Everyday AI Podcast – An AI and ChatGPT Podcast - EP 581: Microsoft and OpenAI renegotiating, Google launches new model, and more AI News That Matters

Episode Date: August 4, 2025

There's a new most powerful AI model in townApple is trying to make a ChatGPT competitor.And OpenAI? Well.... they're in a capacity crunch.Big Tech made some BIG moves in AI this week. And you probably missed them. Don't worry. We gotchyu. On Mondays, Everyday AI brings you the AI News that Matters. No B.S. No marketing fluff. Just what you need to know to be the smartest person in AI at your company. Newsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageJoin the discussion: Thoughts on this? Join the convo and connect with other AI leaders on LinkedIn.Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTopics Covered in This Episode:OpenAI Study Mode in ChatGPT LaunchGoogle Gemini 2.5 Deep Think ReleaseGemini 2.5 Parallel Thinking and Coding BenchmarksGoogle AI Mode: PDF and Canvas FeaturesNotebook LM Video Overviews CustomizationMicrosoft Edge Copilot Mode Experimental RolloutOpenAI GPT-5 Model Launch DelaysApple Building In-House ChatGPT CompetitorMicrosoft and OpenAI Partnership RenegotiationAdditional AI Tool Updates: Runway, Midjourney, IdeogramTimestamps:00:00 AI Industry Updates and Competition03:22 ChatGPT's Study Mode Promotes Critical Thinking09:02 "Google AI Search Mode Enhancements"10:21 Google AI Enhances Learning Tools16:14 Microsoft Edge Introduces Copilot Mode20:18 OpenAI GPT-5 Delayed Speculation22:42 Apple Developing In-House ChatGPT Rival27:06 Microsoft-OpenAI Partnership Renegotiation30:51 Microsoft-OpenAI Partnership Concerns Rise33:23 AI Updates: Video, Characters, AmazonKeywords:Microsoft and OpenAI renegotiation, Copilot, OpenAI, GPT-5, AI model, Google Gemini 2.5, Deep Think mode, Google AI mode, Canvas mode, NotebookLM, AI browser, Agentic browser, Edge browser, Perplexity Comet, Sora, AI video tool, AI image editor, Apple AI chatbot, ChatGPT competitor, Siri integration, Artificial General Intelligence, AGI, Large Language Models, AI education tools, Study Mode, Academic cheating, Reinforcement learning, Parallel thinking, Code Bench Competition, Scientific reasoning, Chrome, Google Lens, Search Live, AI-powered search, PDF upload, Google Drive integration, Anthropic, MetaSend Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info) Start Here ▶️Not sure where to start when it comes to AI? Start with our Start Here Series. You can listen to the first drop -- Episode 691 -- or get free access to our Inner Cricle community and all episodes: StartHereSeries.com Also, here's a link to the entire series on a Spotify playlist. 

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. Meet Firefly AI Assistant, now live in Adobe Firefly, the All In One Creative AI Studio. Just describe what you want to create and the assistant handles the rest, orchestrating multi-step workflows across Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. The assistant accelerates execution. Apple is reportedly trying again to build a chat GPT competitor,

even though they've been trying to do that for years. Open AI and Microsoft are reportedly renegotiating their partnership, which could impact millions of co-pilot users. And Google just released an even more powerful AI model, even though they've been at the top of the AI leaderboards for months. But how long until Open. AI strikes back. Well, we've got updates on that as well. It's been a busy week in the world of AI. And if you can't spend hours every single day trying to keep up like I do, then Mondays are

your day. That's when we bring you the AI news that matters. Cut through the fluff, the marketing and the useless and just bring you the bullet points of what actually matters to grow your business and your career. So if that's what you're trying to do, you're in the right place. Welcome to Everyday AI. What's going on, y'all? My name is Jordan Wilson and welcome to Everyday AI. This is your daily live stream, podcast, and free daily newsletter, helping everyday business leaders like you and me, not just learn AI, but how we can leverage the most important aspects of it to grow our companies and our careers. So it starts here with the live stream, unedited, unscripted.

But if you want to take it to the next level, make sure that you sign up for our free daily newsletter at your EverydayAI.com. So most Mondays, like today, we bring you the AI News that matters on Tuesday, coming in with a hot take. So make sure to join us tomorrow when I'm going to be kind of ranting a little bit on AI agents. On Wednesdays, we do AI at work Wednesdays, a practical and actionable way to use a certain AI tool or feature. And then Thursdays and Fridays, for the most part, we do interviews. So I'm excited to get to the AI news, a lot to cover. Live stream audience. Thanks for tuning in.

Let's get straight to it. A lot to cover today in a little amount of time. First, OpenAI has introduced a new study mode in chat, GBT, aimed really for students just to help students cheat less and actually learn more by encouraging active learning rather than just providing answers that students are going to copy and paste and probably not read. So the new study mode is, well, it's available for everyone. You don't have to have like a student or an EDU account. So it's accessed by logging into your chat chbt account and selecting.

If you look at the bottom toolbar, if you go to tools, study and learn. And then from there, it is kind of an interactive experience. So you can proactively, once you choose that, have chat chbtee explain a certain topic to you by saying something like, help me with this homework. Or if you're not in school, you know, help me better understand, you know, this topic. And you can give it a topic. And then from there, it's going to go through kind of a back and four. interactive kind of experience to actually help you learn and to quiz you a little bit.

So I like this. So this feature asks users for details like their grade level or understanding of a certain subject of interest. Then it tailor's responses to guide users through problem solving rather than just supplying a copy and paste answer. So study mode encourages users to engage in back and forth conversation, requiring them to reflect and participate before receiving guidance instead of just passively absorbing information.

So in contrast to standard mode, which chat GPT just directly gives users answers, obviously. But study mode asks users for their own thoughts or approaches first, fostering a more critical thinking and deeper comprehension. So OpenAI says that this is the first iteration of study mode, with plans for further improvements, such as adding the ability to lock the mode to prevent reverting to direct answer seeking. So Open AI has been talking a lot and making a lot,

making a lot of moves in the education space. So I'm guessing this is something that will be used in pretty much classrooms, I would assume, across the country in the coming years. So this first iteration, my thought is really just going to be about gathering data, about how people are using this versus the standard mode of ChadGGBT and hopefully fostering actual learning. I've said this before. I'm a firm believer.

the more you use AI, the majority of people, the dumber a lot of people get, right? A lot of people aren't using this as a thought partner. I mean, this is the way I do it. I feel that I'm consuming more personalized and helpful information than ever before. But I will admit sometimes the information and how I synthesize it and maybe regurgitate it has almost become transactional. So I'm actually using chatyPD in different. large language models to learn things the right way.

But the problem is it large language models make learning so easy and personalized. It's almost hard to retain the information, right? Because you're almost just, it's like info in, info out. So I'm looking forward to using study mode even more to hopefully just retain information a little bit better. So I think there's a lot of great use cases outside of just students. This is the first version. And make sure to join us Wednesday.

That's what we're going to be doing for the. put AI to work Wednesday to show professionals, you know, how you can use study mode, even though I think a lot of people are just looking at its use cases for students. So make sure to join us on Wednesday for that one. All right. Next piece of AI news, Google woke up and chose violence, you know, as in violence and releasing an extremely powerful and capable of new AI model. So Google has released Gemini 2.5 deep.

think mode just to AI ultra subscribers. So you got to have the deep pockets in order to use this new mode. And here's some of the benchmarks and some details about the new Gemini 2.5 deep think. So it's not an actual new mode per se. If you are on Gemini 2.5 pro, there's a new toggle. So it's almost like you're just giving Gemini a little bit extra juice or telling it to use a little bit more extra juice. So the deep think version of Gemini 2.5 scores. an 87.6 on the live code bench competition coding benchmark, a notable joint jump where it was 80.4 previously. So deep think leverages what Google calls parallel thinking, allowing the AI to consider, revise, and combine multiple ideas before settling on the best answer. So it's almost like a

mixture of experts or like what I called back in January, a mixture of models approach. So it should be pretty interesting on how Google uses the information they gather from initial users of this mode. So Google has extended the model's thinking time and incorporated new reinforcement learning methods to encourage the model to go through a deeper reasoning process. The mode produces much longer, more detailed responses thanks to iteration with tools like code execution and Google search. Early feedback from trusted testers and recent research have contributed to improvements, cross-coding, scientific reasoning, and general knowledge tasks. So deep think excels at iterative development and design, making it especially useful for web development and complex programming challenges.

This is a version of the model that Google used to achieve bronze level performance on the 2025 IMO or the International Mathematical Olympiad benchmarks, with a gold-level version being tested by select academics. So yeah, the downside is you got to be on that expensive Google Gemini Ultra plan, which is $250 a month. I think they still have that $125 a month kind of intro to go on. That's what I'm still on. So if we should do a deep think, last year monies, let me know. Just type in Deep Think or the Spotify comments, FYI.

I can't respond to those, but I will read them. I don't know if there's a big enough appetite for this, just because I know that so few people have that $250 Gemini AI Ultra plan. I've been using it for a couple of days. I'm still even for myself trying to find the right task. I think where it obviously shines is in coding, software development, STEM-related queries. So for me, that's not a lot of what I would be doing on a day-to-day. So I'm trying to push it through some more statistical reasoning and just really just crunching large amounts of data and then combining that with live results from Google search.

So I don't know. Should we do a show on this? Let me know if we should. Yeah, a lot more Google. Google whatnot's this week. So Google just released also a ton of brand new features. Well, kind of brand new features, more borrowing some popular Gemini features and bringing it to AI mode in. search, which Google CEO has said that this will be eventually be the default mode.

So if you haven't used Google's new-ish AI mode, it's definitely worth it, especially with some of these new updates, which they kind of said these new updates inside AI mode were for students and back to school. But I think busy professionals are going to make way more use of these. So Google is enhancing its AI mode in search with new features, designed to help students, educators and parents and just everyday business professionals. So now Google AI mode lets desktop users. So a lot of these new features are just available on desktop.

So you can ask questions about images. And soon you'll be able to upload PDFs and expand on capabilities that were previously only available in Google Gemini. So PDF upload support is rolling out in the coming weeks. So users can query. You know, if you have things like lecture slides, study guides, industry white papers, et cetera, and then it will essentially just do a quick version of deep research based on what you upload. So AI mode analyzes uploaded files and cross references web sources

to provide detailed answers, making it easier to connect classroom materials with up-to-date online information. Also, Canvas mode, one of my favorite things in AI right now is Google Gemini Canvas mode. And this is rolling out to AI mode. So I know that's a little confusing or you're like, what the heck? I don't get it. So essentially what's happening with AI mode is it's kind of blending traditional Google search with Google Gemini. And one of the my favorite parts about Google Gemini is Canvas mode,

which is now rolling out to AI vote, a lot of votes. And that allows users to build and continuously refine different study plans, coding projects, et cetera, in a side panel with the option to add context from files. like if you have class notes or a syllabus. Also, search live with video input powered by Google Lens is rolling out to select mobile users. So some of these are rolling out to mobile users. Some of them are rolling out to desktop users. So search live is rolling out to select mobile users in the U.S.

If you are signed up for it in the AI mode labs experiment, so you have to make sure to toggle that on. So that allows real time help with back and forth. as users point their camera at problems, you know, math problems, science problems, or just objects in your everyday life. Chrome users as well will soon see an ask Google about this page option in the address bar, making it easy to question diagrams or content on a webpage and receive instant AI generated overviews. So all of these updates are initially available in English to users who are 18 plus in the

US and India with expanded file support and Google. drive integration plan for the future. So, uh, I'm pretty excited about these new updates in Google AI mode. And I'm finding myself, uh, using actually it's weird. I'm using so many different, uh, Google AI products at the same time. I'm using like AI studio Gemini AI mode in notebook L.M sometimes all for the same project. It'd be cool if Google somehow could roll these all into one and make it more of like,

a vertical tab or if you could enter information and have it be shared across those four different platforms because these four different platforms, even though they're getting a lot of crossover in terms of features and utility, they all still have their own individual strengths and use cases. So it's kind of hard for me like being someone that follows up like follows with this every day because I'm like, oh, you know, AI mode has this in Canvas, but there's some benefits using AI mode versus Joe and I where I would normally use Canvas, right? So it's great. I love all these new Google updates.

They've been absolutely just shipping features like crazy the last few months. Speaking Google and shipping. Another big update Google's notebook L.M has introduced video overview, which can automatically create presentation style videos with the AI narrated AI audio based on users uploaded documents, pulling visuals as well directly from uploaded PDFs, which is really cool, and also supplementing them with custom graphics. So this brand new feature of Notebook LM allows users to fully customize video content by specifying things like profession, audience, tone, and focus, and enabling highly targeted businesses,

the ability to create customized training, onboarding, etc. So video overviews use a single AI presenter, which is a little different than the very popular in viral deep dive AI audio overviews. So on the video overviews, at least in all of them I've seen and done myself, it just has a single AI presenter. And they can generate multiple versions from the same content base, which is a new advancement to their refresh notebook LMs, refresh AI studio. So on the right hand side, there's the little studio panel. The other great thing is you can create, you know, as an example, custom audio overviews, multiple of them based off the same content, which is pretty cool because previously, you could only create one version of an audio overview or one version of a mind map.

So now you can create multiple versions, which is really cool. So let's say you have a bunch of onboarding documents for your company and, you know, you want to be able to have a personalized learning plan for people who speak different languages, people in different departments, et cetera. So before, you know, you could create like in a single AI audio, audio overview, now you could create dozens of them. You know, you could create a version in Spanish for marketers, a version in Mandarin for HR, et cetera, all based off the same content or a combination of same and different content.

So very cool new update from Notebook LM, which one hour, 24 AI tool or mode of the year. If you want to know more about these new updates, make sure to check out episode 578. Watch the video version on our website or on our YouTube channel. That's obviously one visually that's going to help. So go make sure you check out episode 578 if you want to see more on that. Next piece of AI news, all the big tech companies were just making it rain AI candy on us all. So Microsoft just rolled out their new experimental co-pilot. mode in its edge browser.

So this is a new AI powered feature that lets co-pilot search across all open tabs, as well as assisting with certain tasks more on the agenic side, although that's a little more limited right now versus something like chat GPT's agent mode. So copilot mode inside of Edge brings the copilot chat bot directly to the Edge new tab page, which I really like aiming to streamline browsing and task management. So users can allow copilot to access all their open tabs for comparison, such as evaluating different hotels or summarizing the best product choices from multiple shopping tabs that you may have open.

Also, voice navigation is supported as well, enabling users to ask copilot to find information on websites or open specific tabs for product comparison. So Microsoft planned to expand copilot's abilities, to access browsing history and credentials with user permission in the future. So it could do things like book reservations or handle similar tasks on a user's behalf. So the feature is optional, FYI, so users can turn copilot mode on or off in edge settings, ensuring those uninterested in the AI integration can continue browsing as usual.

So co-pilot mode is described right now as an experimental feature and is currently free for a limited time with usage caps on certain functions hinting at a future subscription model. So pretty pivotal move, I'd say here by Microsoft. I think, you know, kind of the writing has been on the wall, the recent momentum behind agentic browsers. So, you know, Microsoft here at least beats Google. to the punch when we're talking about the big players as well as open AI, right?

So all of them are using a version of Google's browser, Chromium. So the open source version of Google's Chrome to accomplish this. So Microsoft Edge and co-pilot mode is built on chromium. And Open AI is reportedly going to be launching their own AI powered browser as well. But I think actually the leader in the space right now for the most, most powerful, at least among the big names, is perplexity with their comet browser. So if you want to know more, we did cover agentic browsers in episode 572. And then if you wanted to know a little bit more about perplexity comet, we did cover

that in episode 569. So again, you can always go watch those, the video versions of those. Those are a little bit more visual on our website for free at your everyday AI.com. All right. A lot more news. Let's keep it going to chew, chew. So we have some new news on OpenAI's upcoming model, GPT5, as well as some potential hiccups that OpenAI may be experiencing. So OpenAI is obviously planning some major rollouts of its next generation GPT5 model and several new features with CEO, Sam Altman, cautioning users on Twitter that,

Yeah, there's going to be some service disruption. So Sam Altman on Twitter recently said that OpenAI has, and I quote, a ton of stuff to launch over the next couple of months, new models, products, features, and more. Then he asked users to, quote, unquote, bear with us through some probable hiccups and capacity crunches, adding, although it might be slightly choppy, we think you'll really love what we've created for you. So Altman also kind of stir the pot a little bit on Sunday providing some screenshots of his GPD5 in production. So we knew that OpenAI was maybe weeks away from announcing their new GPT5 model or architecture as it may be. But by this tweet, it looks like they may be delaying it.

But we're not sure if this means GPT5 is delayed because previously previous report, We're saying it is going to be released in August, or if this is more hinting at delays of other projects, such as their open model. And we did see some new information and rumors on the internet on that late last week that we shared in the newsletter. Or if this could be just delays on other probable AI product iterations from Open AI, such as their next version of their AI video tool, SORA, or their AI image editor image gen. So regardless, same ultimate here, just kind of saying, yeah, we're going to see some delays. So not sure if we're going to get the hot AI summer yet from Open AI or if we're just going to be seeing more and more delays. But this most recent news comes after reports of another Open AI fundraise because reports showed that Open AI did just recently close a new $8.3 billion funding round. led by Dragonier Investment Group at a $300 billion valuation.

With around being five times oversubscribed, geez, so just swatting previous investors away. And this brings Open AI significantly closer to its goal of raising an additional $40 billion by the end of this year. So Open AI, raking in billions of more dollars in investments. We've seen reports just about their user numbers continuing to skyrocket. But here we have Sam Altman essentially saying, yeah, we're going to be hitting some capacity crunches. It's going to be a little glitchy, hiccups, bear with us.

So we'll see if we have a GPT5 in the coming weeks or if it's going to be delayed or if those delays are about some other open AI products. Adobe just introduced an entirely new way to create, bringing the power and precision of its creative suite into one conversational experience. Meet Firefly AI assistant. now live in the Adobe Firefly app, the all-in-one creative AI studio. Powered by Adobe's creative agent, Firefly AI assistant lets you start with your vision, just describe what you want, and shape the outcome as it takes form with the assistant. The assistant orchestrates multi-step workflows, drawing on 60 plus pro-grade tools across Adobe Creative Cloud apps, including Photoshop, Illustrator, Premier, Lightroom Express, and more

to help bring your ideas to life. You can also get started with creative skills, a growing library of pre-built workflows for common creative tasks, like batch editing photos, creating mood boards, portrait retouching, and creating social variations. Every step the assistant takes is visible so you can refine, redirect, or take over at any time. You stay in the driver's seat as the creative director. Adobe Firefly AI assistant now in public beta. See it today at Firefly.adopi.com. Speaking of Open AI, Apple is apparently coming after them again, even though they're their partners. So according to reports from Bloomberg, Apple has formed a new team called Answers, Knowledge, and Information to develop its own AI chatbot that could compete with chatGVT, according to Bloomberg's Mark German.

So this marks a major ship from Apple's previous strategy, which involved them part. partnering with Open AI and offering chat GPT integration with Siri, essentially because Apple failed to build their own large language model chat bot or a smart version of Siri. So instead Apple was just like, yeah, let's partner with Open AI on this. And reportedly, Open AI did not get compensated for that, which was interesting. So there is obviously probably some revenue pass through from people who, you know, upgraded their version of chat GPT via Siri. because Siri would essentially, you know, have users launch or install chat GPT for queries that were too difficult for Siri to handle, which was essentially everything except like what time is it? And I don't even know if Siri could handle that. Regardless, the new initiative

is right now apparently hiring experts in search engine, search algorithms and engine development. So Apple's reported goal is to create a stripped down rival to chat GPT that can crawl the web and answer questions, potentially as a standalone app and as upgrades to Siri, Mac Spotlight, and Safari. So the company previously delayed its upgraded Siri 2.0 Smart Siri rollout, but CEO Tim Cook said on the latest Apple earnings call that Apple is open to acquisitions as well, to speed up its AI ambitions. So according to Bloomberg, this move reflects Apple's not first but best philosophy and signals an effort to catch up in the fast-moving AI chatbot market. So this new report comes on the heels of Tim Cook telling employees that the company,

quote unquote, must win in AI, which is interesting. That's kind of like a high school basketball team saying they must win the NBA finals. It's not going to happen anytime soon, even with a ton of apps. acquisitions. So I'd say that's a pretty silly quote there from Tim Cook, or at least not very realistic. It's actually funny that he said that, knowing that they can't. Anyways, also that followed his announcement to investors of a significant increase in AI spending. So Cook admitted Apple has fallen behind rivals, especially with delayed Siri upgrades, but stress the company's track record of redefining product categories. So despite launch,

launching new AI features. Apple faces mounting pressure as competitors are just straight up lacking them in every way imaginable. So I don't have any confidence in Apple doing anything competitive in the AI space anytime soon. So, you know, we're seeing a, uh, all these reports of a stripped down AI chat bot, but we're also seeing reports that Apple has lost some of their best in brightest AI engineers recently to meta. has everyone, by the way, but I mean, it's not like Apple has the kind of deep AI rosters that the other companies have. They have the money, but they don't have the vision. I don't see, you know, I follow the kind of AI transactions, the engineers, jumping boats. I've seen hardly

no one over the past two years, at least no one notable jump over to Apple, but a lot of Apple people jumping elsewhere, as is the normal, although recently, essentially everyone's going to meta. So it's going to be interesting to see the meta super intelligence, super intelligent labs, MSL, to see what they develop. And also reportedly someone was offered a billion dollars over a four-year span. But Apple's just not right now. They're not paying the money to retain their talent.

They haven't put out any type of product that's impressive on the Gen A.I. Large Language Model side. So I don't know. For me personally, I have about negative. confidence that Apple's going to do anything in the AI space. All right. Our last piece of AI news for the week, Microsoft is pushing for an ongoing partnership with OpenAI as they're looking to renegotiate their current terms.

So according to reports, Microsoft has a 49% equity stake in Open AI after investments of more than $13 billion many years ago. But right now, according to reports, Microsoft is in advanced negotiations with OpenAI for a deal that would ensure continued access to Open AI's technology, even if the startup achieves what it defines as artificial general intelligence. So essentially, with the current arrangement, there's some weird kind of clauses, like in terms of when does Microsoft stop getting access to Open AIs IP, right? So there's different dates, there's different definitions, but essentially it's not ongoing forever. And even if they may not

be admitting it out loud, obviously the partnership with Open A.I has been a major reason first why Microsoft co-pilot, I think, has largely been a success, even though I know a lot of people grumble about Microsoft copilot. But if I'm being honest, I think it's more of a user access and education piece. Like, if you know what you're doing and if you give all your users the right access and training, Microsoft Copilot is really, really good. But I think they've obviously benefited largely from having that open AI kind of partnership, right?

And we've seen reports last week that we shared in our newsletter that they've also been testing GPT5. in limited capacity. So Microsoft co-pilot is about to get a lot better, but Microsoft is trying to negotiate, renegotiate, and extend its partnership with OpenAI so they can keep that kind of competitive edge for a long time. Because let's be honest, you know,

Microsoft CEO, Sadia and Andella has been saying for a while that even if, you know, their Microsoft or their Open AI partnership were to cease, that they're fine and they have, you know, access to models internally, which they do have, I'd say some, you know, some okay models. Nothing, you know, top tier. But I think Microsoft is heavily reliant on OpenAI, obviously, to keep up with Google.

So some more information on this, on this partnership and what they're trying to do. So like I said, Microsoft has invested about $14 billion in Open AI and currently holds right to some of the company's IP. But right now that actually, is set to expire in 2030 or if OpenAI declares that it has reached AGI and there is a monetary kind of trigger that set that off. So the two companies have been discussing Microsoft taking an equity stake in OpenAI and the low to mid 30% range. So knocking it back a little bit, but that would give Microsoft formal shares and the potential

for greater financial returns. So that would allow them to benefit financially as well. So regulatory scrutiny and Elon Musk's ongoing kind of useless lawsuits for being honest to block OpenAI's for-profit transition could still pose challenges to these renegotiations. So Open AI's current structure as a nonprofit overseeing a capped for-profit company limits how much money it can raise and how fully it can commercialize its technology. But the organization is eager to remove those restrictions. Microsoft has built some major. products like Azure Open AI services, their AI foundry, and co-pilot around Open AIs models. So losing access if Open AI achieves AGI or in 2030 could be a pretty big setback for Microsoft

or they would just have to invest a lot of additional money to keep the partnership going or probably at that point, I don't know, acquiring Anthropic if Amazon doesn't do so first. So OpenAI has reportedly told investors it expects to pay Microsoft a lower share of its revenue as it grows and nears a GI and wants to ensure Microsoft uses its technology safely. So if OpenAI transitions to a standard for profit structure, reports say that Microsoft stands to benefit from increased investor returns in more direct ownership in addition to ongoing tech access. a lot of stories this week in AI. But let's get to our bullet point roundup of what's new and what's next. Live stream audience, do you like these kind of bullet points at the end? What I realize is like I try to only do eight to 10 major AI news stories each week.

But there's so many things that I think deserve your attention between things that were just released and things that are coming. So let me know. Do you like this? Yes or no. the kind of what's new and what's next. But let's go ahead and go over because there's a ton of these kind of random. All right, not so random. So pretty big.

All right, here we go. So Zipu, I believe, the Chinese firm ZAI, has unveiled an open source model that looks pretty impressive called GLM 4.5 for agentic reasoning. And right now you can use it for free. Runway has launched Aleph, which enables real-time text-driven video edits and generation. Adobe has added a harmonized beta to Photoshop, which looks really cool. Essentially, it allows you to blend two objects automatically, so you can drop an image on another image,

and if those don't look the same in terms of composition and lighting, this new harmonize AI makes them look and blend together really well. Luma has introduced a modified. with instructions for its dream machine video restyling. Mid Journey has updated its kind of video product with start frames and end frames. Idiogram has released a new character feature that delivers single photo, consistent character generation for users. Amazon has invested in Fables Showrunner, which is what they're calling the Netflix

of AI, where you can literally go on and just with a prompt, generate a bunch of AI, you know, shorts and series. Amazon has acquired B, a voice recording wearable. Speaking of Amazon, its CEO, Andy Jassy, confirmed that Amazon is considering ads and higher fees for Alexa Plus, their new version and smart version of Alexa that's powered by Anthropic. And notebook LM could, yeah, even more new updates, could be getting customized audio overview flows, which can help you put together.

and piece together in audio overview and an order you want, as well as connections to Gmail and Google Drive. Woo, a lot going on. Let's again very quickly recap the top AI news stories for this week in the AI News That Matter. So first, OpenAI has introduced study mode in check GPT to encourage better learning and to tackle academic cheating. Make sure to join us Wednesday as we give you the putting AI to work on Wednesday

version for business. Next, Google has rolled out Gemini 2.5 deep think mode with a major boost, but only for ultra subscribers. Google also updated a smarter AI mode with different features like Canvas and more. More Google as they updated their notebook L.M with video overviews, which I'm loving. Microsoft is testing out in experimental co-pilot mode in their edge browser. Open AI CEO Sam Altman has warned of some possible hiccups and delays as OpenAI is setting a lot of new features live here in the coming months. Apple is reportedly building an in-house AI team to rival ChatGPT.

And last but not least, Microsoft is reportedly pushing for a new negotiated term deal with OpenAI to extend their partnership or modify it. So much going on. I hope this was helpful, y'all. You don't have to spend two, three, four, five hours a day trying to keep up with AI. The best way to do it is number one, read our newsletter every day. It takes about five to seven minutes. It's free. Go sign up for it at your everyday AI.com.

And then the second way, if you just really care about the big developments, join us on Mondays. That's when we go over the AI news that matters. The no fluff, no BS, no PR, just straight take on here's what to. focus on and here's what it means for your business. I hope this was helpful. If so, don't be a jerk. Go share this with someone. If you're listening here on the live stream, appreciate it, please click that repose button. Share this with your network. Share this with your mom, your mom's sister, your mom's sister's babysitter, your mom's sister's, babysitters,

dog walker. Share it with everyone because we all need to learn AI to get ahead. We spend a lot of time putting the show together. So we'd appreciate you sharing it with your network. If you're listening on the podcast, appreciate that. As always, please make sure to click that follow button on Apple Podcasts and on Spotify and to leave us a rating. We really appreciate that. Then go to your EverydayaI.com. Sign up for the free daily newsletter. Thanks for tuning in.

Hope to see you back tomorrow and every day for more Everyday AI. Thanks y'all. Meet Firefly AI Assistant. Now live in Adobe Firefly, the Allman One Creative AI Studio. Just describe what you want to create in your own words and the assistant handles the rest, orchestrating multi-step workflows across Adobe. Creative Cloud apps, including Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome while the assistant accelerates execution.

Stand control with the ability to step in and refine at any time. See it today at firefly.adobie.com. 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 Everyday. A.a.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.