Everyday AI Podcast – An AI and ChatGPT Podcast - EP 605: Apple calls on Google for AI help, AI execs meet with Trump & more AI News That Matters

Episode Date: September 8, 2025

Although there were no huge LLM updates this week, the big AI companies were still making moves. ↳ Google avoided an AI breakup. ↳ The world's most prominent AI players met with U.S. President Trump. ↳ Apple had to get AI help from its biggest competitor. ↳ Anthropic suffered a $1.5 billion lawsuit loss. And that's barely the beginning of this week's AI news that mattered. Don't waste hours a day trying to keep up or miss an update that impacts your work. Tune in now!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 AI-Powered Job Board AnnouncementOpenAI ChatGPT Certification Launch DetailsAtlassian Acquires Browser Company for $610MAI Agentic Browsers Market Competition UpdateOpenAI Research on LLM Hallucinations ExplainedApple Integrates Google Gemini in SiriApple-Google AI Partnership for World KnowledgeOpenAI Broadcom Partnership for AI ChipsWhite House AI Execs Meeting with TrumpMajor AI Industry Data Center InvestmentsWarner Bros Sues Midjourney Over CopyrightAnthropic $1.5B Author Copyright SettlementGoogle Wins DOJ Search Monopoly Court CaseAI Image Generator Ideogram Launches StylesClaude Rolls Out Chat Memory for UsersTimestamps:00:00 OpenAI Launches AI Job Platform03:43 AI Skills Boost Salaries, New Platform07:59 Atlassian Acquires The Browser Company12:54 AI Models and Guesswork Issues15:08 AI Hallucinations vs. Uncertainty17:49 Apple-Google AI Integration Unveiled22:07 OpenAI Pursues Custom AI Chips28:59 "Warner Bros. Sues Midjourney"32:38 Anthropic Settles Copyright Lawsuit35:04 Google Avoids Chrome, Android Sale37:47 AI News: Rumors and ReleasesKeywords:OpenAI, AI news, Google, Gemini, Apple, Siri upgrade, large language models, AI hallucinations, AI job board, LinkedIn competitor, AI certification, ChatGPT, ChatGPT Enterprise, prompt engineering, Walmart AI training, Microsoft, AI salary premium, multimodal learning, NotebookLM, Perplexity Comet browser, AI agentic browser, Atlassian, The BrowseSend 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. Google got a big AI win in the courtroom while Anthropic took a big AI loss.

And Open AI may have solved hallucinations. Although we didn't get a new large language model update this week, what we did get is more AI news and headlines than the average person can handle. And if that feels like the position, that you're in, don't worry. That's what we do every single Monday here on Everyday AI, bringing you the AI news that matters. What's going on, y'all? My name's Jordan Wilson and welcome to Everyday AI. This is your daily live stream podcast and free daily newsletter, helping everyday business professionals like you and me, not just learn AI, but how we can leverage

all of this information and updates to grow our companies and careers. If that's what you're trying to do, it starts here with the unedited, unscripted livestripped. podcast, but if you want to take it to the next level, you make sure to go to our website. That's your cheat code at your everyday AI.com. So there you can sign up for the free daily newsletter where each day we recap that day's live stream podcast, but we also keep you up to date with everything else going on in the world of AI. So make sure you go do that.

Speaking of what's going on in the world of AI, there was a ton this week. So let's get straight into it. Podcast audience. If you didn't know, you can always join us live if you want to. 7.30 a.m. Chicago time. I'll just say that instead of Central Standard. Let's just call the Chicago time. So live stream audience, great to see you.

Branson and Michelle, big bogey face. Everyone else, Joe, thanks for tuning in. Brian, Jose, Dan, Dennis, Cecilia. Good to see you, y'all. Let's get straight into it. A ton going on in AI news this week. So don't spend hours a day trying to keep up. Just join us on Mondays.

So first, is Open AI going to try to compete with Microsoft with an AI powered job board? Well, that's because Microsoft obviously owns LinkedIn. And Open AI just announced that they're kind of creating a job board that might compete with LinkedIn. So OpenAI has announced that it's building an AI-centered jobs platform to match qualified candidates with employers, with a target launched by mid-2026, according to a company statement. So the move positions Open AI in a potential competition with Microsoft's LinkedIn, despite Microsoft being opens AI largest backer with an estimated $13 billion invested over the years. So OpenAI's head of applications said that the new AI powered jobs platform will include tracks for big companies,

local businesses and local governments to find AI talent. So OpenAI will expand also its Open AI Academy with tiered certifications in AI fluency from basics to prompt engineering using ChatGPT's study mode, which quizzes and coaches rather than just giving direct answers. So OpenAI aims to certified 10 million Americans by 2030 and organizations can embed these certifications into their training programs. Open AI is already working with Walmart on this initiative, one of the largest U.S. private employers. So OpenAI argues that workers with AI skills earn more on average, citing research from labor market analytics firm Lightcast, indicating salary premiums for roles requiring AI competencies.

So pretty interesting move here from Open AI with their announcement on this AI-centered jobs. And it's also said that users will be able to gain their certifications without ever leaving the chat GPT platform, which is pretty interesting. So it's not like there's a dedicated, you know, LMS study, you know, style module where you will study for certain certifications. So it's kind of interesting and ushering in a new era of AI learning where you will reportedly just be able to gain these certifications in chat GPT itself using chat. GPT, kind of meta, and I'm not talking about Facebook. So we'll see how or if this takes off. However, I do see a lot of people, you know, wanting to flock to this because there hasn't been a lot of certifications, I think, that have separated kind of AI talent.

So you have, you know, the big ones from Microsoft, Google, Amazon, etc. But for whatever reason, a lot of people, I think, pin those to more like marketing, right? You're like, oh, okay, well, I'm going to go get this, you know, the Google AI certification or the Microsoft or the Amazon, but it's only going to be really applicable for that company's software. Whereas Chad GPT is the most used AI chatbot in the world. And it's not even really close. So number one, it's kind of interesting to me that Open AI took this long to offer a certification. Number two, I think it's going to be pretty popular when it comes out just because of the sheer number of even Fortune 500 companies in the U.S. that are using ChatGBTGBT

EBT enterprise. But also, it is one of those certifications that I think won't be viewed as only being applicable for one type of work or using one kind of piece of software, right? Because I do think when people think of AI, they think of ChatGBTT, right? Even though it's not. not like that. And I don't think it should be like that. That's how a lot of people think. So I think that this open AI certification outside of just the job board aspect, I think people will view this especially very early on as a huge leg up and just as a sign of AI proficiency. So I'll say my judgments until I can get my hands kind of on this certification. And to see it's like, okay, is this really a good gauge on if someone has actually learned something? or is it just like, all right, you can use chat GPT long enough to just anyone

pass the certification. So we'll see. It's going to be interesting. I'd love to hear what our live stream audience thinks. So Michael from YouTube says, I want to use chat chbtee platform. LinkedIn is so saturated and it's hard to get noticed. I agree with that. 100%.

Joe is saying, I think Open AI is going to have to build in multimodal learning into their new learning platform. If they want to be competitive and successful, that's a great point, Joe. To compete with Notebook LM and their true learning capabilities there with the multi-modality, with the video overviews, the audio overviews,

I agree 100%. It leaves me wanting a lot. So I use Chat Chavit a lot when I'm out, you know, going for a walk or something like that. And I use the read aloud mode. And so many times I'm like, yeah, this would be way better if I could just use kind of true multi-modality in learning.

If I could have a video overview, if I could have an audio overview that I could talk with. So, you know, there's certain aspects of learning something in notebook LM versus learning something in OpenAIs chat GBT. I wish you could just kind of combine the best of both worlds. So we'll see if chat GBT slash open AI does improve, kind of that experience once they do roll this certification out.

But yeah, I would love to hear. Are you going to do it? Are you going to do this Open AI, chat GPT certification when it comes out. And do you think that this will rival LinkedIn? Personally, I don't think it will, but it's making pretty big news regardless. All right. Our next piece of AI news, a pretty big acquisition, maybe from a company or two you haven't

heard of. But Atlasin has said it will acquire New York based the browser company. for $610 million in cash, a move that pushes its products, Jira and Confluence, deeper into the AI-powered browsing for work. So according to reports from Reuters, shares of Atlas in the publicly traded company that acquired the browser company fell about 2% after the acquisition announcement signaling investor caution as the company expands into a kind of crowded category. So this deal adds the very popular Dia browser. So that is their kind of AI agenic browser from the browser company to the Atlason portfolio. So Dia was launched earlier this year and Atlasin reportedly plans to

position this as its go-to-work browser that pulls together tasks and different tools across the web and adds enterprise context. So the AI browser market is heating up. as startups and incumbents raised to add agent-like features that summarize pages and can also take actions for you. So I would say that Dia was one of the main leaders that wasn't a tech trillionaire company, aside from perplexity, which I would say is the leader with their comment browser, but other companies that are building on top of Chrome's, sorry, Google Chrome's Chromium, including Microsoft Edge, reportedly Open AI is going to be announcing an agenetic browser. You have the Brave company there, Leo browser.

So pretty interesting here that you have just an enterprise company acquiring this. And we're going to see what they're actually going to use it for. But it's a pretty interesting acquisition. So I did not have this on my acquisition, AI acquisition bingo card. If anyone would have acquired the browser company, I would have thought it would have been someone like Meta or AWS or even Open AI. So I did not really see a company like Atlison being the one that would acquire this company. All right. Did Open AI solve hallucinations?

Well, according to a new Open AI research paper, they say they at least kind of figured out why. large language models hallucinate to begin with. So according to OpenAI's new research paper, which we shared about in our newsletter last week, that's why you got to read it all. All right. So they're saying that large language models mainly hallucinate because of training and evaluation methods,

reward confident guessing over admitting uncertainty. So Open AI researchers argue that accuracy-based evaluations act like a constant test-taking and that encourages models to guess when they're unsure rather than just being like, yeah, you know what? I don't know. So essentially, when you can confidently guess something, even if a large language model is fairly confident, but not sure, confidently guessing something and getting it right has

kind of trained these models over time to just confidently guess, even when they're not 100% sure because that's what you're rewarded for, right? They're not rewarded in their testing and evaluation phase for saying, I'm not sure or I don't know. So Open AI says that, well, yeah, hallucinations could be fixed and they just need to redesign primary evaluation metrics. So they stop penalizing models that just say, hey, I don't know. And they start discouraging guessing, which they say would reduce these false statements or

hallucinations that are ultimately and very confidently presented as facts. So the paper notes that many models are effectively trained to treat questions as binary. So there's either correct or incorrect answers, and that is sometimes a mismatch with real world scenarios where uncertainty is common and stating limits is actually valuable. So open AI contrasted approaches by noting in a prior blog post that Anthropics Claude tends to express uncertainty more often, which can reduce errors but may increase refusal rates that limit usefulness. All right, this is one of those things. that a lot of people don't understand large language models, literally most of them in their

system prompt, are trained to be a helpful assistant. So what the AI researchers, I would say from the early 2020s through maybe 2023, 2024, probably push that a little too hard. And, you know, this causes large language models, even when they are not entirely sure to be helpful and to solve a user's query. Because saying like, I don't know is not very very much. very helpful. So a pretty promising piece of research out of Open AI. And researchers also talked about the kind of sometimes problematic with leaderboards, right? So if evaluation leaderboards continue to reward accuracy without penalizing guessing, Open AI warned in their paper that models will keep learning to guess, right? So they're kind of going through this, this awkward reinforcement learning

phase where guessing is actually rewarded because many times large language models will guess it correctly even when they are not entirely sure, where they would normally just be penalized if they're like saying, I don't know. So researchers recommended aligning benchmarks with real world needs by rewarding calibrated responses such as just saying, I don't know, or when models say that they aren't sure and then ask clarifying questions to get a more accurate answer. So, I don't know. Is this a good thing or a bad thing? Right?

If you're using a large language model, let's say, I don't know. Let's say you do 100 prompts a day, which I'd say for our audience might be kind of about standard. And let's say right now, you get zero times a large language model says, I don't know. But do you know what it hallucinates, right? You know, most models are usually less than a 1% hallucination rate. but that one may be big. So let's say you use a large language model 100 times a day. And you get one hallucination, but you're not sure it hallucinates because the model isn't like,

hey, I'm not 100% sure, but here's what I think. Instead, it just accurately responds. So over the course of maybe a month, you have 30 hallucinations on your hand that you probably didn't know her hallucinations because a model is trained to be helpful and to be confident in giving you an answer. So would you rather have that? If you're a power user of a large language model, and I'm not just talking about chat UVT,

would you rather have 30 hallucinations over the course of a month and not necessarily no? Or maybe out of those 100 prompts, would you rather have 10 times a day where a large language model is just like, not sure, or five times a day, right? Even five times a day, that's 150 times a week where you're hoping to get a helpful response from a large language model

and it's like, you're on your own, buddy. Which one's better? I don't know. I'd say probably the latter versus the formal because or former because I'd say AI hallucinations are a much bigger problem than people realize, right? Even because that one degree of accuracy and that one hallucination, maybe out of 100, which is still higher, a much higher hallucination rate, but I'm just saying it to round up to the nearest percentile. It's a much bigger problem than people realize because number one, when we talk about an agentic

future, that 1% compounds, right? But not only that, people look at large language models as the ultimate truth, where a lot of times if you're looking at the internet, well, hopefully, you don't view the internet in the same way. You have a certain degree of skepticism where you're like, all right, well, maybe not everything I read online is true, right? That famous quote from Abe Lincoln, you know, you can trust, what is it? Like, everything you read on the internet is true. Ableincoln, right? So I don't know which one's more helpful, but regardless, a pretty, interesting piece of research from Open AI. So make sure if you haven't read that already, we'll put it again in our newsletter. Two tech trillionaires teaming up. That's our next

piece of AI news this week. 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.adobie.com. But it's competitors, and this is not a good look for Apple. So, uh, Apple. Apple has, according to reports from Bloomberg, signed a formal agreement with Google to integrate a custom version of Google's Gemini into Apple's upcoming new version of Siri called World Knowledge Answers. Essentially, an Answers engine. It's going to be Apple's new web base Answers Engine slated for public launch in March 2026.

And this news moved markets. So Google's shares in their parent company, Alphabet, jumped about 9%. Yeah, they had some other good news that week, but instantly responded with a 9% jump, which is crazy. That's like a quarter billion dollars in market cap that Google added because of this news. And Apple's stock rose about 3%, signaling that investors expect Google to gain more upside from the deal than Apple. So the integration is designed. and here's kind of how it works, or reportedly how this new integration, this smarter AI Siri, internally called World Knowledge Answers, will work.

So the integration is designed so that Google's Gemini model, a version that's specifically made for this instance, runs inside of Apple's secure private cloud compute, keeping processing on Apple-controlled servers to preserve Apple's privacy stance. So series architecture will split users' tasks, and Apple's own foundational models will handle queries involving users' personal data, while Google's custom version of Gemini for this Apple integration will power the web summarization for general knowledge in real-time web answers, according to Bloomberg. So the world knowledge answers experience will start in Siri, reportedly as soon as spring 2020. 26, then expand to Safari, its browser, and Spotlight, offering AI-generated summaries from

web results with text, photos, videos, and also local points of interest. So Apple reportedly evaluated Anthropic as well as an alternative and even considered acquiring perplexity, but ultimately chose Google its chief competitor for their speed and reliability. So this move follows a series of years. of Apple's huge AI missteps. So they've had internal models missing benchmarks after reportedly spending millions of dollars a day, tried to develop a capable large language model, which they really haven't done.

And some reporting that said series internal team said that they just quote unquote couldn't get it to work properly. And multiple lawsuits that Apple is facing because of their false advertising over the last year and a half. half since the WWDC announcement of their Apple intelligence in 2024, essentially promoting and marketing all of these AI features that just didn't exist. They didn't work because Apple couldn't get them to work. So now to deal with this, Apple had to spend a lot of money in unreported some, but I'm guessing it's a ton and had to eat crow and ask Google their biggest competitor for health. That is a bad look. So if you want to know more about this, we did cover this one late

last week right after it broke, right after it broke. So make sure that you go check out episode 603 from the Everyday AI podcast where we did break that one down into more depth. All right. Let's get chippy with it, y'all. Our next piece of AI news, open AI might be starting to develop its own in-house AI chip and a deal with Broadcom. So according to reports, OpenAI is planning to start using a custom AI chip co-designed with Broadcom as soon as next year to address surging compute demand and also reduce reliance on Nvidia. So Broadcom's CEO referenced a new.

major customer committing to about $10 billion in orders, which reports essentially identified as Open AI. So this new joint chip is slated for only internal use at OpenAI and not for external sale, positioning chat GPT and other services for steadier compute access during peak loads. So this aligns OpenAI with Google, Amazon, and Meta, which have all built their own accelerators to control cost, performance, and supply risk for AI workloads. So the collaboration reportedly began last year with clearer timelines emerging now that shipments are targeted for next year pending production ramp success.

So Broadcom added a fourth marquee custom AI chip customer, them adding a fourth marquee custom AI chip customer signals accelerating demand for tailored silicon chips as AI model sizes and usage grow. And if successful, Open AI could shift pricing leverage away from Nvidia and speed future rollouts to end users by reducing compute bottlenecks. So there's been actually, Open AI has faced a lot of down, doubtage issues over the last few weeks. They had a pretty bad one last week.

And essentially, I don't think many people, even if you go back to when ChadGBT first came out a few years ago, I don't think you would foresee that Open AI would have this many users, 100 million weekly active users using their platform, which is just an astronomical number. But what this has led to is open AI hasn't been able to keep up with demand, right? So essentially, the larger AI industry earlier on in the generative AI boom, let's just say from 2020 to 2023, so many of the big companies relied solely on Nvidia. But now you've seen some success as an example with Google's internal TPU chips. Microsoft has started to develop their own chips as well, their NPU chips. So pretty bold move here, OpenAI, partnering with one of the biggest chip makers in the game in Broadcom to essentially develop their own chip that won't even be available to anyone else.

So we'll see how this pans out. And if ultimately this leads to less downtime for one of the most popular AI products on the planet. All right. Our next piece of AI news, just about everyone in AI was just at the White House. And some new reports have kind of suggested what this big meeting was about. So US President Donald Trump hosted a White House dinner with top AI leaders to spotlight pledges for huge AI investments. With a focus on fast tracking permits, in ensuring enough electricity for energy-hungry data centers that the biggest companies in the

world were buying. So as an example, Metas Mark Zuckerberg, who was at the meeting, said the company plans to invest at least $600 billion in the U.S. through 2028, including a $50 billion data center in Louisiana. Apple's Tim Cook, also reportedly at the meeting, said Apple had added $100,000. billion to its U.S. manufacturing plans, bringing its total pledge to $600 billion. And Trump signals those commitments could help Apple avoid some upcoming semiconductor tariffs. Interesting now that works and that these things are just being reported out there in the wild by, you know, a big company like Apple saying,

oh, we're going to invest all this money. And then Trump's saying, well, maybe you can avoid the tariffs then. Apparently, that's how politics works now, very out in the open. So Trump reportedly told executives that his administration is making it easier to secure electric capacity and also permits for data centers, essentially greenlighting those from major tech companies. And he claimed the U.S. is well ahead of China in AI. I would argue that that is not necessarily true. I would say China is very close, not very, you know, the U.S. is not very far ahead of China, especially if, we look at humanoids and robotics, China is light years ahead. And if that is kind of one of the endgame goals of AI, right, it being used in the real world, maybe the U.S. is winning the

sprint, but China may be winning the actual race. We'll see. But the guest list included Open AI's CEO Sam Altman, Google, Sundar Pichai, Microsoft's Saudi Nizdala, Bill Gates, and more marking a very rare high-level alignment around domestic AI expansion here in the U.S. Notably absent. It was like, oh, everyone from every single AI company except X-AIs Elon Musk's. So this was Donald Trump's former kind of right-hand man for a while. And they had a very public, very messy divorce. So it looks like, unfortunately, Elon wasn't invited to the table.

So the White House spotlighted a lot of plans that are bringing more demand to U.S. soil and the administration's AI agenda guided by White House AI advisor, David Sachs, centers on easing federal regulation, boosting research and development dollars around AI and expanding domestic energy production to keep the U.S. competitive. and companies are signaling more U.S. expansion in exchange for reducing their tariff exposure, with Trump indicating firms that build domestically could receive relief on some of the recently enacted import duties. Wow. Michael said this dinner scared me. Yeah, it is. It was kind of crazy to see every single AI leader around the table except Elon Musk. And I'm just imagining him watching this, you know, plotting what he's going to do next.

I don't know with robots and taxis, right? They're robotaxies and all these things. So we'll see how this ultimately shakes out. But pretty huge news, right? Essentially saying that the federal government is not only going to be expediting and greenlighting these data center expansions in ways that they normally wouldn't have, Right. So not just raiding all the yellow tape, but essentially waving the yellow tape, but also promising in exchange for breaking ground in U.S. soil on these data centers, essentially saying, yeah, we'll let these whole tariff things slide for you guys if you build more in the U.S. All right. Two more pieces of AI news. So Mid Journey has another big company coming for them. So Warner Brothers Discovery has filed a federal lawsuit in California,

accusing the AI image-generating platform mid-jurdy of using illegal copies of its work to train and sell their commercial AI image service, escalating the broader fight over AI and intellectual property. So Warner Brothers seeks up to $150,000 per infringed work, which would be like, I don't know. what's more than a trillion? That would be like a quadrillion dollars, $150,000 per piece of infringed work.

I mean, you can't prove how many there were, but there's got to be tens of millions or hundreds of millions, presumably pieces of potentially, you know, copyrighted works. And they cite. So Warner Brothers cited a lot of examples of their IP being used inside Mid Journey, such as Bugs Bunny, Superman, Batman, the Flash, Wonder Woman,

Scooby-Doo and the Powerpuff Girls. Yeah, don't forget about the Powerpuff Girls. So the complaint argues that Mid-Jurney creates consumer confusion by suggesting its large-scale copying and the resulting images are authorized for use when they obviously are not. Warner Brothers Discovery says it brought the case to protect its contents, creative partners, in financial investments, framing the dispute as a defense of Story. retailer rights. So the case follows a related case from other media giants, including some pretty big lawsuits by Disney and Universal and comes amid recent high dollar AI copyright settlements,

such as a $1.5 billion settlement from Anthropic. So the core legal question is whether training on copyrighted material and generating outputs that replicate recognizable characters constitutes as infringements, a decision that could set influential precedent. So yeah, this one is pretty big. And again, I'm no legal expert, but it seems kind of cut and dry. And here's the reason. I think with large language models, when you're looking at text outputs, you can argue that there's some gray area, right?

AI images not so much, right? If you pop out a verbatim copy of Bunk's Bunny or Superman or Batman and are using that in any way, it's a little harder to argue that it's not just blatant copyright theft. So some pretty solid cases, I would say, from both Warner Brothers and obviously, the previous lawsuit, which we talked about in depth on this show a couple of months ago from Universal in Disney. All right. One more piece of AI news.

I think I flipped past it in my slides. But speaking of big lawsuits, we can't skip over this one. Anthropic has agreed to pay at least $1.5 billion to authors to settle claims that it used copyrighted books without permission to train its AI systems. So the deal sets a minimum payout of roughly $3,000 per infringe book, giving the author industry a clear per work benchmark for compensating rights holders. So plaintiffs called the terms critical victories and warned that going to trial posed an enormous risk.

And anthropic faced potential exposure of more than a trillion dollars if it lost. So versus a $1.5 billion settlement, it seems that Anthropic was maybe okay with paying that $1.5 billion fine, because if they lost, if they took it all the way to trial and didn't settle, a court documents showed that this could have been a trillion dollar settlement. So the lawsuit focused on infropic's early data gathering practices, including downloading reportedly unauthorized copies of books to build training. data sets for its model. So this agreement is likely to shape future licensing negotiations across publishing, music and media by pricing past use and signaling the cost of unlicensed scraping. All right. Our last big piece of AI news before we get into the Roundup section, pretty good news for Google.

Wow. Mid Journey is facing some bad news in the court and Anthropic. bad news. Google got a piece of good news. So a U.S. district judge has declined to order a breakup of Google following a 2024 ruling that found the company illegally monopolized online search, a decision that could keep Google intact while the market shifts toward an AI-driven answer engine economy. So essentially what this boiled down to and what this allows, at least for now, it allows Google to keep Chrome and Android while imposing restrictions after that 2024 finding. That said Google illegally maintain a search monopoly.

So yeah, there were some reports that Google may have to sell off its Chrome browser and or its Android operating system. But sounds like they won't according to this new ruling from a U.S. district judge. This still could go higher. But right now the ruling also bars Google. from entering into exclusive distribution deals for Google search, Chrome, Google Assistant, or the Gemini app, allowing phone makers to preload or promote competing search engines, browser, and AI assistants. So the court ordered Google to also share certain search data with competitors,

and Google says it will appeal that, which could delay when any, which could delay when any data sharing requirements take effect. So the judge cited fast-changing competition. from artificial intelligence as a reason to avoid the most severe remedies, such as Google having to sell off Chrome, noting advancements by firms like OpenAI have reshaped the search market. So according to trial disclosures, Google paid more than $26 billion in 2021 to Apple and Mozilla and others for default placement deals for making their Google the default search engine in Apple's Safari in Mozilla's Firefox,

a practice that will now face tighter limits rather than just a strict ban. So analysts described the remedies as less heavy-handed than originally expected. And industry watchers say that Google Search is still projected to generate close to $200 billion this fiscal year with tens of billions of dollars flowing to distribution partners.

So this case, though, doesn't get Google completely out of the water as it's separate from an ongoing U.S. Department of Justice action alleging that Google holds a monopoly in ad tech, which was brought with 17 states. And Google says it will appeal that decision as well. So yeah, there's kind of multiple big suits in big cases that Google was facing, and this is one of them, a pretty big win, although there are some bullet points and some changes to their current setup there. All right. That was a ton.

There's more. We can't get to every single big AI piece of AI news, but we will give you the bullet point roundup. So this is kind of our newer segment at the very end of our AI news recaps on Mondays, essentially what's new and what's next. So some of these are rumors that are on pretty good. good terms and some of these are just new releases that didn't quite make our, you know, top 10 AI news stories of the week. So here we go. Get your pencils ready. There will be a test,

but you might want to know some of these things. Here we go. So notebook LM, T's flash cars and quizzes, literally the day after I said that they should be rolling these out. Open AI Aqua hired Alex, a YZBack startup known as the cursor for X code. Chad GPT's projects are now a available in a limited capacity to free users of the platform. Perplexity's comment browser is reportedly rolling out to Android and iOS phones, and it's made free to college students worldwide. Pretty big, great time to be a college student, get all this stuff for free. China's DeepSeek will reportedly release an agent to compete with OpenAI's agent by the end of this year.

Google's Nanobanana launched a dedicated Twitter account where you, can tweet at it and it will create images from text prompts. Very cool. So you don't even have to use the platform. You can just tweet at them and get an image from them, the best image model in the world. More notebook LM updates. They added new default audio overview style. So they only had the deep dive. But now there is the brief, the critique and the debate. So some new modes. Claude can now remember past chats for pro users. After the memory was announced last month, it just started to roll out. Open AI in some news announced that it will burn through $15 billion through 2029.

Claude is teasing a new feature that will create files inside of its chatbot. Chinese giant Alibaba released Quinn 3 Max preview, their newest large language model. It's benchmarking really well. AI image generator, ideogram, launched a new styles feature to get kind of consistent styles across generation. France's mistral closed a $2.3 billion funding round. Open AI is rolling out parental controls inside chat, GPT, over the next month. That one's interesting and will be fascinating to see how that's actually policed. Google has cut its V-O-3 API prices by up to 60 percent,

some models and open AI aqua hire sorry acquired stat sig to bolster its internal applications team. That was a lot. All right. I hope this was helpful. If so, please reshare this with your audience, but lets you a quick rundown of our main stories. So first, Open AI is planning to launch a new AI driven job platform as well as certifications. Appleson acquired the browser company for $610 million.

Open AI kind of came out with a great report investigating why large language models hallucinate the way that they do. Apple has tapped its chief rival Google to power a future version of their AI Siri. Anthropic has agreed to pay $1.5 billion in their author settlement case. A.I. is reportedly going to debut an in-house AI chip in partnership with Broadcom, which may ship next year. U.S. President Donald Trump was courting big tech for a massive USAI buildout. Warner Brothers is suing mid-jury over alleged copyright theft, and a U.S. judge has declined to force Google to break up. All right. That was a ton of.

of AI news. Don't spend hours trying to keep up every day. We do this almost every single Monday, bringing you the AI news that matters, cutting through the fluff, saving you time and telling you, here's what you need to focus on, here's the AI news that's going to

impact your company or career. Thank you for tuning in. If you haven't already, go to your everyday AI.com. We'll see you back tomorrow and every day for more, everyday AI. Thanks y'all. Meet Firefly AI Assistant. 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.adop.com. And that's a wrap for today's edition of Everyday A. 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.