EP 561: Meta’s new AI lab, OpenAI hit with huge copyright blow, AI’s impact on content publishers 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.

Meta is making all the AI headlines in both good and bad ways. Speaking of bad, OpenAI had kind of a bad week, both in its staffing fight and with Meta and in the courtroom. And the quality of large language models could go down in the future after a huge move from Cloudflare. That's kind of dorky, but I don't think anyone's actually talking about speaking of talking about it. That's what we're going to be doing today on Everyday AI, bringing you the AI news that matters. 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 from what's happening in AI, but how we can leverage all of this new information to grow our companies and our careers. So you can spend a ton of hours every day trying to keep up, or you can spend your Mondays with Everyday AI, where we cut through the fluff and everything that's happening, the whirlwind of updates in the AI world and we just cut it to you straight and bring you the AI news that matters. So before we get started, thank you number one for joining us. But number two, if you haven't already, please go to your everyday AI dot com. There you can sign up for our free daily newsletter. But also while you're there, you might as well just go take advantage of the cheat code that is our website, because there's more than 550 episodes that you can go watch, listen and read about for free from the world's premier AI experts, all sorted by category. So make sure you go do that. All right, let's get straight into it, y'. All. Here is the AI news that matters for the week of July 7th. So first, OpenAI CEO Sam Altman has responded again to Meta's AI talent raid from OpenAI and others. So according to reports, OpenAI CEO Sam Altman has responded to Meta's recent aggressive recruitment of AI talent, including from OpenAI, following a series of high profile departures from the company to Meta's new super intelligence team, which we're going to tell you about here in a couple of minutes. So according to reports, Altman addressed OpenAI researchers in a slack message, internally emphasizing the company's commitment to building artificial general intelligence and hinting at a compensation review for the entire research team. So Meta CEO Mark Zuckerberg announced the creation of their new super intelligence team with a lot of OpenAI researchers helping to launch this new team. So OpenAI's chief research officer Mark Chen compared Meta's hiring efforts to someone breaking into their home and stealing something, which we covered last week, highlighting the emotional impact of the departures. Altman though kind of on OpenAI's week off week to recharge during the US holiday of July 4, he downplayed the impact of Meta's hires, claiming that most of OpenAI's top talent stayed and suggesting Meta had to go quite far down their list to fill positions, although obviously a lot of people online are suggesting otherwise. He criticized Meta's approach as potentially leading to deep cultural problems and reiterating that OpenAI's mission driven culture sets it apart from competitors who may be more focused on short term gains. Altman stated that OpenAI is evaluating compensation to ensure fairness, not just for those targeted by Meta, and argue that OpenAI's stock offers more long term upside than Meta's stock. So employees who previously worked at Meta shared their support for OpenAI's culture, describing Meta as frequently shifting priorities and praising OpenAI's quote unquote quirky and weird environment as a driver of innovation. Yeah, also in the Slack messages reportedly Altman said missionaries will beat mercenaries and while Meta is doing is doing well in my opinion, led to very deep cultural problems. So this is one of those stories. It's not going to die down. We've covered it in its iterations throughout the last couple of weeks. But y' all OpenAI, I've said this until about December of last year, had such an insurmountable or a seemingly insurmountable lead in the large language model race until Google just woke up and chose straight up violence from December onward. And Meta throughout this has kind of lagged behind a little a bit. And I think after Meta's Llama 4, their latest large language model kind of came out to luck last year reviews it didn't benchmark as well as they had hoped. Again, Meta being though the leader or a leader in open source or open source ask right if we go line by line, it's not truly open source, but it is open source for most intent and purposes. Meta's just been very far behind. But we've seen recently Meta has been investing billions of dollars and apparently offering OpenAI researchers and other hundred million dollar salaries when you look at stock options. So this is one of those things. I think it's going to continue to grab headlines in the coming months. Just this, I mean, it's unprecedented. Now you have researchers who are commanding more in salary than the world's top athletes, right? So, you know, sometimes you see these headlines, you're like, oh, this, you know, I don't know, NBA player who's not even that good just signed an $80 million contract, right? The contracts that these open AI researchers are signing now, reportedly up to more than $100 million annually, are dwarfing the size of those kind of sometimes gaudy, you know, professional sports contracts. All right, speaking of Meta, something Meta was in the news for that was not necessarily good. But they've been, according to reports, developing and testing AI chatbots that send unsolicited, although friendly, follow up messages to users on Facebook, WhatsApp and Instagram, according to a report from Business Insider. So the initiative is called Project Omni and it aims to boost user engagement and retention for brands, having chatbots reach out to users without data being prompted. So these chatbots are built using Metis AI Studio platform, which launched in 2024 and enables users to create custom AI bots without technical expertise. So originally these bots were supposed to be more for when users wanted to talk to a brand. But now it looks like Meta may be offering businesses the capability to send messages, right? So not just respond to queries like so I'm sure many of you listening, you probably have maybe a chatbot on your website or something like that that responds to users messages. So think of it on the flip side, right? So now you may have on Meta's platforms, according to this report from Business Business Insider, codenamed Project Omni, the ability for your AI Persona, maybe that your company created in Metis AI Studio platform to proactively send messages to users. So Meta's motivation appears to be both engagement and profit. With company projections aiming for 2 billion to $3 billion in AI product revenue by 2025, up from 1 and up to $1.4 trillion by 2035. Yeah, talk about a huge leap in what they're expecting from AI product revenue. And one of the advantages that Meta has over almost everyone else is, yeah, it kind of owns those large social and messaging platforms in Facebook, Instagram and WhatsApp. So we'll see how this is received because other companies have tried this, such as Character AI and Replica, and they've tested kind of the similar proactive AI chats and they've been kind of controversial, including legal action in one more high profile case against Character AI. So I don't know my take on this. I get where Meta's coming from. Right. And how this can be a, an asset to businesses. Right. That you may want to reach out to companies. And when you look at it on the surface, okay, if you maybe like a Facebook page or you follow an Instagram account, is that much different from signing up for a company's newsletter, in which case they obviously might send you, you know, kind of unsolicited messages, although you're signing up for them. So I don't know. This is one of those weird instances and applications of AI that I guess we as a society are going to figure out. Is it going to be weird, right, if a company messages us or texts us unprompted but more personal. Right. So this isn't you're signing up for sms, you know, mass updates. This is presumably using the context of whatever your engagement has been in the past with the brand and offering it up to you in a way like an AI chatbot. So I'm even not sure how to personally feel about this, but this does look like one of those areas of potential revenue that Meta is going to explore. Right? We're talking about wherever you can see crossover between traditional business practices and using generative AI in large language models, we're going to see it. All right? We covered this one last week in our Hot Take Tuesday, right when it happened seems like kind of old news by now, but this happened during the past week. So Apple is in talks and reportedly negotiating with leading AI firms including Anthropic and OpenAI to overhaul their Siri platform, according to Bloomberg. So Apple is in advanced discussions with OpenAI and Anthropic to potentially use their large language models as the foundation for a new, more capable, smarter AI powered Siri. So Apple has requested custom versions of these large language models to run Apple's own private cloud compute infrastructure, aiming to maintain tighter privacy controls compared to using outside cloud providers. Talks remain ongoing with no final agreement reached yet. And according to reports, Anthropic was looking for a multibillion dollar annual fee to use their service, which was reportedly going to escalate in a huge way and spike up year over year. We don't have yet any reports on what OpenAI was potentially looking for from a revenue standpoint. And Apple's internal large language models have reportedly lagged behind just about everyone else's, even though they've reportedly been spending millions of dollars on LLM developments since 2023. Millions of dollars a day trying to develop their own internal AI in large language models. But apparently it hasn't worked very well because as we've seen, Apple has lagged behind ridiculous. Like it's not even close, right? You can't even put Apple in the same conversation as OpenAI, Microsoft, Google, anthropic and Meta. Right? They're not. Their AI capabilities are probably about the same as UN eyes, right? In terms of what we could do because they have nothing. They have some edge models that are okay, definitely not industry leading, that they're rolling out access to developers as well. But like I talked about in our, in our last Hot Take Tuesday episode. So I'll save most of my takes and just have you go, uh, listen to episode. What was that? Uh, episode 558 where I give you the, the full lowdown and my full thoughts on this. But uh, you know, this is just the latest in news surrounding Apple's failure to really create any in house AI at scale. And instead they've spent their time trying to release research papers that debunk AI's capabilities and they're facing a bunch of, a bunch of class action lawsuits for essentially marketing AI that didn't work. All right, moving on. Here's one of those. This might not, you might not feel this impacts you and you're like, oh, this is kind of a niche story. Jordan, why are we talking about Cloudflare blocking AI crawlers? But it's actually a really big frigging deal. So Cloudflare, which powers about 20% of the Internet, will now block AI bots and crawlers from accessing content by default. That is huge. A move that could remove one fifth of open web content from AI training data sets. Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on Genai. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you, you can partner with us too. Just like some of the biggest companies in the world do. Go to your everydayai.com partner to get in contact with our team. Or you can just Click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on gen. So this change means any new website signing up with Cloudflare will automatically prevent AI crawlers from scraping their content unless site owners specifically allow it so. In doing so, Cloudflare is also introducing some new granular privacy controls, letting individual site owners decide which AI bots can access their content and for what purposes, such as training, content generation or search. After verifying the identity and intent of the bots. The company is also announcing a new pay per crawl, a new initiative allowing AI companies to pay for access to web content from certain publishers. With publishers being able to set their own rates in AI services choosing whether to pay for access, this model could create a new revenue stream for publishers that have honestly been struggling since large language models have kind of replaced a huge chunk of their revenue while providing AI companies a more transparent and efficient way to access high quality, upto date content. So major publishers and industry groups including Adweek, buzzfeed, Fortune, Stack Overflow, the Atlantic, Time and Quora have signed on to support the initiative. So Cloudflare CEO Matthew Prince explained that this shift is aimed at giving publishers more control and compensation, arguing that if the Internet is going to survive the a of AI, this is a quote, we need to give publishers the control they deserve and build a new economic model that works for everyone. So the move responds to a long standing complaint that AI crawlers often ignore robots Txt files and other controls, leaving publishers with little recourse against unauthorized scraping. So this development could pressure AI companies to negotiate directly with content creators and pay for the data that powers their tools, potentially changing the economics of AI training and web publishing. So on the surface, I absolutely love this. This is a great move from Cloud Flare, which like we just talked about there powers 20% of the Internet. I don't know if this is going to work though, because like we just mentioned there at the end, a lot of these AI companies just kind of ignore, right? Kind of ignore what's in these robot txt files. So if you're not a geek like me, let me oversimplify this. There's essentially a set of directions or instructions that you can upload to your web server and it tells companies like Google and OpenAI and perplexity and Microsoft, so right, all these big tech companies both originally it was used for search engine and saying hey search engines, you know you can access these pages, don't access these pages they're private or they're whatever. Right. So it was instructions that helped companies better control their search presence and tell, you know, essentially earlier on search engines how in where they can access their site's content that they wanted to show up in search of. So prior to the big generative AI wave, it was kind of the same crawlers that would crawl and also use that data for search and for their AI training. Right. So what we've seen is kind of this separation now over the last year or so, where a lot of companies are trying to straight up block the AI companies from scraping their website and training on their data. But in some cases it's like, hey, if you want to show up in search of, you can't like have your cake and eat it too. Right. To oversimplify something. So I love this move from Cloudflare. Like I said, though, I don't know how well it will work in theory because we've seen just about all of the big companies now face lawsuits from maybe not following the directions set forth in these robot TXT files and sometimes, you know, saying, hey, AI companies, you can't scrape our content. And then the publishers are saying it's happening anyways. So that's one definitely to keep an eye on. All right, our next piece of AI news. The US Senate has voted 99 to 1 to strike down a proposed 10 year ban on state level AI regulation from President Trump's quote unquote, big beautiful bill that was recently passed. So the moratorium would have blocked states from enforcing laws against harmful AI uses such such as deepfakes and political manipulation for a decade. But bipartisan opposition prevailed and that failed. So here's what essentially would have happened and what ended up happening. So the Senate vote against this measure means that US States will keep their power to make and enforce their own AI related laws, whereas what was essentially tagged on as a pork provision to this big beautiful bill, right, which was President Trump's essentially domestic policy, it would have restricted states from enacting any sort of statewide regulation on AI for 10 years, which was absolutely nutty. And we talked about last week, there was essentially an amendment to this that there was some wiggle room and it knocked it down to five years and some individual provisions. But instead the US Senate just straight up shot this thing down. So this means that states will keep their powers to, at a state level to legislate and create laws around AI as they see fit. If the original moratorium had passed, states would have not been able to address problems like deep Fake pornography, AI driven scams, or election interference at the local level for 10 years. And it would only would have been federal oversight. So obviously the AI companies were wanting something like this to pass because this mean it would be easier for them to operate. Right. They wouldn't necessarily have to adhere to different rules and regulations across the U.S. right. It's already difficult enough, I think, for AI companies to roll out consistent products and services worldwide. Specifically. Right. The EU has some very tight privacy laws and things that you just can't do. So it's kind of affected the global rollout of different AI tools, which, you know, it's, it, it actually creates this, this competitive advantage. Right. For those countries that don't legislate AI. And at a federal level, the us, which is obviously where I'm from, it's been the Wild West. There's really no laws, at least actual federal legislation regarding AI where a lot of other countries do have this. So this was really, I think, number one, it would have probably led to a lot more economic progress for the US because they could have just said, hey, we're the Wild west, we can do whatever we want to. But now we're going to see what's going to happen at the state level in terms of legislating AI. And like I said last week, really just keep your eye on California. Yeah. Other states matter, you know, New York, Texas, Illinois. Right. Not just, you know, certain states with large, you know, metropolitans, but also looking at states, you know, even something like Tennessee, right, where XAI has their data center. Right. Because you could look at different levels of taxation, different rules around job creation, about job displacement regarding AI. So it will be interesting to see how this unravels. But y', all, I said this, a lot of people online and news organizations when this first came out, they don't know how politics work, right. I was a former journalist, I covered politics at the state level and a little bit at the national level as well. And I'm like, there is zero chance that this actually becomes a law yet. Everyone, just because it was an amendment, you know, just because it was essentially a pork provision, right. In this domestic policy, everyone thought it was a done deal. And I'm like, no, that's not how this works. Right. So that's why y' all don't just believe even there are some pretty big accounts online and some biggie, some pretty big media outlets that just kind of ran with this as if it was a done deal like a month ago when this was introduced. Of course, it's not right. These first, first versions of these all encompassing kind of domestic policy and budgetary bills, they essentially have hundreds of unrelated, you know, what we call pork provisions. Right? And a lot of them get stripped out and all it is, it's, it's leeway and leverage to get certain people, you know, that the minority and majority whips can, you know, get people to vote for it or vote against it. Right. According to these provisions that end up getting included or excluded from the final bill. So, you know, a lot of people were losing their noodle and being like, oh, my gosh, you know, states aren't going to be able to control what's happening. And I'm like, y', all, that's, that's dumb. It's not going to pass. And obviously it didn't. All right, our next piece of AI news. Bad news for OpenAI. So a federal judge has ordered OpenAI to grant the New York Times, New York Daily News, and the center for Investigative Reporting access to its user logs, including deleted records, as part of a very high profile copyright infringement lawsuit, with the New York Times being kind of the lead there against OpenAI. So this sweeping order allows the plaintiffs to examine whether OpenAI's language models were trained on copyrighted news content and if ChatGPT is capable of reproducing or plagiarizing that material. So the New York Times argues that users may delete their ChatGPT histories after bypassing paywalls, making full access to logs essential for uncovering potential copyright violations. Right. So the New York Times is saying, hey, people were using CHAT GBT to read all of our articles that were behind a pay wall. So we're losing money and people are using CHAT GBT to access all this. OpenAI says, yeah, that's not how it works. You know, there. OpenAI essentially says so much of these, so many of these things that New York Times has reported on over the years, and the New York Times alleged it was millions of pieces of content. They're like, OpenAI is like, yeah, this has just become public records in so many cases. Right. Just because if you report on something doesn't mean it's exclusive to the New York Times. Right? Other companies have reported on it. Other people report on what the New York Times has reported on. As a former journalist, that happens all the time. So OpenAI is saying, yeah, we never will reproduce anything verbatim, Right? That is exclusive to the New York Times behind a paywall. The New York Times is obviously saying, absolutely, yes, you do. And for now, a federal judge is granting what I think is kind of a shocking allowance for these companies to move forward looking at private and deleted chats from OpenAI. So OpenAI has objected to the ruling, saying it undermines its long standing privacy norms, and has vowed to continue fighting the order as both sides negotiate how the data surge will proceed. So the order covers logs from free ChatGPT and API users, but specifically exempts enterprise and educational accounts. So the plaintiffs also hope to use the logs to show that ChatGPT's ability to summarize articles reduces direct traffic to news sites, leading to lost ad revenue and market dilution for publishers. OpenAI has stated it will, quote, unquote, keep fighting the order and is actively appealing the judge's decision. So this stems from the original lawsuit. This thing's been going on for a long time, going back to December of 2023. So what, that, that's more than 18 months now that this has been going through the courts. But a pretty big, a pretty big development here. But again, this isn't the end, right? I believe that OpenAI will appeal this and it's not going to be a done deal, but obviously we'll be covering this. The former journalist and me, you know, I'm, I'm not picking and choosing sides here, but you also have to think, number one, I think this request and the fact that it was granted is absolutely bonkers. But number two, you also have to say, hey, OpenAI and all the other AI tech companies, they did just essentially scrape the open web to train their large language model. So I definitely understand where both sides are coming from. And if you want independent and mainstream journalism to continue, you have to make sure that these companies are able to monetize somehow. I've always said all along the only. And go back, I've been saying this since 2023. I've been saying the only way for this to work in the long run, right? We'll see how this new cloudflare option plays out. But I said you can, number one, companies can sue. That's number one. Number two, you can go into content agreements, right? A lot of you've seen these with news organizations, sites like Reddit, you know, signing these, you know, million plus or tens of millions of dollars agreements each year. So number one, you can sue. Number two, you can sign a huge agreement. Or number three, you can probably just go out of business. So, you know, we'll see. And this is the key case here. Obviously there's been some juicy, you know, AI kind of cases like we talked about the The Disney versus Mid journey one. But this is the biggest one because text is a lot easier or sorry, it's, it's a lot more convoluted to prove right than if you go into Mid journey and you say, you know, draw me a picture of a superhero and it spits out an almost verbatim, you know, something from Disney or you know, Universal or Marvel or whatever. It's a lot more difficult when you look at the, the entirety of the Internet and it's put into this melting pot stirred and something new and unique comes out. So that one's going to be an interesting one to to follow. Which leads to our next news story, which is very relevant to the last one. So new data from similar web shows that 37 out of the top 50 US news websites have suffered significant year over year traffic declines since Google introduced its AI overview search feature in May 2024. So now that has even though it was a gradual rollout, that feature has now been out for more than a year. So Forbes and Huffington Post experience the steepest drops, each losing about 40% of their website traffic, while Daily Mail fell 32% and CNN dropped 28%. So according to this new similar web study. And if you don't know similar web, they are one of the most reputable, you know, companies or services just that studies the Internet and Internet usage at the whole. So according to this similar web study, the rollout of AI overviews had led to to automated summaries appearing at the top of Google's search results, often answering user questions directly and reducing the need to click through to the news sites. So the News Media alliance, which represents over 2,200 news organizations, warns that these changes are making it harder for publishers to invest in quality journalism as their audience shrinks. So the average click through rate for the top organic search results on queries triggered AI overviews dropped from 7.3% in March 2024 to just 2.6% a year later, according to multiple studies. So yeah, that's not good. So all of these publishers that really thrive on clicks, right? Because then when you click on, you know, I don't know, let me pull, what were the examples here? Forbes. Right. So when you go to a Forbes web article, they make money because there's ads on their pages or you may sign up for their email list or one of their services. So obviously the fewer people that end up on the Forbes.com website, Forbes loses a ton of money. So then that impacts their ability to create new and unique information to put out on its website. So it is weird. Just the very nature of this relationship of large language models scraping the entirety of the Internet, it actually hurts the world's ability to create new and unique information. And again, this is another one of those things I've been talking about since 2023, since I started everyday AI, that we are going to see this cyclical slop, right? Because so much of new information that has been Posted since like 2021 has just been GPT spun or syntax created from old information. So the percentage of new high quality, unique information stats and studies that is being put out there is at a much lower rate. So we have to, you know, not like worry about the future of large language models in their utility and everyday business context. But you have to look at that, right? And what happens if something like this cloudflare plan comes to fruition and you know, these AI companies maybe just aren't going to pay up, right? So hopefully, you know, there can be this harmonious, you know, joint effort where, you know, the big AI companies are somehow paying or you know, something on the back end, you know, all of these online publishers that are now just losing clicks and losing, you know, losing traffic. And that's like I said, it's really restricting their ability to create new accurate information. All right, our two last stories. So according to reports, OpenAI is going into consulting, or at least doubling down on their consulting efforts. So according to the information, OpenAI is intensifying its enterprise consulting efforts, now charging at least $10 million per client. So the company's engineers, known internally as forward Deployed Engineers, work directly with third party organizations to tailor models like GPT4O to specific business data and develop custom AI applications including chatbots. So this move places OpenAI in direct competition with established consulting firms like Accenture, Deloitte as well as companies like Palantir. So services extend beyond model customization to include data labeling where experts review and correct AI generated responses, ensuring higher accuracy for enterprise solutions. So OpenAI is reportedly considering outsourcing some of this data labeling to work to specialists like Snorkel AI and Surge AI after companies like OpenAI and Google broke ties with Scale AI, which was Aqua acquired or Aqua hired by Meta. So high profile clients already include the US Department of Defense and Southeast Asian tech company Grab. So these are companies that are reportedly already using this OpenAI consulting services, signaling significant trust and adoption among major organizations. So this shift could make AI implementation more accessible to large enterprises, but also raises the bar for entry due to the high service costs. And I Don't know, you could pay $10 million or you could reach out to us. We'd do it for way less than that. All right, all right, our last piece of AI news. We talked about this at the top of the show, but let's get into all of the nitty gritty details because Meta essentially just released their dream team for the newly created and announced. Here's the official name of Meta's new AI organization. It is msl, the Meta Super Intelligence Labs. So Meta has announced MSL their new organizational structure uniting its foundation, product and fair teams with a dedicated lab for next generation AI models, according to a memo obtained by the Wire. So the move marks a major escalation in the AI talent wars. Like we said, Meta essentially acquired a 49% stake in Scale AI, which all the other big tech companies used before, and also hired its CEO Alexander Wang as the chief AI officer to lead MSL. Nate Friedman, the former GitHub CEO, will co lead the new lab alongside Wang. So yeah, Nate Friedman and Alexander Wang, two of the biggest name in AIs now kind of the one, two punch there. Also Daniel Gross, the former CEO and co founder of Safe Super Intelligence Inc. Is also joining Meta's new AI superintelligence lab. So my gosh, we haven't even got into all the, you know, the researchers, just the three of those, Daniel Gross, Nat Friedman and Alexander Wang. I mean that's like the who's who in the AI world. So these new hires and also include top researchers and Engineers previously at OpenAI, Google, DeepMind, Anthropic and other leading AI firms underscoring Meadow's aggressive recruitment strategy. All right, let me name off some of the notable additions and I'm sorry, I'm probably going to get some of these pronunciations wrong but you know, this thing's unscripted, unedited. So here we go. Some of the notable additions that Meta has added included Trapeet Banzel, a pioneering in reinforcement learning and a leader in the O series model development at OpenAI. Suchow Bai, the co creator of GPT4O voice mode, Huynh Chang, the inventor of Mask It. Sorry, Mask it at Google Research, Jack Ray, the pre training tech lead at gemini for Google DeepMind and several others with direct experience building state of the art AI models. Yeah, so what Meta did here is they essentially and like we've been talking about for the last couple of weeks, they've been quietly poaching some of the top talent from around the globe, specifically going to companies like Apple, OpenAI Google, DeepMind and Anthropic and others, and taking some of the top researchers. So Meta has essentially kind of been lagging behind. And like we talked about, they are taking more of an open approach to AI, which puts them kind of at a technical disadvantage. But in the long run, it may put them at an advantage. Right, because they could commoditize their competitors, proprietary paid models, if this gamble really pays off. And what's interesting here is Meta, unlike companies like OpenAI and Anthropic, they've been making billions of dollars for more than a decade from their Facebook and Instagram and WhatsApp and just their advertising revenue as well. So Meta has straight up cash to play with when it comes to hiring some of the top researchers, where, you know, some of these other more startup companies, yeah, they have money to play with and they have stock options. But y', all, I mean, Meta is just Scrooge McDucking this thing. Like, you know, Zuck is literally sitting on a pile of cash and he's not playing around anymore. Right. It is kind of interesting to see this dynamic play out in real time. And we've been covering it obviously week by week here. But, you know, if you weren't taking Meta's approach to large language models seriously, maybe a couple of months ago, which I think, if I'm being honest, not a lot of companies were. Right. I think a lot of companies were looking at Meta almost as like an option be. Right? Yes. They're still investing in it and it's like, hey, what happens if, you know, OpenAI 20x's their prices? Or what happens if, you know, Google's Gemini, you know, we can no longer use it for whatever reason. So I think a lot of companies were looking at models like Meta's Llama almost as a backup or plan B in case their proprietary model of choice, for whatever reason was no longer applicable for their business use cases. But now I think Meta is trying to change that dialogue. They've got money to throw. They've got the compute. Right. They've got just about more than more GPUs than just about anyone in the world. And now CEO Mark Zuckerberg is saying, we're not playing around, we're throwing hundreds of millions and actually billions of dollars at forming the AI dream team. And now the newly minted Meta Super Intelligence Labs. All right, now let's get to our kind of rumors and rants. You know, what's new and what's next? So just bullet points of some new updates that maybe didn't make the big AI news cut. So Perplexity is rolling out a local search version with maps on its iOS app. So kind of trying to compete with local Google search. Anthropic has been testing a new model on different platforms, potentially an update dated version of Claude 4. Google rolled out v3 access to more Gemini Pro users worldwide, including the EU. The EU had kind of been waiting for that rollout. Perplexity also launched a new 200amonth max plan with increased limits and early access to their comment browser, which looks pretty impressive. We've seen some new hints of Grok 4 but still no official launch. I would be expecting a launch I don't know now, you can say any day now. But Grok 3.5 got delayed to eternity. You know, we were told, you know, Elon Musk said that Grok 4 would come by July 4th. It didn't. But we've seen kind of kind of some hints of Grok 4 in different companies, source codes and benchmarks. So that could be coming any day now. Notebook LM is testing a new Flashcard mode in video overviews as well. And Google Gems are starting to roll out across across workplace apps, which is pretty cool. You still have to build the gems inside Google Gemini, but then you can use the gems in things like Gmail, Google Docs and Google Sheets. All right, that is a wrap y'. All. Let me very quickly recap the AI news that matters. So OpenAI CEO Sam Altman has responded again to Meta's AI talent raid. Then we talked about Meta testing proactive AI chatbots that message that message users unprompted. Apple is in talks to partner with Anthropic or OpenAI to essentially take over Siri since they couldn't get it right with the newer smarter AI powered Siri. Cloudflare has blocked and has introduced the capability for publishers to block AI crawlers by default and has launched a paper crawl for web content. The US Senate rejected a 10 year AI regulation not allowing states to make any AI laws by an overwhelming 99 to 1 vote. So states will be able to legislate AI at the state level. The New York Times first open AI case has an update as a judge has ordered OpenAI to hand over user logs in the New York Times copyright lawsuits, although OpenAI says they're going to continue to fight it. A new study from similar web shows that major US news sites are are losing up to 40% of traffic after Google's AIO reviews launch. OpenAI has expanded its enterprise AI consulting team and are apparently charging a $10 million minimum to work with them. And Meta has officially kind of launched and unveiled its roster for its new AI Dream Team, Meta Super Intelligence Labs, or msl. That's a wrap, y'. All. I hope this was helpful for the AI news that matters for July 7th. If this was helpful, tell someone about it. If you're listening on the podcast, please subscribe and follow the show. Reach out to me. I always leave my information in the show notes. Connect with me on LinkedIn. Tell me what you want to see more or less of. If this was helpful, please consider reposting this on social media as well. And then go to your Everyday AI dot com. Sign up for that free daily newsletter and then make sure to join us 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 your everydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.