EP 625: Sora 2 release and update, OpenAI and AMD partner, ChatGPT Visual Agent Builder incoming 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.

This was another week of AI news updates where OpenAI legit grabbed like every single headline. I mean, this is one of the biggest weeks ever for the company with some new development, product releases, and partnerships for OpenAI that we haven't really seen before, like ever. No exaggeration. Look, we had the viral hit that was Sora 2, hot off the press's partnership with AMD that could see OpenAI with a 10% stake in the company. And oh, speaking of the company, OpenAI became the most valuable private company in the world. Oh, and they're set to release a visual agent builder in the next few hours. But don't worry, it's not just OpenAI. There's plenty of other AI news from the other big players like Anthropic, Meta, Google, and Microsoft. Sheesh, what a week in AI. Let's get into what actually matters. What's going on, y'all? My name's Jordan Wilson and welcome to Everyday AI, 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 the world of AI, but how we can use all of this information to grow our companies and our careers. If that's what you're trying to do, it starts here with the unedited, unscripted livestream podcast. But if you want to take it to the next level, make sure to go to our website at youreverydayai.com. There you can sign up for our free daily newsletter. Each day we recap, uh, the highlights from the livestream podcast as well as keeping you up to date with all of the other AI news. So speaking of AI news, this is our weekly Monday segment. So if you're brand new to the podcast, welcome. All right. We do this thing live, 7:30 AM Central Standard Time. And on Mondays, most Mondays at least, we recap all of the week's news. On Tuesday, bring you a hot take. Wednesday, we do AI at Work Wednesdays, which is essentially showing you a new tool or mode and learning together live. And most Thursdays and Fridays, we do interviews. So let's get straight into the news that matters this week. And the big one, which we'll have final confirmation on in a couple of hours, is OpenAI is previewing and launching a new visual agent builder with a drag-and-drop canvas for creating AI agent workflows, according to reports from Testing Catalog. So this could put them into a little bit of competition with platforms like Zapier in their new agent builder. So we'll have this kind of, uh, in stone, or at least finalized, probably here in a couple of hours, as OpenAI is, uh, kicking off their Dev Day today at noon Central. So yeah, we record the livestream podcast at 7:30 AM, so we should have final confirmation in a couple of hours. But it looks like all reports are showing that OpenAI is coming out with a legit game-changing release with the Visual Agent Builder. So the tool's main new angle is its promise to lower the barrier for building production-ready AI agents by replacing manual code writing with a new visual drag-and-drop canvas that speeds prototyping and deployment. So yeah, you've— obviously been able to build agents on the backend with OpenAI's SDK, but now it looks like they're priming it more for frontend kind of deployment. So if you can drag and drop something and build a visual canvas, it actually looks fairly straightforward and easy. Also reportedly, users can start from predefined templates such as customer service bots, data enrichment pipelines, Q&A agents, and document comparison flows. Flows to rapidly assemble solutions without writing extensive code. So the canvas offers modular building blocks, including logic nodes. What that means is kind of like if this, then kind of connections, as well as connectors like the popular MCP protocol, user approvals, and built-in visual guardrails which I know is going to be a huge area of concern when we talk about building agent flows. So a responsive user interface includes a sidebar palette of components plus a preview and testing modes. So deployment begins within a dedicated OpenAI platform, making it easier for developers and even non-experts to create, test, and operationalize agents across departments with consistent security and infrastructure. So pretty big news, y'all. And let me know, is this something that you all want us to cover a little bit more on the Everyday AI Show? This is one of those where I think we probably would. You know, what we really try to appeal to, there's plenty of, you know, podcasts and resources out there for highly technical people. So as an example, you know, OpenAI has always had the ability to build agents with their SDK and API, um, platforms. But we haven't covered that too much because it does require a little bit of technical ability. But with this, it looks pretty easy, uh, you know, drag and drop interface. I'm thinking we're probably going to spend a little bit of time, uh, depending on the release, rather this week or next week, going over. So, uh, livestream audience, number one, it's good to see you. Robert saying good morning from Dallas. Jose joining from Santiago. Sarah joining from London. Got people from all over. But yeah, let us know if we should be covering this over the next, next week or two. It's getting harder, right? Because I know our audience for the most part. Yes, there are some very technical people, but Everyday AI, it's for everyday business leaders. So this is one of those areas where it's like, okay, I think this is something that we should probably pay some attention to. Uh, yeah, love this. Phil said probably it will make it to the UK in 2028. Yeah, we'll see. We'll see how long that takes, Phil. Uh, all right. More OpenAI news, y'all. Sorry. If you're not an OpenAI fan, they dominated the headlines this week. Uh, so more breaking news. According to a new report from the Financial Times, OpenAI and Johny Ivy's teams are incurring, uh, encountering significant technical hurdles that could delay their planned 2026 launch of their new palm-sized screenless AI device. So we've been talking about that over the last few months, this new partnership between OpenAI and famed Apple designer Jony Ives. This new kind of Apple-esque OpenAI hardware device. But it looks like they may be running into some technical snags. So the new AI device is designed to take continuous audio and visual cues from the environment and respond without relying on a traditional screen. So the key challenge, according to reports, is the proposed, quote unquote, always-on listening approach, as sources say the team is struggling to ensure the assistant engages only when truly helpful and ends conversations at appropriate times, which is critical for privacy and usability. So the project faces unresolved questions about the device's personality as well, including how it expresses tone, initiative, and boundaries, all of which shape, trust, user comfort, and ultimately decide whether people are actually going to buy this thing, right? If it's going to be truly helpful or more of a privacy concern, because the privacy handling, according to the Financial Times, has been a central concern that may push some delays, particularly how the continuous sensing will protect user data and manage local versus cloud processing, as well as offering transparent control to prevent unintended recordings. Also, computing infrastructure requirements are likely involved in the advanced on-device processing, low-latency edge services, and scalable cloud inference. So OpenAI, as we've been talking about over the last couple of months, did acquire Johnny Ive's device startup called I/O in May for $6.5 billion, with OpenAI CEO Sam Altman saying the goal was to create a new generation of AI-powered computers. So we'll see what happens here with this device. When they originally mentioned a 2026 timeline, I was like, oh, aggressive, but also seemed like OpenAI has been able to ship faster than traditional kind of the tech trillionaire companies. So I think a lot of people were like, okay, yeah, 2026 at the time, a year out, maybe that's feasible. We saw that OpenAI, we reported a couple of weeks ago that OpenAI had essentially poached dozens of Apple hardware leaders as well as had already established official partnerships with Apple's hardware suppliers as well, which made it seem like, okay, this could actually be one of the fastest-to-production AI hardware projects ever. But now we are hearing some first reports of some potential delays. All right, next, which buried in all this OpenAI news, this actually might be one of the pieces of news I'm most personally excited about because Perplexity has made its Comet AI browser free for all. Yeah, got buried in the news cycle, so I had to make sure to highlight this one. So Perplexity has officially opened access to its agentic AI browser called Comet, for any member, even if you are a free Perplexity user at no cost, removing the waitlist and invite codes as well. So you do have to have a little bit more of a powerful computer to use this. So it's currently available for desktop, so it's not yet available for mobile apps. So it's available on Windows 10 and 11 devices as well as macOS if you have an Apple M1 or newer chip. So Comet's key feature is a built-in AI assistant that lets users ask questions, assign tasks, and interact directly in the browser without opening a separate chatbot tab or app. So the move positions Perplexity to capture probably a little bit more of search traffic because what Comet also does, if you go into the, the URL or the address bar, you type something, if you type a query in there, Well, it obviously opens up Perplexity instead of Google Search. So they're really looking to cut into Google's browser share there because right now Chrome currently commands nearly 72% of global browser market share. And Google pushed AI mode into its search experience as well. An important thing to know if you're a big Chrome user or Chrome fan, Perplexity's Comet is going to seem like the same thing because it's built on top of Chromium. So you can actually sign in with your Chrome credentials and everything will be synced. So your bookmarks, extensions, etc. And that's the same thing for Microsoft's Edge browser as well. So it's also worth keeping an eye on if there's going to be any of the major agentic browsers that aren't built on top of Chromium, because right now the biggest ones are obviously Google Chrome, which we've covered over the past couple of weeks, has started to roll out some agentic features. Perplexity, probably technically the leader in this space with Comet, built on Chromium. Microsoft Edge, built on Chromium. And reportedly OpenAI's browser as well, their new agentic browser, which maybe we'll hear word of on that in a couple of hours, maybe we won't, but also reportedly built on top of Chromium. So, even though some others are starting to claw back a little bit of share, ultimately Google still owns the browser market, even if it's through their, their open source chromium. All right, next piece of AI news. How, how did this get buried in like our fourth biggest story? Uh, my gosh. So OpenAI is now the most valuable private company in the world. So according to Bloomberg, OpenAI facilitated a $6.6 billion share of sales last week held by current and former employees pushing the company's valuation to $500 billion, the highest ever for a privately held company. So purchasers in the transaction included SoftBank, Dragonair Investment Group, Thrive Capital, and others, signaling continued investor confidence in OpenAI's trajectory. So without getting into too technical of details, you're like, well, what the heck does this mean? Well, essentially, early employees, whether they're still working with OpenAI or previously, right, they get shares in a company. And a lot of times there are certain points where employees are allowed to sell those shares to investors. And this was one of those instances. And then you essentially get a new valuation on a company because it's dictated by what the, you know, both your last kind of price, but also by what investors are willing to pay for it now. So it was a secondary transaction, meaning that the proceeds went to insiders rather than OpenAI itself. But it does serve as a major retention tool among, amid tightening competition for AI talent. Because, yeah, those OpenAI current and former employees that didn't sell their shares or held on to the majority of them are now sitting nice, right? Yeah. There's a lot of kind of jokes now on the internet that, hey, if you ever want to live in San Francisco again, you're probably not going to be able to because current and former OpenAI employees all essentially just got rich, right? Probably thousands of new millionaires were just minted and the price of all property in San Francisco is going to go through the roof. So OpenAI's last primary funding round closed in August at a $300 billion valuation. So this secondary sale pushed the new valuation to $500 billion. So the previous most valuable private company was SpaceX at a $400 billion, uh, valuation. So no surprise here. And another strange thing. Oh wait, I told you guys this would happen back in 2023. That's the cool thing about doing a daily AI podcast. When I say something that people think is kind of crazy and I'm like, no, go back and look. Same thing I said back in 2023, NVIDIA, which at the time wasn't a top 20 company in the US. I said NVIDIA is the most important company in the world. No one believed me. It happened. OpenAI, same thing. I said they're going to be the biggest private company in the world until they go public. And here we are. So also, there was a recent non-binding agreement, which we mentioned a couple of weeks ago, between OpenAI and Microsoft, which had reportedly invested about $13 or $14 billion into OpenAI for a 49% equity stake. But the new agreement between Microsoft is widely viewed as a as the groundwork for OpenAI's conversion into a for-profit entity, which could allow them to, in the future, actually go public. So huge news, right? And even if you look at public companies and their current valuations, let this sink in. Ready? And yeah, maybe that was a little, uh, not so punny, uh, play at open or at, uh, Elon Musk's takeover of Twitter. Let that, let this sink in. Ready? But even if you look at public companies, this puts OpenAI in the top 20 companies, private or public, in the US, even going ahead of companies like Netflix that have been around for decades. Right. So pretty big news here on the OpenAI front. All right. Let's take a very, very quick break for a word from our partners. This podcast is supported by Google. Hey folks, Steven Johnson here, co-founder of NotebookLM. As an author, I've always been obsessed with how software could help organize ideas and make connections. So we built NotebookLM as an AI-first tool for anyone trying to make sense of complex information. Upload your documents and NotebookLM instantly becomes your personal expert. Uncovering insights, and helping you brainstorm. Try it at notebooklm.google.com. All right, let's dive back into it. Um, are you using Meta AI? If so, well, those conversations are soon going to become the baseline for future targeting within Meta's platforms. So according to Meta, the company will begin using people's interaction with its Meta AI assistant to shape ads and content recommendations across Facebook, Instagram, Messenger, and other Meta apps starting December 16th, as the user notifications are going to start rolling out tomorrow. So Meta says that Meta AI now has more than 1 billion monthly active users, which make this change pretty significant as it affects a good chunk of the global population. So a Meta spokesperson said that conversations with Meta AI about topics like travel or shopping will also inform what your Reels or posts and ads that show up on your feed. Also, voice interactions for those that are using Ray-Ban, the Meta Ray-Ban smart glasses will also feed into Meta's recommendation engine. So, you know, I know people have always been talking about for years like, oh, I was talking around my house and, you know, now all of a sudden I'm seeing, you know, these ads on Instagram or whatever. Well, now this is confirmed and it's going to be a quote unquote baked in feature. So love it or hate it, can't really get around it. So also, what's worth noting, WhatsApp interactions with Meta AI will not be included unless the user links their WhatsApp account to sibling apps like Instagram. Also, users cannot opt out of this recommendation change. But if you don't use Meta AI right now, obviously the updates will not apply to you. Live stream audience, what's your thought on this? I don't use Meta AI. I don't really go on Facebook, Instagram, so this won't personally affect me. However, I'm weird. I like hyper-personalized advertising, right? Like as an example, I have, you know, I use Chrome, I use Google products, you know, we use YouTube TV in my home. I like hyper-personalized ads. I don't care. I would actually prefer it if I'm talking to, you know, a chatbot. And I think this is ultimately what OpenAI's new Pulse feature is about, right? It's, it kind of, right now, Pulse from OpenAI for ChatGPT Plus users delivers this like morning daily brief based on all your chat interactions. But I do think that's the future for OpenAI to be a big a big advertiser. But personally, I like this. But I know people's, you know, some people's viewpoint of Meta in terms of their privacy, not gonna like this. But I don't know, if I was a Meta, Facebook, Instagram user, I probably wouldn't mind. All right, this seems like weeks ago, y'all, but there is actually some huge breaking news stories last Monday right after our show aired that we have to mention this week, even though in like AI news time, it seems seems like years ago, but Anthropic did release, uh, Sonnet 4.5, their newest and most powerful AI model. Uh, so, uh, Sonnet 4.5 has some new practical features designed to make real-world coding and automation faster, safer, and more effective. So one of the biggest new features is the improved agentic capabilities. So Anthropic says Sonnet 4.5 excels at autonomous long-duration tasks, allowing it to maintain focus for extended periods, reportedly over 30 hours in some trials on complex multi-step projects like developing an application from start to finish. Also, the coding performance. Obviously, this is where Anthropic has really carved its niche in the software development space. The new benchmarks here, obviously industry-leading with Anthropic claiming it is the best coding model in the world against benchmarks like SWE-Bench Verified. We'll see how long that lasts because not only is today Dev Day from OpenAI, so we may see an updated model even though OpenAI did release a version of GPT-5 specifically for coding, a new one that no one really talked about like a week or two ago. Well, I think we might be seeing something from Google Gemini here in the coming days or weeks as well. So we'll see how long Anthropic's Claude Sonnet 4.5 can kind of hang on to the title of, you know, hey, this is the preferred coding tool, as well as Grok's new coding models have done pretty well as well. All right. So much more breaking news. This is minutes before the show started. Yeah, we actually kicked the show off at like $733 this morning because of this piece of breaking news. OpenAI just— yeah, there is some controversy with Sora too, which we're going to get to in a minute, but just a bunch of friggin home runs from OpenAI. Here's another huge one hot off the press. OpenAI and chipmaker AMD entered into an official deal which sent AMD's shares soaring. And I'm sure that $500 billion valuation is a little bit more for OpenAI. So OpenAI and AMD struck a multi-year partnership to roll out AMD Instinct AI chips at a massive scale for OpenAI, pushing AMD stock up 25% in pre-market trading. Yeah, unheard of for a company to shoot up 25% on a single piece of news. Like this. So here is how this partnership works. And well, ultimately, this is going to impact Nvidia and maybe the rest of the AI landscape as well. So OpenAI will now be able to deploy 6 gigawatts of AMD GPUs over multiple years, starting with 1 gigawatt in the second half of 2026, the company said in a joint statement. So AMD granted OpenAI also the right to buy up to 160 million AMD shares if OpenAI hits deployment and stock price targets fully used, that could give OpenAI about a 10% stake in AMD based on current shares outstanding. So very interesting partnership here that gives OpenAI, now the most valuable private company in the world, a 10% stake in AMD, which has been one of the, you know, a been a mainstay, uh, in public trading for a long time. This is kind of weird, right? Like, if you would have told someone 5 years ago that, you know, not saying AMD needs OpenAI, but you couldn't make that case, right? AMD's stock shot 25%, and their stock has been struggling against, uh, their chief competitor Nvidia over the last couple of years as Nvidia has just become a rocket ship gone to the moon. AMD and Intel have struggled a little bit, but we just saw Intel actually enter into a strange-ish partnership, official partnership with the US government, where the US got a 10% stake in Intel. And now OpenAI, a private company, potentially getting up to a 10% stake in AMD. So some pretty big news here. And this is also a multi-billion-dollar agreement, with some estimates placing the total revenue for AMD over the next several years in the tens of billions of dollars in terms of new revenue. So this makes AMD a core supplier to OpenAI for all their future projects and stands as one of the largest AI chip rollouts to date, aiming to ease supply constraints domestically and reduce reliance for OpenAI on a single vendor. The move also comes 2 weeks after OpenAI's separate $100 billion equity and supply agreement with Nvidia, which set aside 10 gigawatts of capacity. And Nvidia's— with this recent news with OpenAI and AMD, Nvidia shares fell about 1% in pre-market trading. So if you care about the stock market and AI and chips, keep an eye on AMD and Nvidia over the next 48 hours to see how this shakes out. Jobs news. That's going to be coming today. So in an exclusive report, Axios got a hold of a new Senate Democrat report that is going to be released that shows that AI could cut up to 100 million US jobs. So according to a forthcoming report not yet released from the Senate Democrats on the Health, Education, Labor, and Pensions Committee, a ChatGPT-based analysis projects that AI AI and automation could eliminate nearly 100 million US jobs within the next decade, making this a major flashpoint in the national debate over how to regulate AI. So this report, which is expected to be released today, estimates that 89% of fast food jobs, 64% of accounting roles, and 40% of trucking positions could be replaced by AI-driven artificial labor, signaling potential disruption across both white and blue-collar sectors. So the analysis argues that AI's current trajectory is not only about productivity gains, but also about concentrating wealth and power by reducing labor costs and increasing corporate margins. Oh, weird. Another thing I've been saying for multiple years, right? Everyone's saying like, oh, AI is going to create creates, you know, more jobs than it takes away. And I've been very honest with you guys from the start. I'm like, no, it won't. AI is definitely going to replace more jobs than it will create. And I do think this is going to be one of the first large-scale studies that might actually show that. So Republicans here in the US counter that the US must lead in AI development and caution that heavy regulation could cede advantages to countries like China, framing the debate as a balance between innovation and worker protections. So this Democratic-led report criticizes the current Trump administration's approach as favoring deregulation and favoring Silicon Valley insiders, alleging that executive actions from the Trump administration weakened federal oversight and discouraged state-level regulation. So to respond to the projected displacement, the report proposes policies such as a 32-hour workweek, which I'm sure most US workers wouldn't mind, as well as profit sharing and a robot tax to fund worker support and transition programs. Robot tax. Interesting. Yeah. So maybe robots will be doing more of our work, but then also getting taxed for it. You know, who's looking out for the robots? All right. More AI news. Microsoft, pretty big announcements as they rolled out two new AI agents, their Agent Mode and Office Agent for Microsoft 365 Copilot. So the biggest limitation, though, well, right now they're not available to everyone. So the Office Agent is only available to US-based customers who have a paid plan, but personal and family Copilot subscribers, not to most enterprise customers, which is interesting. Also, Agent Mode. So you have the Office Agent and then Agent Mode. Agent Mode is live on the web to use, to be used inside of Word and Excel via the Frontier Early Access Program. And the Excel version requires the Excel Labs add-in to enable agent mode, while Word shows agent mode as a selectable option in the Copilot pane. So, well, what the heck do these agents do? Well, make sure to go watch our episode 621, maybe the video version on our website at youreverydayai.com. So you can kind of see, I think we did a demo, we showed a demo, but essentially This is one of the first, I think, big moves from Microsoft to bring agentic capabilities inside the normal version of Copilot. Right. I think Microsoft has actually been a leader in the agent space. Well, until today. We'll see in a couple of hours from OpenAI with their Microsoft Copilot Studio. However, even though Microsoft Copilot Studio is geared more towards low-code and no-code, it seems like a lot of the non-technical business business world hasn't really caught on yet to Microsoft Copilot Studios. So this is a big push from Microsoft to bring in agentic capabilities inside of the actual programs that people are using most. And, you know, one of the capabilities that I think most people are looking forward to is the Office Agent, which you use inside the Microsoft Copilot Chat, and it can essentially just build PowerPoints, right? That's what everyone's waiting for. So we'll be keeping an eye on the rollout to see when this goes out to these two agents go out to enterprise customers. The other noteworthy thing is this is one of the first big public releases from Microsoft where they are offering and using a model that is not OpenAI. Because Microsoft says that agent mode on the agent mode side, it does use OpenAI's latest reasoning models while these the Office agent inside Copilot Chat actually uses Anthropic's Claude Sonnet 4 and Opus 4.1. So pretty big news there from Microsoft. And I have been using agent mode inside of Microsoft Excel on the web. It was actually pretty impressive. I did briefly touch on this in episode 621, so if you want to know more, make sure to go check that one out. Actually, the Microsoft Copilot agent mode in Excel handled some things that I tried it inside in Profit Cloud. I tried it inside GPT-5. I tried it inside Gemini 2.5. Actually, it handled some things that those other platforms couldn't. So especially if you are a heavy Excel user, you're definitely going to want to check out more on that. All right. And our last big piece of AI news for the week, and this was big and we have some updates. OpenAI officially launched their Sora 2, their updated AI video generator with integrated audio and a mobile-first app. So OpenAI says their new Sora 2 also produces synced dialogue, music, sound effects, and entire soundscapes inside the platform, while there's also improved physics and realism and longer multi-shot narratives that maintain temporal consistency. In other words, you can go in right now with a simple prompt inside of Sora 2 and get a 10-second output that actually uses multiple camera angles with synced audio background noise, and it is character consistent, which we haven't seen even from previous platforms that were leaders in the space like Google's V0.3. Also, probably the key feature that a lot of people are talking about and using is the new Cameo feature that lets users insert verified likenesses of themselves or use their friends' Cameos and essentially create AI videos of yourself. So we went through kind of the pros and the cons of that, right? So you can essentially create an AI avatar that you can then put anywhere with a simple text prompt. You can create a video. I could create a video of myself walking the streets of Chicago, talking to people about everyday AI, right? So it depends on how you set up your Cameo and who you give it access to. So you have to do this inside the iOS app right now. To set up your Cameo. And then you can say, oh, I only want to use it myself. I only want to use it with people I'm connected with on the new Sora app because, yeah, it kind of rolled out as a social network, kind of a TikTok-esque kind of feed. So you can create a Cameo, keep it only for yourself, only people you're connected with on the Sora social app, or you can open it up to everyone. Like, as an example, OpenAI, CEO Sam Altman made his Cameo available to everyone, which, you know, obviously added to the viral launch that was Sora too, right? You know, as an example, one that went viral on all social media networks, which I'm sure a lot of people saw, was Sam Altman, you know, quote unquote, going in and stealing GPUs from Target. So also it went to number 1 on the App Store. So right now for free apps, actually OpenAI has the number 1 and the number 2 spot with Sora, even though Sora did debut and is still invite only, right? So we actually gave out a couple invite codes in our newsletter last week. But once you do sign up in the iOS Sora app, you can use it on the web, right? Which is great because because I don't want to be logging in to something essentially, let's be honest, it's a little bit of brain rot, right? And OpenAI did acknowledge that, right? But it is essentially the TikTokification of AI video. So to sign up, yes, you do have to use the app, which is a social media AI brain rot feed by default. But after you do that, you can then go use Sora on the web. So that's not it. There were actually a ton of updates over the weekend and some good or bad depending on who you are and what your view is on this. And one of the biggest thing is, well, one of the reasons this thing went so viral so quickly is because it allowed people to use copyrighted content, at least temporarily. Right. I mean, literally, people were making their own entire episodes of shows like South Park, right? Yes. Piecing it together, 10-second clips at a time. But you literally had people putting out entire South Park shorts. So, yeah, that kind of is gone now. So let's talk about some of the updates that happened to Sora 2 over the weekend. So OpenAI tightened Sora's safety while also adding more robust Cameo controls, clearer watermarks, and monetization signals for rights holders, which is what they call that now, right? Instead of just saying like copyright infringement, they're saying they're going to work with rights holders. So let's talk about the Cameo preferences. So the updated Cameo preferences let people restrict how their likeness is used, including you can just you know, block certain type of political content, certain words, certain scenarios. Essentially, you can assign custom instructions if you are allowing others to use your Cameo, which a lot of people just wanted to do that. I would never do that. But with natural language like custom instructions, it will then restrict how others can use your likeness, which will hopefully in the long run cut down on deepfakes misinformation and disinformation. The other, the other big thing is they're updating their watermark because it didn't take long. And I called this, I literally called this once it dropped. I'm like, by the weekend, you're already going to have multiple companies that have working watermark removers. Obviously, that happened. So, you know, OpenAI is going back to the drawing board because people were able to very easily remove those watermarks. Also, OpenAI introduced IP, intellectual property opt-outs, as well as opt-in revenue sharing for rights holders. So there was also moderation filters that were tightened to better catch violence, racism, and copyrighted material, though community reports say that benign prompts are increasingly flagged and their creative flows are disruptive. Yeah. So essentially, essentially all Twitter was over the weekend was users saying, you know, that Sora 2 was nerfed because it refused now to generate any, you know, material, right? Between all of the now tight copyright, you know, restrictions and the new restrictions on things like violence, racism, essentially people who went absolutely wild generating the most obscene and, well, copyright infringement, short AI videos ever. Now people are like, well, I can't even really create anything. So we'll see how this ultimately shakes out. But it's something, especially if you work in media, to see what happens, right? CEO Sam Altman over the weekend kind of jokingly said in an interview, I hope Nintendo doesn't sue us because, yeah, like half of the feed was like Pikachu and Mario. So we will obviously continue to cover this as it rolls out. All right. Yeah, we got to wrap this up because there's a ton more. So here is the quick bullet points of what's new and what's next. So some pretty important and noteworthy updates around the AI world that maybe didn't make our, you know, top 10 stories. All right, here we go. So GPT-5 has been updated. The GPT-5 Instant model now better detects distress from users and guides to crisis resources. Reports from The Information say that OpenAI is targeting 1 billion weekly active users by the end of the year, which is crazy. Google's Nano Banana went generally available and has new aspect ratios default. In Google AI Studio. Great feature there. NotebookLM added the ability to customize response length and conversational style. Microsoft launched a Microsoft 365 Premium tier at $19.99 a month, which people are saying is to compete with ChatGPT. OpenAI updated the Sora and ChatGPT connection because there was maybe a bug or unintended feature where if people signed up for Sora, you have to sign up with your ChatGPT account. And then when they went to go delete Sora, if they didn't like it, it would delete their ChatGPT account. So OpenAI addressed that and kind of unchained those two things. Synthesia released V3 of their AI avatar software, which included a new agent. Microsoft is reportedly going to be releasing connectors for Copilot Online. OpenAI announced ChatGPT checkout mode. Yeah, crazy that didn't make our top 10 news stories because that's going to change the future of shopping. Google is prepping a new feature in Flow for VO3 frames to video. A new coding model called Cheetah is being tested on Cursor, which many are saying is Grok Code 2. Google launched new visual experiences in AI mode. Microsoft made Copilot free for college students who need to sign up by 10/31. So no tricks, go get that treat. And ChatGPT polls will be rolling out to the web instead of just the mobile app for ChatGPT Pro users. My gosh, that was a ton, y'all. If this was helpful, please repost this if you're listening online. We spend a ton of time trying to keep you up to date and also cutting through the BS, the marketing, and telling you, hey, Hey, here's what you should actually be focusing on. So if this was helpful, if you are listening on the live stream, please repost this. If you're listening on the podcast, we'd appreciate you following, subscribing to the show on Apple Podcasts as well as Spotify. If you could leave us a rating, I'd appreciate that as well. Also tomorrow, we're going to be covering, you know, normally Hot Take Tuesday, maybe I'll have a hot take, but we're going to be covering all of OpenAI's latest announcements. So make sure you tune in tomorrow for that. So thank you for tuning in. And if you haven't already, make sure to go to youreverydayai.com, sign up for that free daily newsletter, and we'll see you back tomorrow and every day for more Everyday AI. Thanks, 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 youreverydayai.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.