AI Daily News June 13 2025: 🤖Mattel and OpenAI team up for AI-powered toys 💥 AMD reveals next-generation AI chips with OpenAI CEO Sam Altman 💰Meta is paying $14 billion to catch up in the AI race 🎬 Kalshi’s AI ad runs during NBA Finals
15:32 Share

Welcome to AI Unraveled. This deep dive is created and produced by Etienne Neumann, senior engineer and passionate soccer dad from Canada. Hey, if you're listening right now, please take just a second to like and subscribe. Honestly, really engaging with these topics, trying to understand the, well, the incredible speed of change in AI. It's just fundamental if you want to navigate what's happening today. And welcome specifically to this deep dive. We've got a

Quite a stack of daily news updates right here. They're all focused on AI innovations and they're all from one specific day, June 13th, 2025.

Our mission really is to cut through all the noise from that single packed day. We want to zero in on the most significant developments, pull out the key insights and really explore what these rapid changes mean, not just for AI in general, but for you. It's definitely a dense set of sources for just 24 hours. We're seeing news covering some really critical ground here. Things like major privacy concerns, ethical debates around new AI products.

The high stakes chip wars that are driving everything underneath, huge investments from the big players and some pretty fascinating creative uses of AI hitting the mainstream. Exactly. It's one of those days where you can easily get lost for hours just reading articles. So our goal here is to slice through that, navigate it together and get you straight to the crucial stuff you need to know fast.

Let's just jump right in with the first big theme that really popped out. Okay, sounds good. Let's start unpacking it. The sources, they really focus early on these concerns bubbling up around the Meta AI app. And we're not talking small things here. According to the sources, privacy experts are, well, they're sounding pretty loud alarms.

The core issue seems to be just how much data this app collects. We're talking voice inputs, maybe location data. And with what's being called minimal transparency. That's the key phrase. Right. Minimal transparency. And what really struck me from the sources here is this

fundamental user experience problem it created. You know, people are using it thinking it was like a private chat, like texting someone or typing in a search bar. But the default setting, well, that was the issue. It seems those interactions were being made public. People were just unknowingly publishing their chats. Wow. That sounds like a recipe for disaster.

The examples mentioned in the sources are pretty chilling, actually. Users accidentally sharing really sensitive stuff. Yeah, things like home addresses, details from court cases, even like questions that could look incriminating later, all just out there because the user had no idea the chat wasn't private.

Imagine typing something deeply personal, expecting just a helpful AI answer, and then finding out it's, well, public knowledge. It really throws a spotlight on a critical challenge, doesn't it? As these AI tools get woven deeper into our daily lives, this lack of clear, really intuitive privacy settings...

it's going to face immense scrutiny. You can't just bury it. Absolutely. It needs to be obvious. And, you know, speaking of AI getting integrated into maybe surprising parts of life, the sources also brought up a related ethical point.

this collaboration between Mattel, the toy company, and OpenAI. That's right. Yeah. The Barbie folks and the ChatGPT folks, they're teaming up. The idea is they're co-developing smart toys using natural language processing. The stated goal, anyway, is creating educational and imaginative interactions for kids. So we're talking about toys that could potentially understand what a child says and respond dynamically. Yeah. Just...

pre-recorded phrases. Exactly. Telling unique stories, maybe helping with homework, that kind of thing. The sources say the first product might even be out later this year. And it covers both physical toys and digital stuff across Mattel's whole range. And there's this other detail sometimes missed. Mattel's own employees are

also getting access to chat GPT enterprise internally. Oh, interesting. So it's not just the toys. No, it shows how this AI is seeping into the corporate side too. Maybe for design marketing, who knows? Now, the companies, naturally, they're stressing safety, age appropriateness, Mattel keeping IP control, which all sounds good. But the sources also highlight the potential flip side.

What this means or could mean is, well, a fundamental change in how kids learn and play. And right alongside that potential upside are some pretty significant ethical flags being raised. You know, introducing this kind of sophisticated data collecting tech into a child's environment. It brings up serious questions about surveillance.

data privacy for minors, the psychological effects. Interacting with an AI personality during those formative years. Exactly. It's a really tricky area. It's going to need a very careful thought and probably new rules down the line. Definitely makes you pause. Okay, let's shift gears quite dramatically here because AI isn't just about apps and toys, right? It runs on incredibly powerful, incredibly expensive hardware. And that brings us to the, well, the escalating arms race in the tech world.

Indeed. And the sources had a pretty big piece of news on that front. The AMD and OpenAI announcement. Yeah, this wasn't just your standard chip launch.

AMD unveiled their next-gen AI chips, the Instinct MI400 series. But the real kicker, highlighted in the sources, was OpenAI CEO Sam Altman being right there on stage with AMD. Yeah, co-announcing it. Confirming that OpenAI is actually going to use these new processors. That joint appearance speaks volumes, doesn't it? It signals this isn't just AMD hoping for a sale. It's a validated partnership. And they talked about this Helios server rack.

concept, basically thousands of these MI400 chips working together like one giant compute engine. Think AI supercomputing on a massive scale. And OpenAI wasn't just signing off on the final product.

The sources said they actually gave direct feedback during the MI400's development. That suggests a really deep strategic collaboration. So looking at the implications, what this means really is a couple of things. First, it seriously ramps up the AI chip war. AMD is making a major play against Nvidia's dominance. Second, it shows OpenAI is actively diversifying who they get their hardware from, spreading the risk, gaining leverage, maybe getting better access. Smart move. Makes sense in a market where these chips are like gold dust.

and this focus on huge compute power, it ties directly into the money side of AI.

The sources also detailed Meta's just staggering investment, $14 billion in AI. $14 billion. It's an almost unbelievable number for one company in this time frame. The sources kind of frame it as Meta really trying to hit the accelerator, trying to catch up in some areas, even though there were reports of them losing some top AI researchers. Right. Seems like a massive bet to make if talent is leaving. But it just shows how vital they think this is. So where's that money actually going?

The sources mentioned their superintelligence group and specifically custom hardware development. Which probably links back to needing chips like AMD's or building their own. And they're hiring like crazy too. Alexander Wang from Scale AI is leading a new lab focused on superintelligence reporting straight to Zuckerberg. Wow. Plus there was mention of a massive new investment in Scale AI itself and Zuck apparently personally recruiting top people with huge pay packages. It sounds like a total full court press.

The sources explicitly linked this big push to how their last big model, Llama 4, performed. Apparently, it didn't quite hit the mark. Yeah, seems that way. So Meta's stated goal now is, well, it's incredibly ambitious, building full general intelligence AGI and what they're calling its leading personal AI. So

So the takeaway here seems to be that Meta spending $14 billion just highlights how absolutely critical AI dominance is seen by these tech giants. It's not just tweaking things. It's a massive strategic gamble. Exactly. It shows you need incredibly deep pockets, huge ambition, and the ability to grab the best talent and the best hardware to even compete at this level. And, you know, understanding the whole competitive landscape, seeing where the money's going, knowing about the underlying tech like these chips, right?

That's really crucial knowledge. It's the kind of foundation you need, whether you're actually building AI or maybe looking to, say, boost your career by getting certified in these areas. Understanding the battlefield is key. That's a really good point. The tech and the strategy, they're completely locked together.

OK, so speaking of building with AI, let's look at how it's starting to change creative fields, sometimes in really dramatic ways. We saw a fascinating example in advertising mentioned in the sources. Kelshi's AI generated ad during the NBA finals. Yeah, that felt like a significant moment. A national primetime TV ad slot, right? And the ad was entirely AI scripted and AI voiced.

No traditional actors, no film crew as we know it. And get this, the sources gave some incredible details. An AI filmmaker, PJ Exituro, apparently put the whole 30-second spot together in just two days. Two days? Yeah. He generated something like 300 to 400 different video clips using VO3 just to get the final 15 or so he needed.

And he used tools like Gemini and ChatGPT for everything, brainstorming, script writing, even crafting the prompts for the video generator. He was apparently using VO3's newer feature where the characters can speak, though the sources did note some glitches, like weird subtitles popping up or the voices not being consistent between shots. Still early days for some features, maybe. But here's the kicker, the really disruptive part.

Exeter estimated the cost was around 95% less than a traditional ad shoot for something similar. 95%. Wow. And his prediction, he thinks these high dopamine VO3 videos will be the ad trend of 2025.

Seeing it on national TV, you could think he might be right. The implications for advertising are just huge. Generative AI is officially playing in the big leagues, primetime marketing, the speed, the cost savings, the ability to iterate so quickly. You have to think more brands are going to jump on this. It could completely change how ad agencies and production companies work. And this speed isn't just happening in ads. The whole field of generative video is moving incredibly fast.

The sources highlighted another big player making waves, Bykedance TikTok's parent company, with their new AI video generator, Seedance. Yes, Seedance. That's the name suddenly shooting up leaderboards. Looks like a really serious competitor to models like Sora that got so much attention earlier. The sources flat out said Seedance 1.0 has now hit the top spot on the artificial analysis video leaderboards.

At a VO3, Cling 2.0, even Sora, that's a meteoric rise. And the specs sound impressive too. Generating 5-second clips in 1080p and understated

in under a minute. And apparently it has key features like multi-shot storytelling, keeping characters consistent across shots, smooth transitions, things you really need for actual video production. ByteDance even made their own benchmark, Seed Video Bench, which, surprise, surprise, they claim shows Seed Dance is better than others in motion quality, sticking to the prompt

and just looking good aesthetically. And their plans are pretty clear. They're going to fold C-Dance into their existing apps, their Dubow chatbot, and their Jemang video platform later this year. Right. So the core insight here is TikTok.

TikTok's parent isn't just building good AI tech. They're poised to dominate how it gets used and distributed. With their massive user base, they could reshape the whole generative video landscape very quickly. It really shows that blistering pace again, doesn't it? New models jumping ahead almost month by month. It absolutely does. And it just reinforces why, you know, staying current, continuously learning, actually building skills with these rapidly changing tools. It's essential for anyone working in AI or even thinking about getting into it. Definitely a race.

Exciting one, though. OK, before we wrap up this deep dive, the source has actually covered a ton of other AI news just from this one day, June 13th. We should probably just quickly mention a few to give a sense of the sheer breadth of activity. Right. Just a quick snapshot. ByteDance researchers also introduced something called SeaLead APT2, a different model aimed at real time interactive video streams.

up to five minutes long at 24 frames per second, moving towards live AI video. Microsoft started rolling out co-pilot vision highlights in the U.S., basically letting co-pilots see your screen to give you better help. Google DeepMind launched its Weather Lab, using AI for hopefully better storm predictions. Apple is apparently targeting spring 2026 for major AI upgrades to Siri. That's almost two years after their big announcement at WWDC 2024, which...

feels slow compared to some of this other stuff. Interesting timeline, yeah. Runway released chat mode for its creative tools, letting you just describe what you want in chat, and it makes it. And connecting back to science, Los Alamos, Meta, and Berkeley Lab released Open Molecules 2025.

It's a huge data set, over 100 million molecular simulations, specifically to train AI for chemistry and drug discovery research. It's just incredible the volume and variety packed into a single day. It really is. Okay. Okay, so for a final maybe provocative thought, something that popped up in the sources that really pushes the envelope. Right. There's this claim mentioned coming from scientists in China. They reportedly state that their AI systems have achieved...

human level cognition, that these systems have spontaneously developed reasoning abilities like ours. That is quite a claim.

And the sources point out how significant it would be. What this means, if it's validated, is potentially enormous. Not just another step, but maybe a true paradigm shift in AI globally. And strategically, of course, it would be a massive boost for China's AI ambitions. It would change the entire conversation. So this really raises a big question maybe for you, the listener, to think about. When you see the incredible speed we've discussed just today, from superpower chips and AI ads on TV, all the way to these kinds of...

claims about human level reasoning. What does this rapid evolution really mean for how we even understand intelligence itself? And maybe more personally, how does seeing this pace just from one random day in June 2025, how does it change your own perspective on what's possible and maybe what we should expect in the well, the very near future?

It's definitely a lot to chew on. And those are the key insights we've pulled from the sources provided for this deep dive into AI news from June 13th, 2025. We covered a lot, privacy concerns, ethics, the hardware race, huge investments, creative disruption, quick hits across science and software, and ended on that really startling claim. Yeah, we hope this discussion helped highlight what was most significant in today's sources and maybe spark some new thinking for you.

This has been a deep dive on AI Unraveled, created and produced by Etienne Newman. Now, if today's talk about just how fast AI is moving has you thinking about where things are going and how you can get involved, well, understanding these technologies is absolutely key. Etienne has put together some fantastic resources specifically to help you get certified in AI, boost your career, and really learn how to build with these tools.

You should definitely check out his AI certification prep books over at DJAMGATE.com. That's D-J-A-M-G-E-C-H dot board com.

He's got guides covering Azure AI Engineer Associate, Google Cloud Generative AI Leader Certification, AWS Certified AI Practitioner Study Guide, Azure AI Fundamentals, and the Google Machine Learning Certification. Really comprehensive stuff. And also, don't miss Etienne Newman's toolkit. It's called the AI Unraveled Builders Toolkit. It's full of practical AI tutorials, PDFs, audio, video, plus certification guides. It's really designed to help you get hands-on and actually start building with AI right away. Yep.

And all the links for those books in the toolkit are right there in the show notes for this deep dive. Super easy to find. Thanks so much for joining us for this exploration. Keep asking questions, keep learning, and we'll dive deep again very soon.