← Everyday AI Podcast – An AI and ChatGPT Podcast
Everyday AI Podcast – An AI and ChatGPT Podcast
Ep 739: OpenAI building Superapp, NVIDIA’s trillion dollar AI play, Microsoft’s big AI shakeup and more
00:33:26
Loading summary
Transcript5 lines
- [00:01]
A
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.
- [00:16]
B
What does OpenAI have up its sleeve? Apparently something big after the company said they'd slow hiring in January. Yet here we are and now they're reportedly looking to double their headcount by the end of this year. Doubling headcounts wasn't the biggest number of the week. That was probably Nvidia CEO Jensen Huang talking about the more than trillion dollars in real inference demand by 2027. Oh, and while no one was looking, the White House is trying to rewrite the rules of how AI is governed in the US and oh, did you miss this one. Microsoft might sue OpenAI and Amazon while it has also put a new person in charge of running Copilot. My gosh. And the craziest thing, those might not even be the biggest pieces of AI news this week that'll actually impact you as new reports say that OpenAI is completely changing the ChatGPT app this year and they're going super. Yeah, if you need to make decisions that impact how your company uses AI, you have to pay attention to all of these AI news stories and developments. But it's pretty much impossible. You'd have to dedicate multiple hours a day to do it. Or you can just join me on Mondays as we bring you the everyday AI news that matters on Everyday AI. Well let's get into it. Welcome and what's going on? If you're new here, my name is Jordan and this is Everyday AI. This is your daily live stream, podcast and free daily newsletter helping everyday business leaders like you and me not just keep up with AI, but how we can make sense of all the non stop developments, know what matters, what doesn't and use that information to get ahead grow our companies and careers. So make sure if you haven't already, go to our website at your everyday AI.com we're going to be recapping today's show and a whole lot more. So if you are new here, most Mondays we bring you the AI news that matters. On Wednesdays we do a deep dive, usually going hands on on probably usually the biggest, you know, kind of AI update of the week. And then on Fridays we're trying out a new series called Friday Features. Just going over a handful or so of more practical large language model updates that you can use right away. And then on Tuesdays and Thursdays, you know, we'll splice in some other shows so without any further chit chat, here's the AI news that matters for this week. Well, the first one is pretty big. That's because this happened on like Friday afternoon. And sometimes as a former journalist, I know it's those news stories that sneak in on Friday, usually, you know, doing that for a reason so they don't get a lot of news coverage. Not sure if that's what's happening here. But regardless, it's pretty big. That's because the White House on Friday released a six point legislative framework seeking a single national policy for AI, aiming to set uniform safety, security and economic guardrails and to stop states from passing their own conflicting AI laws. That's the big one here. So the Trump administration proposed a federal single standard for AI regulation and is asking Congress to preempt state AI laws that would create a patchwork of rules, arguing this will avoid undue burdens of industry and preserve US Competitiveness in AI. So the framework covers those six broad categories, including child safety protections, standardizing permitting and energy rules for AI Denisetters and protections for creators, intellectual property, and limits on using AI to silence lawful political expression. So the White House said it wants Congress to turn the framework into laws this year and aims for bipartisanship support, with the Office of Science and Technology Policy Director saying the plan will unleash American ingenuity. So the timing is obviously politically sensitive. Congress is deeply divided and Republicans are holding right now. Narrow majorities and lawmakers are currently occupied with other high profile bids and the election year calendar that's right around the corner. So the administration's push though, directly responds to recent state level efforts. We've seen some different state laws in New York, California and others that have been advancing their own AI rules, which the White House and industries say could fragment the market and slow innovation. So this will be pretty, pretty interesting to see how this ultimately unwinds. So here's kind of the nutshell of how it's all gone since this, the kind of state versus federal debate on AI first rolled out. So essentially this past December, the Trump administration passed or not passed, but they came out with their executive order on national AI policy, which kind of was trying to preempt state laws from contradicting anything on the federal side. So the Trump administration said, well, we just want to kind of make the rules on AI, not let states make their own, which is not normally how things would work. And then obviously since that time, different states have both introduced and passed and will soon enact new state laws and regulations on AI. So Those two things obviously contradict each other. And the state laws will continue on until they are shot down by, or if they are shot down by a federal judge. So we will see. But I think the biggest takeaway here, and this is kind of the sticking point, I think it will be one of the sticking points because it's going to be one of the things that's going to be the most divisive. So in the new framework, the White House essentially says that training on AI, training AI models on copyrighted material does not violate copyright laws. So that's obviously something the big tech companies are going to be very much so in favor of, but pretty much every, everyone else is going to be against this. So that's why I don't necessarily expect anything in here to become law, because that would have to be passed by a very divided Congress. Right. So do a very brief background of reporting, you know, on Illinois and, you know, some congressional politics way back in the day. So I know how slow these things go. So I do think that this is probably a step in the right direction. Turning an executive order, which essentially has no teeth, into a frame framework, you know, that starts to get it there, but it ultimately does have to get passed by a very divided Congress. I don't necessarily see that happening, at least not in its current state. And probably because of what I take it as the biggest takeaway is essentially the White House saying, yeah, these big tech companies can train on copyrighted data, which I don't think is going to fly over well with a lot of these big media companies that, well, they are the ones that are controlling a lot of the advertising dollars. So it's going to be two different big sectors at play here. But we'll obviously be following any new updates on this. So next, OpenAI is reportedly going super well. What does that mean? They're going to try to throw all their apps together into one super app. So according to some recent reports, OpenAI is reportedly building a unified desktop app that merges its ChatGPT, Chatbot, its Codex coding tool, and the Atlas browser into one product. A move at aimed at simplifying development and improving user experience. So the reported plan would bundle ChatGPT, Codex and Atlas into a single desktop application, with the goal of just making it easier for users to come in and share their information, get better outputs, and reduce product fragmentation. So the Wall Street Journal first reported the bringing all of these together into one, and the company executives confirmed the initiative internally, saying the company had been spreading EFF across too many apps and stacks and needs to simplify. So that's according to a memo from applications head Fiji Simo. Fiji did also I asked on one of her tweets if this, if she was kind of confirming that this is happening and she did at least like that tweet from me. So she didn't confirm it, but it seems like this is actually going to happen, which is great. It's something I've actually been, you know, telling people at OpenAI and, you know, advocating for. Like, I actually love the way that Anthropic does it. In their desktop app they have their, their chat, their cowork and their Claude code. So this is a, what I think is a step in a great direction for OpenAI. So OpenAI also amidst all this, they did recently introduce smaller, faster variants of their GPT54 model, the Mini and Nano. And these coding focused versions reinforce the firm's push to better serve developers and enterprise users. So analysts so far are talking about this combined super app and they're saying it will allow stronger personalization, which I think is why ultimately OpenAI is going to want this to happen. And the system could learn a user's coding style from Codex and research interests from Atlas, then deliver more tailored ChatGPT suggestions. But I think what this ultimately is going to be a big play for and we'll see how they, if, if this does come to fruition, what's going to be telling is the tiers that they allow this for. So my assumption is this will be available for free, this new, you know, desktop super app, and most of the features will be available. My thought is, well, it's really just marrying and maybe popularizing a little bit more the Atlas browser, which is a really good browser from OpenAI. But I think after its initial launch it's, it's, it's one of those things, I think it's so hard to continue.
- [10:20]
A
Foreign.
- [10:26]
B
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. 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. 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. Momentum for different projects after they get released. So like I said, I think Atlas was a very powerful agentic AI browser and they've actually been getting a lot of good updates in there as well. But it seems like over the last few months some other kind of products, maybe even Codex, has taken away maybe some of the spotlight from ChatGPT's Atlas browser. So I do think and assume that they'll be rolling this all out to all users. But the main thing is actually I think going to be for ads, right? I do think that by putting all of this into one app, right? So bringing your Codex, Chat, GPT and browser information, history and data all under one roof and giving that to free users who are now being served ads, I think that is ultimately going to create a jackpot for OpenAI. That's just my take on it. So we'll see if that is ultimately where this leads. All right, our next piece of AI news. Well, Microsoft and OpenAI's relationship may not be as nice as we think. So according to the Financial Times, Microsoft is considering legal action against OpenAI and Amazon over agreements tied to the roughly $50 billion deal that could conflict with Microsoft's exclusive previous cloud arrangement with OpenAI. So according to reports, Microsoft believes the new agreements between the companies, including a pact making aws the exclusive third party cloud provider for OpenAI's Frontier Enterprise Platform, may breach exclusivity deal with OpenAI that centers on Microsoft's version Azure. So the possible legal move follows public statements by OpenAI and Microsoft that Azure remains the exclusive cloud provider of stateless OpenAI APIs, a Microsoft spokesperson said, stressing Microsoft's confidence that OpenAI understands its legal obligations. So the core dispute is whether OpenAI can offer this new Frontier program via Amazon Web Services without violating the Microsoft partnership clause that requires OpenAI's models to be accessed through Azure. So financial scale and timing make this particularly high stakes. So the Reuters summary of the Financial Times article links the agreements to a package of deals worth about $50 billion, raising competitive and contract risk across the cloud market. So Microsoft executives reportedly view the AWS frontier approach as likely inconsistent, the spirit, if not the letter of their contract, and say they prefer to resolve the matter in talks but are prepared to sue if necessary. So OpenAI and Amazon did not immediately comment to Reuters and Microsoft emphasizes it will enforce its contractual rights if breached. So yeah, this is going to be one of those kind of continue to munch on the popcorn as OpenAI and Microsoft's once very tight knit partnership does start to, you know, go in other directions. So they obviously still have a lot of ongoing business together, right? A good majority of Microsoft copilot's models run OpenAI, right? And they did so pretty much exclusively until they started to debut some of their, some of the anthropic CLAUDE models into their Office products late in 2025. Now obviously Microsoft has come on as a pretty heavy investor to OpenAI's chief rival in Anthrop Tropic. And now OpenAI on the cloud side is starting to enter into a lot more cloud partnerships, right, with other big companies including aws, which is one of Microsoft's biggest competitors on the cloud side. So yeah, we're going to see how this one shakes out. Obviously any big news we'll have in our newsletter. All right, well, 50 billion might sound like a lot of money, but how about $1 trillion in real money? Well, that's the amount that Nvidia CEO Jensen Wong says Nvidia will see in real AI demand. So he highlighted that 10, sorry, 40 million fold increase in compute over 10 years. Which is, yeah, not a misprint. 40 million fold increase in compute over ten years. So that came from Jensen Wong at the Nvidia GTC conference where he did announce that number that grabbed so much like so many headlines because I think a lot of people are talking about, you know, AI. Is there really actually money out there or is it just all these, you know, circular financing deals? Well, Jensen Wong said how about a trillion actual dollars through 2027. So aside from that big number that grabbed headlines, Wong also officially unveiled the Vera Rubin platform. That was Integr introduced as the new integrated stack with Reuben GPUs, Vera CPUs and a new GRO3LPX inference accelerator designed to run model queries. So yeah, Nvidia had that kind of aqua hire of the GROK technology. That's Groq, right? The inference company, not G R O K, the XAI chatbot. So Nvidia claims up to 35 times higher inference throughput per megawatt versus prior generation hardware. So Huang framed the core economics of what's happening right now in the AI world, he said. Data centers have fixed power budgets, so tokens per watt is a new critical measure for companies to focus on. So Nvidia claims that Vera Rubin delivers roughly five times the revenue per gigawatt compared with Blackwell and Grok. Equipped racks deliver much higher tiers. So Wong predicted token pricing tiers from free to about 100 $150 per million tokens, and said every engineer will likely receive an annual token budget, signaling a change in how companies budget for AI compute. So aside from all those chips and numbers and all that fun stuff, Nvidia also had some other announcements. We covered a lot of them in our newsletter, but probably one of the ones that's going to continue to be talked about a lot is while they launched Nemo Claw, which is their secured version of OpenClaw for building agent based personal and enterprise AI. So Nemo Claw pairs open shell security and has deploy options including a 13 cent an hour instance on Nvidia's Brev Brev console and an open GitHub repo for self hosting. You do need a Windows PC with a decent Nvidia chip to go that option. Also Nvidia also showed off new advancements in robotics, space use cases, Robo taxi rollouts with Partners projected across 28 markets by 2028 Disney Train robots demo, and the Vera Rubin module that's planned for orbit, illustrating Nvidia's push beyond data centers all right, our next piece of AI news. Yeah, a big shake up at one of the biggest companies in the world and its leader is kind of out. Well, kind of not really, but still kind of. At least when it comes to the day to day operations of Microsoft's Copilot. So Microsoft announced an executive reorganization that unifies consumer and commercial copilot efforts and refocuses key leaders to new areas. So Microsoft is consolidating part of its consumer and commercial copilot efforts into a single unified program to create a more cohesive assistant across business and personal use. So Jacob Andre will lead the copilot experience across both commercial and consumer products and will report directly to Microsoft CEO Satya Nadella, taking responsibility for design, product and engineering. So this change makes Ontario accountable for aligning Copilot's features, look, feel and roadmap across different customer segments after years of separate teams and inconsistent features. And well, maybe the bigger news in all of this is the movement of Mustafa Solomon. So he is the still the Microsoft AI CEO and will shift his time away from leading co pilot on the consumer and enterprise side and instead will focus on building Microsoft's own AI models that they're hoping will kind of lead them to super intelligence. So he will reportedly remain involved in day to day operations and retain a dotted line connection to Andre. So Nadella framed the reorg around four connected pillars. The Copilot experience, the Copilot platform, Microsoft 365 apps and AI models with the goal of moving from multiple products to an integrated system that is simpler and more powerful for customers. There's also a new Copilot leadership team that will obviously include Andre, Ryan, Roslonsky, Perry Clark and Charles LaManna as they will lead the different platforms under the new umbrella. So this shift, I mean it's pretty big because this is Microsoft kind of acknowledging that their current approach, well, it just wasn't working. Whether it's from a personnel or how they just structured this internally. And it is kind of them acknowledging that having separate consumer and business copilot experiences failed to just work. Right. So if you didn't know this, I think most people use the enterprise or the business version of Microsoft Copilot which is, you know, integrated into their operating system. But there was the consumer version of Copilot which looked very much like, well, if you've used the product before, was very much similar to Inflection AI. And that's Microsoft kind of Aqua hired that company. So it had two very just different looks, different designs, different feels. It worked two very different ways, which I always found confusing and I think a lot of people did as well. So we'll see if this new move, you know, improves. Improves kind of what Microsoft is hoping to accomplish here. And Solomon's move is to focus on model also raises immediate questions about future leadership for Edge, Bing, MSN and Microsoft's advertising businesses which had reportedly or sorry, which had previously reported to Solomon and were tied to prior Bing to Copilot efforts. So yeah, a lot of shakeups happening at Microsoft, all right. But the biggest news when it comes to staffing is not actually Microsoft kind of reorgang and having a new person run its Copilot flagship. It's probably actually what OpenAI is doing as a company. And that's because we heard back in January from CEO Sam Altman that the company was going to drastically slow down its pace of hiring. Well, maybe not so much anymore. That's because a new report by the Financial Times said that OpenAI is planning to increase its workforce from about 4,500 employees today to roughly 8,000 employees. By the end of the year. So yeah, that's very not normal for a company of that size to double, you know, over the course of what? That would be like eight and a half months or something? No, like seven and a half months. Yeah, that would be absolutely wild. So the hiring drive is notable because it does run counter to broad tech industry layoffs that are happening right now. Right. We just saw a huge layoff at Meta, we've seen some big layoffs at Microsoft. So it looks like, like OpenAI might be going counter to what everyone else is doing. So according to the reports, new hires would span across product development, engineering, research and sales, with the company also recruiting what they're called technical ambassadorships, specialists, which would essentially help businesses implement OpenAI tools. So the ramp up appears aimed at strengthening OpenAI's competitive position against its other big competitors, including Anthropic and Google, which have started to gain traction from different parts of OpenAI's kind of pie that they had dominated, you know, since early 2023 and 2024. So OpenAI's recent deal as well with the Department of Defense and reported advanced talks with private equity firms such as Brookfield Asset Management underscore demand from both government and corporate buyers. And like we talked about earlier with the new, the reported potential lawsuit between Microsoft, OpenAI and Amazon. Well, OpenAI also has the Frontier program. So OpenAI, not saying they're turning into a consulting company, but that does seem like it's going to be a big part of their position moving forward. Because if they want to increase stickiness across the enterprise. Right. They're already pretty much, at least right now, untouchable when it comes to weekly active users. Right. They're nearing a billion weekly active users. No one else is really close to that. But if they want that to stick in the enterprise, well, they're gonna need enterprises to actually understand how to use them. Right. This is something that, you know, we've been doing at everyday AI for companies and it's surprising, you know, that companies, even larger multi billion dollar revenue companies still don't really know the basics of using, you know, the enterprise version of ChatGPT because it changes so often. So a pretty big shift here between, you know, OpenAI potentially completely reshaping and going into just one big super app to now. Well, they're saying they're going to double, almost double headcount. So this is a pretty big, big statement from OpenAI, especially since we heard in January they're going to slow down hiring and there's been all this talk out there on, you know, anthropic catching up, Google catching up, and a pretty bold statement here from OpenAI saying, yeah, we're gonna quite literally double down, bring all of our main products under one roof. And their advertising business is really picking up as well. All right, yeah, there's a lot, there's a lot more, but that's our main stories. Now. We're going to roll into what's new and what's next. Kind of our bullet point roundup of all the other stories, some of them big, some of them rumors, some of them smaller updates, but we're going to cover them all quick bullet point. And remember, we might be covering some of these things in a little bit more depth either on Wednesday as we do our AI at work on Wednesdays or on Friday as we go over our feature Friday. All right, here we go. So Elon Musk announced the company's terrafab chip plant in Austin, Texas. Microsoft, excuse me. Still fighting the sickness, y', all, but hey, this is live. It's, it's unedited, unscripted, so excuse me. Microsoft started rolling out its Copilot task features for those that were on the waitlist. HHSBC is considering 20,000 job cuts as AI reshapes back office work. Grok is reportedly working on a Grok computer, their version of Perplexity personal computer which is kind of just open Claw cloud. Claude code released scheduled tasks. Yay. But only on the cloud, not on the desktop. Next, Perplexity's agentic computer browser landed on the iPhone. We actually went over that this past Friday on our Friday feature show Google AI Studio some pretty big updates. They released an updated full stack experience inside Google AI Studio. OpenAI is reportedly in late stage talks for a $10 billion enterprise AI joint venture. Nvidia unveiled their Nemo Claw agent Runtime. Manus released my computer, which is their version of openclaw. Lovable released a general purpose agent. So not just in the coding game anymore. Claude code added live channel integrations for Telegram and Discord. So again, very similar to openclaw, you can start chatting with Claude code in Telegram or Disc. Super Micro execs were charged over alleged Nvidia chip smuggling to China. Pretty big news there. OpenAI built a near real time monitor that reviews internal coding agents, thoughts and actions. Microsoft released their Mai Image 2 which did pretty good on the Arena AI charts. Google's Stitch relaunched as an AI native software design canvas. It is really good. I've been enjoying using the new update Update. Google expanded their personal intelligence across search, Gemini app and Chrome. Anthropic is hiring a specialist to stop AI guided chemical and explosive misuse. Huh. Talk about a concerning job description if you ever see that one on the Internet. Right? Nothing to be afraid of there. Codex added sub agents for parallel clean workflows. I've been loving using the new sub agents in Codex. FYI Anthropic rolled out Dispatch for remote control on iPhone. I've been enjoying that too when it works. We went over that this past Friday. If you missed that next Meta signed a $27 billion AI infrastructure deal with Nebias. ChatGPT gained right actions for Google and Microsoft apps. Nvidia confirmed its restarting manufacturing of a China tailored variant of its powerful AI chips. Google DMine hired Bridgewater's China Chief Scientist Andy Jassy. Amazon AWS CEO said AI could double AWS revenue to $600 billion. Two dictionary companies sued OpenAI. Anthropic also got sued by the music conglomerate BMG because they said Anthropic allegedly used copyrighted lyrics. Mid Journey released their V8 Alpha so in Google's AI studio you can now mix built in tools such as search and maps with your custom functions in one request. Atlas laid off 1600 employees, or about 10% of the workforce to fund AI. OpenAI released GPT 54, minnow, sorry, mini and Nano. Those aren't in the Chat GPT app. Those are more for developers. Although ChatGPT or OpenAI did update ChatGPT's model picker. I'm not the biggest fan, but I think I'll get over it. Minimax launched their M27 or sorry M2.7 model and agent on the API and platform. Mistral AI released both their open source Mistral Small 4 and they launched their Forge program which allows for enterprise model training. Anthropic made their 1 million token context generally available for Claude Backend, anthropology and Claude Code. Anthropic published a new study about what 81,000 people want from AI and Python tooling company Astral joined OpenAI as part of the Codex team and the latest acquisition, Aqua Hire. All right, that was a lot. So like I said, if you can't keep up with this every single day and you just need to know what is the AI news that matters? That's what we do here on Mondays. Give it to you. No bs, no marketing. A little fluff from me telling you what I think of it, but I hope this is helpful. If it was tell someone about it. So thank you for tuning in. If you haven't already. Please go to your everydayai.com Sign up for the free daily newsletter. Thanks for tuning in. We'll see you back tomorrow and every day for more Everyday AI. Thanks, y'.
- [33:05]
A
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.