Everyday AI Podcast – An AI and ChatGPT Podcast - Ep 798: SpaceX goes public, OpenAI declares pricing war, Microsoft stops Claude Fable use and more AI news

Starting point is 00:00:00 This is the Everyday AI Show, the Everyday Podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life. This week was a wild one in AI news. I mean, we have the world's first trillionaire thanks to AI. We also have the world's most powerful model, but one tech Titan refuses to use it because of their data. and we have another big AI company that's not only going public, but they're reportedly going to start an AI pricing war to make AI cheaper. All the biggest names in AI made some big moves this week.

Starting point is 00:00:47 And if you took your eyes off the road for more than a few minutes, you probably missed some pretty impactful news that will move the needle for your company or your career. But I get it. You can't spend every hour of every day trying to keep up. That's what I do for you. And that's what we do here with our Monday show called the AI News That Matter. So if you're new here, welcome to Everyday AI. My name's Jordan Wilson and we do this thing every day.

Starting point is 00:01:16 And it's for you. This is your daily live stream podcast and free daily newsletter, helping everyday business leaders like you and me, not just keep up with everything that's happening. But I help you separate the real from the fake, the fluff from the meaningful. I tell you what matters. You use that information to grow your company and your career. So if you haven't already, please make sure to subscribe to the podcast. If you are listening on the podcast side, and then make sure you go to our website at your

Starting point is 00:01:40 everyday AI.com. Sign up for the free daily newsletter. We're going to be recapping not only the big AI news stories we're going over today, but everything else that's happening in the AI world that's going to impact you. So without any further ado, because we have a lot to get into and one other, you know, FYI. So normally I love doing this show live on Monday mornings. I am having to record this one a little bit early just because I'm doing some traveling. So if maybe if there's some big news that broke late over the weekend, might not have it in this episode, but make sure you do check out our newsletter for that. All right. So we're actually going to start with a breaking update in the Anthropic versus

Starting point is 00:02:22 White House Fable Development. Like I said, I'm traveling this week. And we actually, actually didn't get the story in the live stream version of today's show. So this is kind of a podcast exclusive that will kick off the show with. So the White House's urgent action to halt Anthropics' latest AI model highlights the growing concerns about national security and the rapid pace of AI development. So the Trump administration imposed sweeping export controls on Anthropics' new AI model Fable 5 just days after its public release. citing national security risks, according to Politico. So senior officials, including Treasury Secretary Scott Besant and White House Cyber

Starting point is 00:03:07 Director Sean Cairncross, held multiple tense calls with Anthropic CEO Dario Amati, according to reports, urging the company to voluntarily remove the new Fable 5 model that they had just released. Well, why? Well, according to reports, that's because Amazon, CEO Andy Javis, whose company is obviously a big investor in Anthropic raised concerns to the White House about the new Fable 5 and Mythos model's ability to bypass safety guardrails after being asked for feedback.

Starting point is 00:03:42 So the administration felt Anthropic was not taking the security threat seriously enough, and officials were frustrated by Emote's reluctance to immediately pull the model. Anthropic argued its safeguards were strong enough and that no universal jailbreak had been found, but the government officials remain unconvinced, eventually setting a 90-minute deadline for removal. Again, this is all according to reports. So the export controls abruptly disabled Fable 5 and Mythos 5 for all customers worldwide,

Starting point is 00:04:16 banning their use also by foreign nationals and forcing Anthropic to comply. Anthropic criticized the government's action as disproportionate. and called for a more transparent and technically grounded regulatory process. So the White House's intervention follows months of debate about how to regulate advanced AI models with previous productive meetings between Emote and senior White House officials leading to an executive order on voluntary model submission. Anthropic issued a statement on Friday but has not yet publicly responded to the new accusations that came out over the weekend, at least as of this recording,

Starting point is 00:04:55 now Sunday afternoon. So here's what Anthropic said Friday afternoon. So they said the U.S. government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside of the United States, including foreign national anthropic employees. The net effect of this order is we must abruptly disable Fable 5 and Mythos 5 for all of our customers to ensure compliance. Access to all other anthropic models will not be affected. We receive this directive from the government today, this is Friday, at 5.21 p.m. Eastern. The letter did not provide specific details of its national security concern.

Starting point is 00:05:45 Our understanding is that the government believes it has become aware of a method of bypassing or jailbreaking Fable 5. We reviewed a demonstration of this specific technique being used to identify a small number of previously known minor vulnerabilities. These vulnerabilities all appear relatively simple, and we have found that other publicly available models are able to discover them as well without requiring a bypass. So there will surely be an update to this story on Monday, and we'll be sure to include this in our newsletter so you can make the right decisions for your. your company. So this obviously has both Anthropic Insiders and connected investors kind of shook, right? So previously between the Department of Defense's decision to label Anthropic a supply chain

Starting point is 00:06:40 risk and to prohibit the Department of Defense and certain branches of the federal government from using any anthropic models. So this kind of back and forth between Anthropic and the U.S. government has been going on for a while now, but this is pretty bad timing for Anthropics. So there was obviously a lot of good and bad reaction to their new Fable 5 in Mythos 5 release. And we will be covering that in a later kind of segment here on this very show here in a couple of minutes. So there's a lot of good hype and bad hype behind it.

Starting point is 00:07:13 But this is obviously a pretty bad look for Anthropic. We'll see who was ultimately in the wrong. We haven't had kind of the ability to hear properly from Anthropic after they've gone back and forth with the government on this. So I'm sure we'll hear more on their side of the story. So we'll give them the proper time before we dig into this too much. But regardless, you know, the kind of rollout of Fable 5 and Mythos 5 was kind of supposed to be this last big positive splash before Anthropic ultimately decided to go public as they did kind of file their S1 paper. about two weeks ago now. So a pretty big move. And like I said, we'll have more on this in the newsletter as it is a developing story. But let's get to the rest of our AI news. My gosh, what are we

Starting point is 00:08:02 can AI when Apple's big AI announcements did even make our little teaser intro. But Apple's WWDC showcase a series of big AI advancements aim at closing the gap between Apple and their competitors. So, man, like in a. AI time. This, like I know this happened Tuesday. It seems like it happened a month ago. But we got to talk about it. Apple introduce a major update to Siri, powered by Gemini, AI under the hood, right? So they're making their new AI assistant more capable, conversational, and visually intelligent, now available as both a standalone app as well as integrated across Apple's ecosystem. So it's not out yet, right? So as is, most of the case with Apple's AI announcements over the past few years. It's in beta now coming fall. So we'll see when the new smarter Siri and all of these Apple intelligence features roll out both to the iPhone and the Mac OS and iPad, I guess. But regardless, a lot to take in here. So the company has obviously emphasized and continued to preach this at WWDC,

Starting point is 00:09:15 their privacy first approach to AI, stating that data used for AI. request will not be stored and can be externally verified for privacy compliance. So Apple intelligence is the company's integrated AI system and it's being expanded with some new features like cross-app context awareness, one-tap password updating, and smarter tab management in Safari. Messages and phone apps are getting AI-powered reply suggestions and the ability to pull contacts from other apps during calls. That's what I'm looking forward to, streamlining communication for all users. So AI-driven improvements are coming to search across iOS, iPadOS, and MacOS,

Starting point is 00:09:57 promising faster and more accurate results in spotlight, mail, and photos. So the Photos app debuts some new generative AI features, including Reframe, which I personally thought was the only real novel use of AI that I personally saw out of Apple's WWDC announcement. That's essentially where you can kind of retake a photo that you already took. from a different angle, a different vantage point. That seems pretty cool. They are also rolling out some extend features, cleanup,

Starting point is 00:10:23 nothing that we haven't seen in a dozen or so other, you know, kind of apps before. But the shortcuts app, that's what I might actually use. It will now let users create automations using natural language prompts, making the tool more accessible for those without technical backgrounds. AI moves too fast to follow, but you're expected to keep up. Otherwise, your career or company might lag behind while AI native competitors leap ahead. But you don't have 10 hours a day to understand it all.

Starting point is 00:10:59 That's what I do for you. But after 700 plus episodes of everyday AI, the most common questions I get is, where do I start? That's why we created the Start Here series, an ongoing podcast series of more than a dozen episodes you can listen to in order. It covers the AI basics for beginners and sharpens the skills of AI champions pushing their companies forward. In the ongoing series, we explain complex trends in simple language that you can turn into action. There's three ways to jump in. Number one, go scroll back to the first one in episode 691. Number two, tap the link in your show notes at any time for the Start Here series.

Starting point is 00:11:38 Or you can just go to start here series.com, which also gives you free access to our inner circle community where you can connect with other business leaders doing the same. The start here series will slow down the pace of AI so you can get ahead. So obviously the highlight here is Apple rolling out a dedicated Siri AI app. So not only is the company saying Siri will actually be smart this time, for real, for real, but all of the conversations that you have with Siri, whether they're on your phone, on your computer will be synced in a dedicated Siri app. So we'll see how all of this goes.

Starting point is 00:12:23 Obviously, Apple has gotten into a little bit of trouble for promising things in terms of Apple intelligence in the smarter Siri. In 2024, they didn't deliver and then they had to pay a $250 million lawsuit. So I did go over this in much more depth in episode 794. So make sure you go check that out if you want to, you know, hear more on any of this. But the big picture here is Apple says the new Siri AI will understand personal context across messages, emails, apps, and what is shown on screen. And then that, you know, sinks across the Siri app if you're using ICloud.

Starting point is 00:13:01 So when might you get this? Well, the broader Apple intelligence updates are expected this fall and the most important Siri AI. So the final thing that they demoed is likely arriving later this year. So presumably that would be late fall, early winter, in English first. So yeah, not rolling out everywhere. We'll see how many times that timeline gets pushed back. There is a beta version out now, although even if you got on immediately like I did,

Starting point is 00:13:33 you're probably still on the wait list. It seems like they're letting very few people in on the beta. But then the kind of the rollout then after that, so some of the other Apple intelligence features coming out in the fall. and then the broader Siri AI rollout later this year. All right. Is chat dead? Well, maybe.

Starting point is 00:13:53 And there's probably something else bigger coming, which we've been talking about for a long time. But I think this is an important news story to talk about just because of the shocking headline here, that chat is dead. So Open AI is preparing to radically transform chat GBT into a super app in the coming weeks. according to a report from the Financial Times. So the company aims to shift ChadGPT from a simple app chat interface to a gateway for advanced agents, coding tools, and productivity apps, focusing on features that could drive more financial revenue. So a senior, and here is the reason why I'm thinking this is a top AI news story here,

Starting point is 00:14:34 so according to the Financial Times, a senior OpenAI staff member told the Financial Times that quote unquote, chat is dead, signaling the major change in direction for the company's core product. So OpenAI's new model will reportedly encourage users to rely on its apps, coding tools, and image generation features, moving away from the traditional prompt-based chat experience. So the company is also planning to introduce enterprise-oriented pay schemes, including per token pricing, to attract more business clients and boost revenue as it competes with Anthropic in the race toward a public offering. Anthropic, OpenAI's main competitor, has recently pulled ahead in some people's view by targeting

Starting point is 00:15:17 enterprise customers with its clawed AI platform rather than focusing on the general public, which is what Open AI's been doing, as they did officially just hit one billion weekly active users. So we've known now for a couple of months that this has been, you know, Chad GPT's priority in terms of its plan, but pretty shocking headline nonetheless to see that an open AI seems your staff member, obviously on background or off the record saying chat is dead, which is interesting, right? And it makes me think how important it is to start changing your mindset, right? So, you know, taking a pause here from the news piece of it, you know, even for myself, I'm using just the chat products way less, right? I'm using chat GPT online,

Starting point is 00:16:05 less. I'm using claw.com.A.I less, right? I'm moving more and more into the super app territory, right? It's no surprise. You know, you guys all know what I'm, you know, doing these shows. I usually have codex and ClaudeCode running right now. I actually have Codex working now for more than a day going and testing out things in the new Claude Fable model. So, you know, let me know if like, do you guys think we should focus a little bit more on super apps? Maybe I'll do a dedicated show on super apps going over what they are, why they matter, how to use them, what constitutes a super app, you know, just let me know, As always, I always have a number, y'all.

Starting point is 00:16:43 So if you are listening on Spotify, you know, say super app, yes, super app no. Same thing, live streamed audience. Let me know. Personally, I think it's smart to start looking there. So if you are using these desktop autonomous agents, right, codex, clawed desktop, cursor, anti-gravity 2.0 desktop, you know, that is kind of, I think, the next wave. All right. Let's keep it going.

Starting point is 00:17:08 Speaking of Open AI, another piece of big news, they have officially filed for their IPO. So OpenAI has confidentially submitted paperwork to go public on the U.S. stock market. A move expected to result in one of the highest valued IPOs in history. So OpenAI's confidential S1 filing with the SEC signals a potential IPO valuing the company at nearly a trillion dollars or maybe more, but at least as of now, we're seeing a lot of numbers in the mid-900s to a trillion dollars in terms of the company's current valuation. So the company says it has not decided on the timing of the IPO, citing ongoing internal considerations that may be easier to address as a private entity.

Starting point is 00:17:55 So yeah, we saw some reporting this week that said Open AI may weigh up to a year to go public because there are some big projects that they want to kick off that they say would be easier to do as a private company versus a public company, but obviously filing the paperwork now gives them the option and the flexibility to go public sooner if they so choose. So OpenAI's chief rival Anthropic has announced plans to go public last week, highlighting the surge of AI firms entering the public markets so far in 2026. And yeah, we're going to be talking about Space X, obviously. But it's not only that Open AI is saying, yeah, we're going to go public at some point, but at the same time, they also said, yeah, we're going to start a pricing

Starting point is 00:18:40 war, but in a good way for consumers. So the Wall Street Journal reported that OpenAI is considering major cuts to token pricing as it prepares for tougher competition with Anthropic and as enterprise customers complain that their AI usage bills are rising too fast. Yeah, now all of a sudden, all those companies that were going hard on token maxing, uh, quarter one and blew through their budgets, all of a sudden, they're like, wait, we should probably be focusing on token efficiency. And weird, I told you all about this. When do we go over this? Go listen to episode 789, our Start Here series, when we talked about this shift from token maxing to token efficiency, but this move from Open AI signals that they're, well, acknowledging that

Starting point is 00:19:27 and that they need to start highlighting, well, they're right now the most token efficient company when it comes to price. So they're looking to even maybe potentially undercut a little bit more after Infrappix' recent fable model release, which we will also talk about here in a second. But next big piece of AI news, Jeff Bezos saw Elon Musk become a trillionaire and he's like, that sounds good. Let's raise more money for another one of my companies. So Prometheus, that is the AI startup led by co-CEO Jeff Bezos in Vic Baja. They announced a massive $12 billion. Yes, billion with a B. They announced a massive $12 billion funding round this week, valuing the company at more than $40 billion, according to Axios. So Prometheus aims to revolutionize

Starting point is 00:20:23 manufacturing and engineering by building what Bezos calls an artificial general engineer. So, Yay, we have one more acronym, right? So not only do we have HGI artificial general intelligence and artificial super intelligence, but now we have to keep track of an artificial general engineer because that's apparently what Prometheus is working on. They call that an AI system designed to speed up the process from design to production for physical objects, including bridges and microchips. So major investors in this round include J.P. Morgan, BlackRock, Goldman Sachs,

Starting point is 00:20:56 DST Global and Arc Venture Partners. So Bezos is the largest backer of Prometheus's early series A round that brought in $6.2 billion, participated again and said the company's capital intensive nature requires significant investment. So Baja described how the AI could tackle complex engineering projects like designing and manufacturing a jet engine, traditionally a decade-long process for human teams as an end-to-end AI problem instead. So Prometheus dropped the project from its name. So previously it was called Project Prometheus. Now it's just Prometheus.

Starting point is 00:21:41 And they now have more than 150 employees with offices in San Francisco, London, and Zurich. So when asked about an IPO, yeah, asked about an IPO already. Bezos said it's too early to think about that, but confirm that he's seeking to raise up to $100 billion for an affiliated fund to buy manufacturing companies that could benefit from Prometheus's technology. So Bezos began as a founding investor in late 2024, but was so impressed that he stepped into the co-CEO role, making his first kind of big appearance since leaving Amazon as CEO in 2021. One. There's more, y'all.

Starting point is 00:22:27 This wasn't just about anthropic and Open AI and Apple and Jeff Bezos because now we have big news from Google. Little curveball here. So Google will appeal a landmark German ruling, a court ruling that found the company was legally responsible for false information generated by its AI. overviews feature. So according to Reuters, this case centers on AI generated summaries that appear above traditional search results, right? So we've seen all those like when they accidentally, you know, said eat glue or something like that. That's not what this lawsuit is about, but we remember all those Google AI overview snafus of the early days. Well, now some German companies took Google the court and won the case, but Google is promising to appeal. So here's what

Starting point is 00:23:21 happens. Two German publishers brought the case after the Google AI overviews falsely associated them with scams and questionable business practices. So Google says the vast majority of AI overviews are accurate, but admit that some summaries can miscontextext or misinterpret web content similar to other search features. So the ruling is expected to set the precedent that could impact other AI developers and how they deploy AI generated content in search and other online platforms. Google stated it takes quick action when AI overviews violate its policies and emphasize that this case involves specific and narrow errors, not the overall design of the feature.

Starting point is 00:24:05 So the integration of AI into Google search results has faced criticism from publishers and content providers who say it reduces their web traffic. readership and revenue. Antitrust regulators are also investigating the impact of AI generated content in search, raising broader questions about competition and content ownership online. So this decision comes as Google opens a new AI center in Berlin signaling continue investment in AI in the region despite the recent legal and regulatory challenges. So this one will be one that we're going to be keeping an eye on because as I understand, understand this is the first time Google's anything when it comes to Google's AI search or they've

Starting point is 00:24:51 lost any kind of lawsuit or case about anything being false or not correct. So a pretty big ruling. We'll see what happens on the appeal. All right. Yeah, big Microsoft news is now as well. This one technically half Anthropic, half Microsoft. So let's start at least with Anthropics new model. And yeah, I'm not going to spend too much time on this because we already did a dedicated episode on Anthropics new Fable 5 in Mythos 5 models that were released. Pretty big release. So go check out episode 796. All right, we did a dedicated episode. So we'll go too much on the actual release and the benchmarks and the use cases and where it shines, right? We covered all that in that episode. But let's at least get the basics for those that only maybe tune in on Monday. So

Starting point is 00:25:42 Anthropic has released Fable 5 and it is generally available. So both in Claude A.I. In subscriptions only until June 22nd, when Anthropic said it won't be available for subscribers anymore, at that point, you will likely have to pay the API cost, which is very expensive. More than twice the cost of Opus 4.8 at $10 per million input tokens and $50 per million output tokens. So the company is calling it its most capable. model to date. And the headline feature from Anthropic is long horizon work. So Anthropic says

Starting point is 00:26:20 that early testers who used Fable 5 for major software engineering, finance, vision, legal spreadsheet and research tasks said that it took months of work or engineering into days or hours. So in terms of the actual capabilities of the model, it is pretty impressive. terms of what the model can actually accomplish, especially across these specific domains. So that is the Fable model, but Anthropic also launched Claude Mythos 5 for a limited group of cyber defenders in the Project Glasswing group, and then they later kind of select biology researchers. So essentially, here's what the Mythos Fable is. So Mythos 5 is the kind of the big unrestricted model that's only available for a select group of people in Project

Starting point is 00:27:16 Glasswing, but they did make Fable 5, which is essentially Mythos 5, but with some more guardrails, but it's in the Mythos family. So that is the one that's available to right now, at least subscribers using it on the API Claude Code wherever you use it. But like I said, four subscribers only available until June 22nd. But here's the big news and the update that I think gets worth paying attention to. So although Fable 5 got plenty of benchmark love and applause for its ability to, I mean, do some really impressive work, you know, even replicating complex software, the model has also received the harshest backlash, probably of any model ever released.

Starting point is 00:28:01 So Anthropic has faced ongoing controversy due to Fable's aggressive safety controls. It's once hidden model rollback. and its mandatory data retention policies, according to reports from the Verge, Reuters, Forrester, and more. So here's where we get into the backlash that was really just dominated the news cycle this week. So Anthropic has essentially eliminated zero data retention agreements for Fable 5 because now they require all prompts and completions. So they require anything that you send to Claude and anything that Claude produces from what you send it to be stored for 30 days. So yeah, there's no more zero data retention agreements, HIPAA, all these things. Yeah, it's bad.

Starting point is 00:28:52 So now Anthropic stores that for 30 days with flagged conversations potentially held for up to two years and safety scores for up to seven. years. And if something does get flagged, that means a human has the ability to read it. So why is this such a big deal? Well, you'll see now. Finally, we'll get to the story I have on my screen here. So this mandatory 30-day data retention has actually led Microsoft to block internal use of Fable 5, citing risks of proprietary code exposure and conflicts with its strict internal compliance standards, even as Microsoft continues to sell the model to external customers. Let me talk real quick just about how big of a deal this is. So Microsoft is one of the biggest investors in Anthropic.

Starting point is 00:29:46 Not only that, but Microsoft makes money when you use Anthropic in its co-pilot. So copilot is powered by both OpenAI's GPD models, as well as Anthropics Clawed models. So not only does MySQL models. Microsoft make money when people use Claude, right, via the foundry and everything else, but they are one of the biggest investors in Anthropics. So you would think if nothing else, Microsoft is pushing anthropic use out the wazoo. And it is the opposite. So apparently, according to reports, they cannot use it internally for that very reason,

Starting point is 00:30:22 the 30 day data retention. And, you know, I think actually for me, I kind of took it easy on Anthropic on the 30 day data retention. Because I didn't want people to be like, oh, my gosh, Jordan, you're so biased. And I'm like, no, I just, this is the truth. So now that, you know, one of the largest companies in the world has done this, I can say, for me, I don't care. And I've said this on the show all the time, right?

Starting point is 00:30:47 I'm a small business owner. I don't care. Anyone can take my data. You can take all my inner, you know, my, my books, my revenue, you know, my, my game plan, my internal. I don't care, right? If I'm a company, if I'm an enterprise, I'm following Microsoft's lead right now. And I'm saying if Anthropic has this 30-day mandatory data retention, I'm not touching it. This is unprecedented, y'all.

Starting point is 00:31:15 Like this doesn't literally doesn't exist, has never existed to where if you use Fable, not only anything that you put in there is going to be retained by 30 days. But if it gets flagged, which there's not a lot of permission on what could get flagged, because it seems like anything is getting flagged, and I'll get to that here in a second. But at that point, not only is your data, you know, kept by default, anything that you send for 30 days, but also it can be, if the conversation is flagged,

Starting point is 00:31:44 it can be potentially held for up to two years. All right. So not only that, but there's been a lot of other issues that have caused some of these, you know, quote unquote false flags for reasons. Well, Anthropic has at least addressed some of them. So early users, reported that Fable 5's safety system aggressively downgraded or blocked benign prompts,

Starting point is 00:32:06 often without notification, which led to frustration among software developers, security researchers, scientists, etc. So there's essentially, there were these areas where, you know, Anthropics said, yeah, you can't use our models in these areas. So even when people went in with no context and just said, you know, hey, Fable 5, how are you, they would get blocked, right, without saying absolutely anything because also, depending on your settings, it carries over your past context and your past chat history. So even if you decide, oh, I'm not going to share anything, I'm just going to use it for public

Starting point is 00:32:44 facing information, right? We don't really know what's happening under the hood. But clearly, infropic is bringing in your past content if you have that, you know, remember chat history enabled. So dicey stuff. That's all I'm saying. And I feel better saying that now that Microsoft is reportedly saying, yeah, we're not going to use this model because we don't want Anthropic to have immediate and exclusive access to literally all of our data. So, yeah, if you don't know how that works, you know, Anthropics previous models, you just didn't happen.

Starting point is 00:33:16 You know, all other models, it doesn't happen. If you're on a paid business plan, companies can't use your data, right? They can't store your data unless ordered by a court. So the first time an AI company has ever chosen to say, yeah, if you use our model, we're going to keep your data for 30 days. Wow. So not only that, but the model's reliance on these intensive reasoning loops means that it consumes far more tokens per query, compounding costs for developers and making it less accessible for startups or smaller businesses. So critics argue that Fable 5's powerful mythos core is effectively muzzled by these heavy guardrails, forcing users to pay premium prices for performance that they cannot fully access.

Starting point is 00:34:01 So, yeah, essentially, there's all these rollbacks as well. So previously, if you were using Fable, it might have rolled back to Opus 4.8, but not told you, right? So if you were using this for researching as an example, anything on machine learning, artificial intelligence, if Anthropic thought that, oh, maybe you're distilling our model information, even if you were using it for legitimate research purposes, it might roll back to a quote unquote, dumber opus 4.8, right? Opus 4.8 is still an amazing model.

Starting point is 00:34:32 But you would think that you were getting a mythos level fable output and it would be instead feeding you opus 4.8. So a lot of people in the AI community, researchers, people who are trying to even, right, on the bio side, biotech, biomed, right, they're getting these things flagged or they can't use the model.

Starting point is 00:34:51 So it's like, okay, if one of the big, you know, upsize of AI is we can use it to cure diseases and medicines and protein discovery and all these things. And we can't use the model or, you know, we think that we're using this big, strong model that's amazing that we're paying all this money for. And we don't know, but you're actually feeding us Opus 4.8 instead. Not a good look, which is why I think this has been the biggest week of backlash, maybe ever, right? I think there's a lot of belly aching when GPD5 came out.

Starting point is 00:35:20 But I think that was not like backlash per se. It was just people were underwhelmed, right? They were like, we thought it was going to be way better and it was just a little bit better, right? People were mad that they got, that they were going to get rid of GPD 4. So I don't think that was necessarily backlash. That was just people, you know, maybe addicted to the sycophantic GPD4O and they're like, oh, we want it back, right? You know, this 5-0 is, or, you know, GPD5 isn't that great. But this is different.

Starting point is 00:35:45 This is, you know, really unprecedented uncharted territory. So I think, you know, I don't know. We'll see how this plays out. I assume, right, maybe even by the time. Like I said, I'm recording this a day or two early. Hopefully by Monday, Tuesday, this is gone because this is not good just for the industry. I don't think any other companies would follow suit, but who knows, I could be wrong. Yet, a pretty big story to keep an eye on.

Starting point is 00:36:09 All right. So that is our big AI news stories for the week. Like I said, if you want to know more on the Fable 5, make sure to go listen to episode 796. We also did the entire rundown on Apple's WWDC and, announcements. Go listen to that in episode 794. And also on Wednesday, we went over the new Codex sites doing a hands-on on that one. I thought it was a really fun show. So make sure you go check that one out as well. But here is the roundup of what's new and what's next because these seven big stories weren't the only big things that happened. There was a lot more.

Starting point is 00:36:46 So some of these are rumors, leagues, just smaller stories. Here we go, bullet point style. Here's everything else that you need to know. So meta is reported. keeping distance from Trump's AI ownership idea, which could delay the president's plan for a public sharing wealth plan with AI companies. We talked about that in last week's AI News That Matters. Kimmy released their Kimmy for work with native agent swarms and 300 local agents. President Trump did say, aside from meta shying away, that all the top AI firms may give the government stakes ahead of these blockbuster IPOs that we talked about. Speaking of those IPOs completely forgot to get the space X story in. So we'll actually tag that one on at the end.

Starting point is 00:37:32 But back to our what's new and what's next. Sam Altman, speaking of IPOs, reportedly told staff at OpenAI that they could go public within a year. Next, New York lawmakers approved a one-year pause on large data centers. Next, OpenAI agreed to acquire ONA, adding secure cloud execution technology to Codex. Google began rolling out Gemini 3.5 Live Translate for speech translation across how many languages. I think that one's really cool. We went over that one in Friday's show where we go over our Friday's features. Next, Chad GPT rolled out interactive charts on mobile and web. Google is rolling out Gemini 3.5 and code tools to select notebook LM users. Right now only Ultra. So we'll cover that once it goes out to everyone else. Next, a Mississippi federal judge removed all

Starting point is 00:38:23 lawyers from a case over AI hallucinated citations. Not good. Gemini Omni Flash takes first in Arena's video leaderboard. Next, OpenAI is planning a $500 billion Ohio infrastructure deal, according to reports. Speaking of Arena leaderboards, Fable 5 from Claude took first place in almost all categories, including text and agents. Oracle customers can soon use their universal credit. for Open AI models and Codex.

Starting point is 00:38:56 IBM in a new study found that two-thirds of tech leaders accountable for AI that they don't even control. China reportedly is preparing a nearly $300 billion AI funding plan for nationwide infrastructure buildout. Amazon secured a $17 billion delayed draw loan for AI infrastructure. Open AI added email drafting, editing, and sending pretty cool. We went over that one on. Friday show as well. And J.P. Morgan said that AI agents drove a 20% increase in private banking

Starting point is 00:39:31 gross sales. All right. So that's all for our kind of bullet point round up. Now, let me not forget, you know, one of our biggest stories. We'll actually do this one to wrap here. So space X. Yeah, space X is public. And what a big splash making Elon Musk the world's first. trillionaire. So SpaceX went public officially and live on Friday. And the SpaceX shares jumped about 25% since they're opening in their NASDAQ debut, trading near $170 and pushing the rocket company's market value to roughly $2.2 trillion. Yeah, talk about a astronomical in terms of their reported revenue right now. So the stock opened. at $150 per share above its 135 IPO price, but below the 170 level initially suggested by

Starting point is 00:40:34 trading desk interests showing strong demand, but some caution around the evaluation. So the debut reportedly makes Elon Musk the world's first trillionaire on record and is expected to create thousands of new millionaires at Space X. in several new billionaires among the earliest SpaceX employees and investors. So Goldman Sachs rose about 3% as investors rewarded its lead role in the IPO, while space-related stocks, such as Redwire and Rocket Lab, fell sharply as attention and capital shifted toward SpaceX. So CNBC reported the retail traders showed heavy interest in the offering, but SpaceX

Starting point is 00:41:20 allocated a smaller than expected share of IPO stock to retail buyers, leaving many to chase shares after trading begin. So this is big for a lot of reasons that are outside of what you think Space X actually is. That's because in May, essentially Space X became a lot of different things, but GROC and XAI, kind of Elon's other companies on the AI side, are a part of SpaceX. So now not only does Elon Musk have two companies that are top 10 in market cap with Tesla in there as well. But I think space acts and we'll see this more how this plays out over the next year or so. But what we've seen aside from them being the rocket company and now housing XAI and Rock,

Starting point is 00:42:09 well, there maybe could be a big compute player as well. Because we've seen stories in recent months about how, as an example, they have a multi-billion dollar longer-term partnership with Anthropic. So all of these other AI companies that maybe can't get the compute that they need, it looks like Space X is going to be a major player in that. So not only from the data centers that they've already built, the Colossus series, but also with the group's plans to eventually have compute in space, right?

Starting point is 00:42:44 Which sounds crazy, but a lot of companies are talking about it. And obviously, you would assume the AI slash space company would be one of the first to actually accomplish that at scale. So this is a lot more, right? So as we talk about SpaceX in the coming months here on the show, it's not just about the rocket company. Right. Space X also includes XAI slash rock as well as just the ability to serve that AI computes that thousands of enterprises. companies need because right now if you want to bring a big AI project online, it could be difficult

Starting point is 00:43:22 unless you're like a Fortune 100 company to be able to get the compute that you need some time to run the world's most advanced and agentic models. All right. So that is a wrap all of our big AI news stories. So I hope this one was helpful. So like I said, we do AI news that matters on Monday. On Wednesdays, we go hands on and we bring you AI at work on Wednesday. is usually a demo of a new, you know, new, you know, AI release or sometimes we do it by category. So make sure you tune in for that one. And then on Fridays, we go over AI feature Fridays.

Starting point is 00:43:56 So all the new features that you can actually use live that day, we do that on Fridays. And then on Tuesdays and Thursdays, we kind of rotate the show. So that's our lineup. If you only listen on Mondays, which I hear a lot of people do, I encourage you to look outside to just Monday. So thank you for listening. Make sure you go to your everyday AI.com.

Starting point is 00:44:15 Thanks for tuning in. We'll see you back tomorrow and Everyday 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.

Starting point is 00:44:31 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.

Everyday AI Podcast – An AI and ChatGPT Podcast - Ep 798: SpaceX goes public, OpenAI declares pricing war, Microsoft stops Claude Fable use and more AI news

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.