to introduce ads in ChatGPT, where they use words like betrayal and deception to highlight concerns about ads interrupting AI interactions. That ad was obviously not well received. A couple of, you know, industry benchmarks show it was one of the least liked ads in the last five years. And anthropic, I believe, only had like a 7% brand recognition. So most people didn't like the ad and they didn't even know who Anthropic was yet. It is Anthropic seemingly changing their tone and their messaging and approach and kind of going on their bully pulpit, pulpit a little bit. So Altman publicly at the point, called Anthropic's ads clearly dishonest and accused the company of doublespeak, defending OpenAI's approach to making AI accessible to billions, including those who cannot pay subscriptions. So for the actual AI Summit, Impact Summit itself, that was held this past week in New Delhi, India, and it concluded with the adoption of the new a declaration on. In. On AI impact. So that's a non binding agreement endorsed by 89 countries and international organizations, including the U.S. china and the EU. And it focuses on AI for all, seeking to bridge the digital divide. Right. Another thing that grabbed a lot of headlines, you know, a lot of people are saying that, you know, Sam Altman did a pretty good job kind of, you know, talking and communicating that message on AI for all. And a lot of people were kind of roasting Dario for he went up and gave a kind of a more robotic speech, reading off an iPhone the whole time. So, you know, I guess if nothing else, nothing really happened at this global summit. Everyone thought something would, right, oh, we're going to see some new announcements or breakthroughs in technology. No, there was a non binding agreement that company signed that doesn't really mean anything. And then an awkward photo op where everyone on the stage was holding hands except the CEOs of OpenAI and IT. So there's your, there's your daily dose of the, the AI drama. Yeah, big bogey face says it was super awkward. I agree at least though I did like, I think it was the, the chat GBT developer account on Twitter posted a picture instead where they put claws on Sam Altman. So OpenAI was at least, you know, having fun with it and you know, referencing their recent acquisition of OpenClaw. So that was fun all right. But I mean, poor Google. All this stuff and Google released the world's most powerful model and no one seemingly batted an eye, which is crazy to me. And you know, Speaking of those two episodes, I tell you to keep listening to 7:12 and 7:13. I said this in one of my predictions. I said at any point Google can release the world's most powerful model. And since I did that show, they've already done it twice with their Gemini 3 Pro Deepthink and then their new Gemini 3.1 Pro. So let's talk about the now world's most powerful model that even though it grabbed a few headlines, it doesn't seem like people are talking about it. So Google released an update to its marquee model with the new Gemini 3.1 Pro, introducing a three tier adjustable reasoning system that allows the model to scale its thinking effort from quick responses to deep multi minute analysis. So this new feature makes Gemini 3.1 pro act more like a lightweight version of Google's specialized deep think reasoning system. So here's some details on what's new from Google's new model. So Gemini 3.1 Pro has those three thinking levels. There's low, medium and high. So if you're on the developer end or using it inside AI studio, you'll see that. And that enables developers to balance speed and depth of reasoning depending on the complexity of the tax. The high setting now kind of mimics the capabilities of Google's DeepThink. I'd say it's like a DeepThink Mini, which I know might be kind of confusing, right? If you think of like the most powerful version of each company's models that's available on front end AI chatbots, right? So let's just use the big three anthropic Claude 4.6 opus Gemini 5.2 Pro and let's just say Google Gemini 3 before this, right? If you use the Pro or the most powerful versions for front end consumers, I'll say this, the GPT 5.2 Pro can work for more than an hour and do very impressive work. Right, Claude? Four, six, opus. You know, 10, 15 minutes. You know, if you give it a pretty hard task before Gemini 3 Pro a minute, right. So now with this 3.1 Pro and I'm not saying that thinking time always corresponds with better outputs, I will say a lot of times it does. But even for myself, sometimes when I'm, when I was using Gemini 3 Pro and I gave it a very complex task and I'm constantly, you know, split testing things across, you know, three or four providers at once. Even just for the fact of not needing the outputs but understanding how each of the highest reasoning levels work, I was always like wait, how did Gemini 3 get this done so quickly? Right. And sometimes the, the quality of the response is more nuanced, right. It's not always finite. Oh yes, this model won. But I do and I have personally enjoyed, you know, the new Gemini 3.1 update with a little bit more of that thinking. But from the benchmark side, yeah, it's absolutely just absolute dominance once again by Google DeepMind. So benchmarks results so show dramatic improvements for the new 3.1 variation where they scored a 77.1 on ARC AGI2. So yeah, if you follow all the, the dorky benchmarks, you know, I don't know, the best models in the world like a year or so ago were in the sub 20s on ARC AGI2. So now Gemini3.1 Pro scored 77.1 which is more than double of Gemini3 Pro's 31%. And it does also outp outperform competitors including the two marquee models from OpenAI and Anthropic. On academic and scientific reasoning benchmarks like Humanity's last exam and GPQA diamond. Gemini 3.1Pro also leads with scores of 44% and 94% respectively. On the agentic side, huge improvements as well on those multi step workflows which is what everyone's paying attention to that and tool use as Gemini 3. One pro scored a 68.5 score on terminal coding and 85.9 on web search capabilities. So pretty interesting though here Google's shifting over to a 0.1 version which they haven't really done often. Right. So usually Google's been, you know, going to, you know, Gemini 2, Gemini 2 5, Gemini 3, you know, so going with an incremental update. But I think that's where we are right now in the state of AI where companies can and probably will be releasing these more incremental versions. Right. So we saw it with GPT5.2 and with their GPT5.3 codecs. We saw it with Claude's 46 models and now we're seeing it with Google. So I do believe that's the first time we've seen this. So the new update is currently available in preview across multiple Google platforms, including the Gemini API, AI Studio, Vertex AI and Android Studio. Foreign. Moves too fast to follow, but you're expected to keep up. Otherwise your career or company might lag behind while AI native competitors leap ahead. But you don't have 10 hours a day to understand it all. That's what I do for you. But after 700 plus episodes of everyday AI, the most common questions I get is where do I start? That's why we created the Start Here series, an ongoing podcast series of more than a dozen episodes you can listen to in order. It covers the AI basics for beginners and sharpens the skills of AI champions pushing their companies forward. In the ongoing series, we explain complex trends in simple language that you can turn into action. There's three ways to jump in. Number one, go scroll back to the first one in episode 691. Number two, tap the link in your show notes at any time for the Start Here series. Or you can just go to start here series.com, which also gives you free access to our inner circle community where you can connect with other business leaders doing the same. The Start Here series will slow down the pace of AI so you can get ahead. Yeah, comment here saying Google for the login. Yeah, that's what I've been saying for a long time. Google has more access to cash websites or websites that they're able to crawl. They have the advantage for YouTube data. Right. And just for search and search intent. And so yeah, I've always said that Google is playing the long game and they've been playing the long game for the past, I'll say six months. But at any time they can drop in and win the short game. Right? So if we see a GPT5.3 from OpenAI, because right now we just have a GPT5.3 Codex, right? So if OpenAI comes in with a world shattering GPT5.3 for chat GBT, which I think is very possible and could in theory, you know, happen any week now, I think that Google, if they wanted to, could come out with a Gemini 32 Pro, that my topic on top it on most benchmarks. Can't speak today. That's fun. All right, speaking of new model releases, again, poor, poor anthropic. This is the second time in just as many weeks I believe that they've come out with a new model and instantly got swiped up by their competitors. Because, you know, last I think it was last week or the week before, they came out with their premiere Opus 4.6 update. And then literally I think an hour later, OpenAI came out with GPT5.3 codecs, which wiped the floor with Opus 4.6. And then here Anthropic released Sonnet 4.6. And you know, for two days it looked really good on the benchmarks. And then Google came out with Gemini 31 Pro. Anyways, Anthropic has introduced Claude Sonnet 4.6, its new default AI model that sits in the middle. It is both faster and less expensive than their Opus4.6, but it also improves coding capabilities in computer use, skills like clicking and navigating software interfaces. So this new upgrade narrows the gap between Anthropic's premium and mainstream AI models, making advanced features accessible even to free users, which could broaden the company's base. So yeah, if you didn't know, right now Anthropic kind of has three different tier. So the tiers are Haiku, which is their smartest and cheapest model. Right. And not as smart. And then you have Opus, which is the big one. That is their, you know, much slower, much more expensive if you're using it on the API side, that is their, you know, powerhouse model. And then in the middle sits Sonnet. But what is interesting here with the Sonnet 4.6 upgrade, well, it did better on some things than its bigger brother opus 4.6, such as certain coding benchmarks in computer use. So yeah, Claude Sonnet4.6 outperforms the recently released Opus4.6 model on certain real world office tasks, Meaning users might need to pay for the might not need to pay for the higher priced Opus tier. Yeah, I've seen a lot of people that were still using anthropics models for OpenClaw as an example, although I think people are kind of moving away from Anthropic models for that in mass. But I did see a lot of people kind of move away from Ops 4.6 for certain tasks and just move to Sonnet4.6. So the new Sonnet 4.6 is better at handling long chunks of code by understanding context before editing, avoiding redundant logic, and delivering quicker, smarter answers. So a couple things on the performance side, sonnet 4. 6 matches the performance of the previous opus 4. 5. So across all benchmarks. I know that's confusing. So Opus 4.6 still leads Sonnet4.6 on many of the major benchmarks. Although Sonnet 4.6 now eclipses the previous Opus 4.5. So if that kind of levels it out in your head. Also the extended context window for developers, it does a 1 million token context window in beta, which doubles the previous maximum. All right, Brian here says Sonnet 46 does really well on most tasks I've tried. Yeah, if I'm being honest, I haven't done as much testing on Sonnet 46 as I would have liked to. Right. The release cycle happening now, right? It's kind of conference season. You know, we, we had the Nvidia conference coming up. We, you know, Google's conference coming up. So, you know, kind of in this, you know, first quarter, the releases are just so fast, right. Because companies are gearing up for their kind of major announcements, you know, usually in the March, April, May. So it's, it's pretty, pretty tough even for someone that just does this. This is all I do. It's tough to do all this. Right. All right, here is what I think is low key, maybe the biggest story of the week because yeah, Google released Gemini 3.1 Pro, which like I said, is now the world's most powerful model. But that wasn't even the thing that got the most traction from Google this past week. It was actually not even a new mode or a new app. Right. It was an update to a piece of Google AI software that I think a lot of people have never heard of. But I think it got a lot of love for the right reasons and it's really, really good. So Google Labs has introduced a new feature called Photo Shoot within its free marketing tool Pumeli, aimed at helping small to medium sized businesses generate professional product images easily using AI. So Photoshoot uses Google's generative AI, including Nano Banana, to transform any uploaded product photo, regardless of the quality, into polished market ready images. So users can select from various shot templates such as studio, floating, ingredients in use, et cetera. The last of which includes AI generated models to show products in real life scenarios. So the tool also offers lifestyle image features, you know, that features the product in typical use cases, replacing the need for some, maybe to hire professional photographers in some marketing tasks. So businesses can upload a photo or simply provide a URL of their product online, making the service flexible and accessible without requiring photography skills or equipment. And then after selecting the preferred shots, Photoshoot generates new images that users can download for free to use in their marketing campaigns. Aiming to reduce the time and cost burden of digital marketing for smaller companies, Google Emphasizes that Pumeli and Photoshoot are designed to help smaller businesses keep up with the demands of online marketing by producing authentic on brand visuals without needing to hire professionals to do it. So currently Photoshoot is available in the United States and Canada, Australia and New Zealand. Expanding Google's suite of AI tools that impact creative industries. Let me know, live stream audience, have you all used this yet? I actually have a lot of thoughts on this and I'll tell you why. And again, maybe this is one of those shows that I should do later. But the interesting thing here, before I give you my personal take on this, this thing went absolutely viral and that's why a product update news, right, made it into the top news story. I believe it had more than 23 million views on Twitter, right? Whereas their Gemini 3.1 Pro announcement had like a small fraction of that. Here's, here's my take. Okay. I think that. And this is a maybe, right? Maybe if this thing takes off, if Pumelli photo shoot takes off, I think this could be looking in five to 10 years down the line, this could be one of those instances where people look back at and they say, oh, this is when a lot of roles changed, right? I'm not saying this one, you know, Pumelli photo shoot, right? But I think this could be one in a series of many where it made data. Is it at least on the creative side, day to day creative professionals change their workflow. And I think this is one of those things, right? I've talked about for years this upcoming shift from operator to orchestrator, right? Go listen to the 2026 AI prediction and roadmap series for more on that. Right? But essentially humans, even over the last four years with very powerful AI, we've still been the operators, right? Pushing the buttons. This might be one of those. Yeah, we're still pushing a button on P Photoshoot, but this might be one of those that turns a lot of creatives into orchestrators, right? Orchestrating AI to get the job done. So, you know, in the same way that, you know, Photoshop was seemingly at the time very disruptive to the photo industry, I think Pumeli Photoshoot could do that, right? I think it's going to be a very polarizing feature. You have to also say that this technology and these capabilities have been out for at least a year plus in many other popular, at least popular in AI circles, you know, AI apps. But this is the first time we've seen from a big company, and by big company, I mean Microsoft, Google, Anthropic OpenAI that we've seen a tool like this that brings drag and drop creative powers to the masses. Right. So obviously I'm talking about more than an AI image generator. Again, that's still, I think, orchestrating, right? You know, uploading multiple, you know, photos into something like GPT image 1.5 or nano banana Pro. Right. And then, you know, doing the work to try to get something else to be able to drop a photo that's not that great and to get a variety of product shots that your business can use. I do think maybe if this does take off, we might look back at this in five years. You know, similar to a, you know, chat GPT type moment for creatives. All right, speaking of potential job loss and doing things differently, Andrew Yang is sounding the alarm on AI and jobs. So the former US Presidential candidate Andrew Yang has warned that AI could cut up to 50% of white collar jobs as companies pursue efficiency and cost cutting measures. So Yang explained that the stock market will reward companies that reduce headcounts and penalize those that don't, creating pressure to automate office jobs. Weird. That's exactly what I said in the very first episode of Everyday AI in 2023. Anyways, the ripple effects will extend beyond office workers to service industries like dry cleaners, dog walkers and hair stylists. Well, according to Yang, as fewer people go to offices and reduce demand for those services. So he had previously advocated for a $1,000 monthly Universal Basic Income, or UBI, to support workers displaced by automation and AI technologies. So a related stat here. So a recent YouGov poll found that 63% of US adults expect AI to reduce jobs, while only 7% believe it will create new ones. And if you're a longtime listener on this show, you already know my take. AI will ultimately quote, unquote, take more jobs than it creates. And I do think the future of full time work is very different than we've grown accustomed to over the last few decades. So this debate comes amid a broader discussion about AI safety and risks highlighted by some top researchers that have resigned from leading AI organizations over the last month or so as AI's capabilities continue to change. Right. One thing, and I'm glad that we're seeing more economic based benchmarks out of the big AI labs such as gdp, Val. Right. But we're starting to see how even a base model that you use on the front end can complete full knowledge worker tasks front to back in a single prompt. If you know what you're doing, if you're using the Right model. If you're providing the right context, yeah. AI models are better than almost all of us at knowledge work. Right. If you can sit in front of a computer and you have to do something. And that's probably most of us listening. Yeah. Today's AI models. According to benchmarks that are judged blindly by experts, AI models are better across the board. All right, in our last big AI news story of the week, we got some more details on what OpenAI is reportedly working on for their consumer hardware. And at least one of the things that they're working on is a little surprising. So OpenAI is preparing to enter the hardware market with a new family of AI powered devices, starting with a smart speaker. Expect to launch no earlier than February of 2027. So according to reporting from the Information, OpenAI currently has, get this, more than 200 employees working on AI devices including, here's the three the smart speaker, smart glasses and a smart lamp. Yeah, that's what I need. Light. That tells me what I'm doing right or what I'm doing wrong. All right. Anyways, the first device according to reports could be released or might be the smart speaker that could be priced between 200 and $300. Featuring a built in camera to gather data about users and their environment. Yeah, I know a lot of people if that is the ultimate how the device lives, a lot of people aren't going to like that. But we've had that for a long time. Y' all like Meta's had their, you know, kind of laptop or not laptop, they're kind of iPad esque, you know, swivel smart assistant that has a camera. There's multiple versions of the Amazon smart device that has that. So a lot of people are going to say like oh my gosh, a camera. It's been happening for like close to a decade. Anyways, the smart speakers camera capability suggests a focus on enhanced interaction and contextual awareness, potentially changing how users interact with AI at home. So mass production of OpenAI's smart glasses is not expected until 2028, indicating a longer development timeline for the wearable tech. So the hardware push follows OpenAI's 6.5 billion acquisition of IO products. The startup founded by famed and former Apple designer Jony I've showing a strategic investment in design and augmented reality. So the move places OpenAI in direct competition with companies like Meta, which has found success with its Ray Ban AI enabled smart glasses that feature cameras for recording and streaming. Also, we've seen recent reports that say Apple and Google are close to developing and releasing their own smart glasses. So OpenAI has not yet publicly commented, commented on these plans, but the company's entry into hardware could impact how AI tools are integrated into daily work and life. So this goes back to that initial report, right, that shows, oh, Anthropic's revenue may catch or surpass open AIs. And I'm like, not in the long run, I don't think so. Right, because OpenAI, a lot of people have maybe called them distracted over the past year or so as they started to work on their own infrastructure, right? On their own chips, partnering with others, but on their own chips, on their own consumer hardware devices, right. Going different places on the software side. So people, you know, I think have been saying, and plenty of reports are saying, oh, OpenAI is not focused. And I would say the opposite. I'd say OpenAI understands the writing on the wall that if you just rely on the majority of your revenue to come in through API, which I think I could be wrong, right? But I think that's where Anthropic has really been, you know, taking it to the bank at any point, right? The way systems are built now, they're built modularly, right? So I, I would think Anthropic at any point could lose billions of dollars with people snapping their fingers and saying, oh, now all of a sudden Gemini is untouchable, right? It's, it's a third of the cost and twice as powerful. I'm switching off, right? Or as these, you know, Chinese models that may or may not be distilling from American US companies offer free or free 99 cheap AI models that are outperforming Anthropic models on certain benchmarks. I don't see what's going to, you know, lock people into anthropic. Whereas OpenAI and Google and technically, you know, Microsoft by proxy, I think that have much more long term stability when it comes to future revenue production, at least compared to Anthropic. So I don't know. There's my random $0.02. All right, so yeah, Joe says when they invent a wearable Star Trek communicator pin, I'll start listening. But seriously, a smart speaker, Hey, I personally I can't wait for the new Google Gemini at home to come out because even the new, you know, A L, E xa. Sorry, I know I say that out out loud sometimes and then I get emails about how I trigger everyone's smart device at home, you know, I have the new Alexa, you know, plus it stinks. It's not good. It's absolutely terrible. Right. So we've been waiting for this new smarter version of Amazon's Assistant for like two years. It comes out, I can't use it. I mean, I can use it. It's absolutely terrible. So I'm fine with the smart speaker. Maybe that's just my personal use. Right. My wife and I are constantly, you know, asking smart speakers for things. So I'm happy that OpenAI is reportedly getting into the game. I can't wait for maybe it's released. I know it's supposed to be released in spring with the new Google Gemini home. Right. Give me an actual AI powered smart assistant that works. That's all I want. Right. All right, enough of that rant. Let's go over the bullet points on what's new and what's next. So we went over the top. You know, each week we usually do the top seven or eight stories, give them a little more depth, but there's always so much to cover. So at the end we go over the what's new and what's next. Sometimes these are smaller news stories that are have already happened. Sometimes they're rumors or leaks. All right, but let's get into them because there's a lot and some actually pretty big ones if I'm being honest. Okay, so OpenAI rolled out interactive code blocks in chat gbt. They're really good. Anthropic released Claude code security which finds hidden code flaws. Xai dropped Grok 4. 2. 0 beta. Yeah, of course. It's called Grok 420. With four AI agents working together for faster, smarter responses. Microsoft is testing a unified tasks feature for Copilot. Meta expanded their Nvidia deal to buy millions of AI chips from the chip maker. Figma partnered with Anthropic to more easily transform AI code into editable designs. Replit released their Gemini 3.1 powered animation feature where you can just vibe code. A video now. Right. It's more of like a typography based video, but still really, really good. Google released Lyria 3. Yeah, this could have been a news story. Really good. The updated version of its AI music generator Quad code on desktop. Can now preview the code and running apps. That's been a great update, FYI that I've been using a lot. Google released its Google AI Professional certificate from the grow with Google program. Reports show OpenAI is rolling out a new Pro Light plan. All right, so reportedly going to be $100 for a Lite version of the Pro plan versus the normal $200 Pro plan. We'll see, I've been on the 200 Pro plan for a while. I use it a ton. I don't like running into limits, especially on ChatGPT, so. But I. I'm guessing it's going to be pretty popular. World Labs raised $1 billion to advance 3D spatial AI. Microsoft is testing their version of Notebook LM's deep dive in Co Pilot Online with essentially, you know, kind of two virtual or AI characters debating something. Speaking of that, Notebook LM rolled out slide revisions. Actually, looking back at it now, that should have been its own news story, if I'm being honest. So many people use different AI tools to get slides right and. But the problem is is there's always some things wrong or some things you want better. Well, now Notebook LM that is powered by Nano Banana Pro has slide revisions big. All right. Pika Labs released Pika AI selves, which is supposed to be kind of AI avatar versions of yourself with persistent memory. OpenAI updated its invite only Aardvark security tool and rebranded it as Codex Security Anthropic made Claude in PowerPoint available to Pro users. So now to all paying users and not just those on The Max plan, OpenAI added tab organization to its Atlas browser. ChatGPT ads are now live and being tested in some free and also ChatGPT Go accounts. Meta's Ad Manager now features Manus AI tools. So yeah, Manus features rolling out in Meta, Meta features rolling out and Manus a lot there since the acquisition. Some new leaks show that OpenAI may be closer to releasing its adult mode and an actual something. My Gosh, thank you OpenAI for this one. OpenAI expanded the context window to 256,000 tokens in GPT5 thinking modes. This is fantastic. I love this. All right, so a lot happening in today's show, so I hope this was helpful. So if you do find value, whether it's in our Monday AI News that Matters update or you're listening throughout the rest of the week, you know what, I would really appreciate if you did one thing real quick. If you are listening on the podcast, please, whether it's, you know, Apple or Spotify, Player fm, wherever, wherever you're listening, please take a couple of seconds to subscribe to the show. That really helps us and helps others that are looking for an unbiased way to keep up with AI news. And then also if you could leave us a rating on those platforms, I'd really appreciate it. And then like I said, if you are listening online, repost this, help other people out, then go listen to 712 and 7 13, our 2026 AI predictions and roadmap series. So thank you for tuning in. Hope to see you back tomorrow and every day for more Everyday AI. Thanks, y'. All. Foreign

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 everyday AI.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.