AI Daily News May 08 2025: 🧑‍💼OpenAI Hires Instacart CEO Fidji Simo to Lead Applications 💳Stripe Unveils AI Foundation Model for Payments 🌍OpenAI Expands 'Stargate' AI Infrastructure Project Globally 💰Mistral AI Launches Cost-Efficient Models
12:31 Share

Welcome to a new deep dive from AI Unraveled. This podcast is created by Etienne Newman. He's a senior engineer and, fun fact, a passionate soccer dad up in Canada. Great to be here. Yeah. And hey, if you're liking these insights into AI, maybe take a second, just like and subscribe on Apple Podcasts. It really helps us out. Definitely does. And also, if you're looking to get

you know, more productive with AI tools, check the show notes. We've got a referral link and a discount code for Google Workspace. Oh, yeah. Was included. You get Gemini, PRO, Notebook Limb, Teams, basically a whole suite of tools to keep you ahead of the curve. Nice. And speaking of staying ahead, also check out Etienne's own AI-powered Jamga Tech app. It's designed to help you ace like over 50 different PBQs, quizzes, those tricky simulation certs. Really useful stuff. Sounds good.

So, yeah, welcome, everyone. This is where we dig into a whole bunch of sources, pull out the key stuff so you don't have to wade through it all. Exactly. And today we're looking at AI news specifically from May 8th, 2025. We've got a sort of daily chronicle of what happened. Right.

And the mission today is to unpack what these updates actually mean for you, for tech in general. Let's do it. What jumped out at you first from this chronicle? Okay, so first up was OpenAI. Big leadership news there. Seems like they're really pushing on the application side now. That's right. Yeah. Fiji Simo, the CEO of Instacart. She's been appointed as OpenAI's CEO of applications. Right. Instacart.

And Facebook before that, wasn't it? And she was already on their board. Uh-huh. Exactly. So she's got this, you know, deep background in scaling up products people actually use. It feels like a very strategic move. So it's more than just hiring a big name then. What's the, like, the deeper strategy here? Well, what's really interesting is how they're splitting things up. Simo focuses on applications. Getting the products out there. Yeah. And that lets Sam Altman really zero in on the core AI research.

And crucially, AI safety. Oh, okay. It signals, I think, that open AI is maturing. You know, they're realizing that building the tech and deploying it globally. Yeah. Well, those need different kinds of focus, different leadership. That makes sense. Yeah. Like building the engine versus actually driving the car and making sure it's safe on the road. Exactly. Okay. So switching gears a bit. Apple, there was this comment from Eddie Q during that Google antitrust trial saying

It really made me stop and think about the future of, well, the iPhone. Oh, yeah, that quote. Yeah. He basically said...

You may not need an iPhone 10 years from now because of where AI is going. Which is huge coming from a top Apple exec, right? Absolutely. He even compared it to how the iPod got phased out eventually. So does that mean they think the iPhone's days are numbered? Are they planning for that? Well, the analysts mostly seem to think it's not quite that drastic. Like, don't expect the iPhone to vanish next year. It's still central. Okay. But it's an acknowledgement, isn't it, that the tech landscape is shifting fast.

AI could fundamentally change how we interact with technology. Maybe the phone as a central hub becomes less essential. Yeah, I see what you mean. It might not be gone, but maybe its role changes, like its functions get spread out across different AI things. Potentially. And related to that, the Chronicle also mentioned Apple is looking at AI for search in Safari. Right. Potentially ditching traditional search. That's the buzz. They're reportedly talking to open AI, perplexity, anthropic.

you know, the big AI players using their tech instead of maybe Google. Wow. That would be a massive shift. Shows how disruptive AI is even for something as basic as web search. Totally. Okay. So from potential future tech,

Let's talk policy. Looks like there might be changes coming to the rules about exporting AI chips. Specifically, the Trump administration signaling a shift away from the Biden era rules. Oh, the AI diffusion rule. I remember that aimed at restricting chip exports mainly to China, right? Exactly. National security concerns were the main driver. But it was due to kick in soon. And now the Trump administration seems keen to pull it back and replace it. And what's their reasoning for wanting to change it?

The official line from a Commerce Department spokesperson is that the Biden rule was overly complex. Too complex. Yeah. And that it might actually hinder American innovation in AI. So the goal is supposedly a simpler rule, one that they say will boost U.S. innovation and dominance in AI. So it's that classic tension, isn't it? Security versus innovation speed. Precisely. So if they do ease up on these restrictions...

What happens? What's the immediate fallout? Well, the big chip companies like NVIDIA. Yeah. This would be huge for them. Potentially reopening markets like China. Okay. And of course, it would stir the pot geopolitically. Big questions about controlling access to this really powerful hardware. Right. Definitely one to watch. Okay, let's pivot to...

specific tools. Figma. Lots of designers use Figma. They seem to be making some pretty big AI plays. Oh, yeah. Their Config 2025 announcements were full of AI. It shows how AI isn't just like background tech anymore. It's getting baked right into creative workflows. So what did they announce? I saw something called Figma Sites.

Sounds like it could challenge website builders. Yeah, exactly. AI-assisted website building could definitely compete with things like WordPress. And then there's Figma Make. Make? What's that? It's an AI coding assistant.

Uses Anthropix Cloud 3.7, apparently. It can generate code, build prototypes from prompts, potentially a massive speed up for developers. Wow. Okay. So web design, coding, what else? They didn't stop there. Yeah. Figma Buzz is for AI-powered marketing content. Sort of puts them in the ring with Canva. Yeah. And Figma Draw.

for vector graphics. That's a direct shot at Adobe Illustrator. So they're really trying to become this like all-in-one design platform, keep you in their world for everything? Seems like it. Very ambitious strategy, taking on some established giants. Definitely ambitious. Okay, next up.

Something maybe a bit more sci-fi and maybe a bit worrying. Meta's super sensing AI glasses. Yeah, this one definitely brings up the privacy alarms, doesn't it? They're reportedly working on next gen smart glasses with this super sensing mode. Which means? Continuous real time data collection. Always watching, always listening and potentially using.

Facial recognition. Always on facial recognition in glasses. Okay, yeah, that's a lot. What's the point? What's the benefit they're pitching? The idea is enhanced contextual awareness.

proactive assistance, like the glasses recognize someone, pull up info about them instantly, maybe remind you of stuff based on where you are or who you're talking to. I can see the potential appeal, but yeah. The battery life in trials is apparently terrible in this mode. So there are practical hurdles. And then there are the massive ethical questions, constant monitoring, data use, facial recognition. It's a minefield. Absolutely. Huge questions there.

OK, let's shift again. Fintech. Stripe is getting into the AI foundation model game for payments. Right. They've launched their own payments foundation model trained on, get this, billions of transactions. Billions. Wow. Yeah. Using self-supervised learning. So the AI learns patterns directly from all that raw payment data.

The goal is to make their services smarter. Smarter how? What does that mean for businesses using Stripe? Well, the big one is better fraud detection. They're claiming something like a 64% improvement in catching certain kinds of fraud for big companies. That's significant. It is. Plus, optimizing payment approval rates, making sure legitimate payments go through smoothly, and creating more personalized checkouts. Okay. They also mentioned plans for stablecoin accounts and a new orchestration product.

It really shows how AI is becoming this core intelligence layer for critical infrastructure like payments. Fascinating. Using AI behind the scenes to just make things work better, safer. OK, let's circle back to OpenAI. But this time, they're global plans. OpenAI for countries sounds geopolitical. It definitely has that flavor. It's basically taking their Stargate supercomputer project global.

Stargate. That was the massive U.S.-based AI computer plan, right? Exactly. Now they want to partner with other governments, co-finance, and build sovereign AI infrastructure, local data centers in other countries. So building the hardware abroad, but it's more than just hardware, isn't it? Oh, yeah. The plan includes offering customized versions of their models, like ChatGPT, tailored for local languages, cultural contexts. Their initial pitch is focusing on public services. Think healthcare, education. They're calling it

Democratic AI Rails. Democratic AI Rails. Interesting phrasing. Isn't it? They're aiming for about 10 international projects to start, focusing on democratically aligned nations. It's a big strategic play, expanding their tech's reach, their influence, and shaping the global AI ecosystem around their principles. Yeah. Building a global footprint. Very deliberate. Okay. Something maybe a bit more immediately practical for many of us.

Email. Superhuman is adding AI features. Please tell me it makes email less painful. Uh-huh. That's the goal.

Yeah, they rolled out several AI things. AI email triage, auto labeling to help you prioritize. Okay, that sounds useful. There's an instant reply feature. It generates draft replies based on context, supposedly in your style. Ooh, I like the sound of that. Give it a few bullet points, get a draft back. That's the idea. Plus, automated follow-up reminders and natural language search for your inbox. Nice. The whole point is just cutting down that time suck that is email, right? Freeing you up for other stuff.

A clear example of AI tackling everyday work friction. Definitely needed. Okay. Nearly there. Mistral AI. The European player. They've got new models, an enterprise platform. Sounds like they're making waves. They really are. They just launched Mistral Medium 3. It's a family of models, high performance, especially for coding and STEM stuff. But the key thing is the cost. Reportedly, like...

Eight times cheaper than comparable models from competitors. Eight times. Wow, that's disruptive. It is. And alongside that, they launched LeChat Enterprise. It's their AI assistant platform specifically for businesses. What does the enterprise version offer? Things like search across company data, no-code tools to build AI agents, data connectors, and crucially, flexible deployment, cloud infrastructure.

private cloud, even on premise. Big focus on privacy and control. Okay. So aiming for businesses that want power, but also cost efficiency and control. Exactly. And they also hinted they might open source their biggest model large eventually, even though Medium is closed for now. It positions them really strongly in that enterprise market.

Makes sense. And just quickly, there were a few other rapid fire developments mentioned to you, right? Yeah, a few quick hits. Antropic gave Claude web search via an API so it can browse and cite. Google updated its Gemini 2.0 flash image model. Better quality, fewer restrictions. Netflix added open AI powered natural language search. LinkedIn's got a new AI job search tool. And Ace Studio released a super fast open source music model called AC Step.

Whew. Okay. A lot happening. So let's try and wrap this up. Looking back at all this news from May 8th, 2025,

What are the big themes here? What stands out? I think it's just the sheer pace and the breadth. AI isn't just one thing anymore. It's advancing in foundational models and the hardware in specific apps and enterprise tools and policy. It's everywhere. Yeah, it really feels like it's weaving itself into everything. It is. It's becoming deeply integrated personally and professionally. So for you listening, what's the main thing to take away from all this? Well,

just recognizing how much potential there is here to genuinely reshape things, how we work, how we live, how we interact with tech. It's all potentially up for grabs. Staying aware of these trends isn't just...

interesting. It's becoming kind of essential to figure out what's next. That's a really good point. Okay, so final thought to leave everyone with, something to chew on. Go for it. Given this incredible speed of AI innovation we just talked about, think about a piece of technology you rely on every day, something indispensable right now. What if AI completely

completely transforms it or even makes it obsolete in the next 10 years? And maybe even more interestingly, what totally new things might pop up that we can't even really imagine yet? That's the big question, isn't it? It is. Mull that over. We'll catch you next time for another Deep Dive.