AI Daily News April 01st 2025: 💥OpenAI to Launch its First 'Open-Weights' Model Since 2019 🎬Runway Releases Gen-4 Video Model with Focus on Consistency 🤖Amazon Launches an AI-Powered Browser Agent 🧠AI Instantly Converts Brain Signals into Speech
09:27 Share

Welcome to a new deep dive from AI Unraveled, the podcast created and produced by Etienne Neumann, senior software engineer and passionate soccer dad from Canada. If you're enjoying these explorations of AI, be sure to like and subscribe to the podcast at Apple. Today, we're going back in time, taking a look at a single day in the world of AI. Yeah. March 31st, 2025, to be exact.

a pretty remarkable snapshot of what was going on in the field. Really gives you a sense of just how rapidly things were evolving back then. Exactly. So for you, our listener, we've got a collection of news reports from that day, kind of like a time capsule. Our mission is to unpack it all,

highlight the key developments, and try to understand what it all means. So we've got a lot of ground to cover from big tech moves to some surprising projects that might have flown under the radar. We'll see what's aged well, what's changed course, and what still resonates today. Where should we start? Let's start with something everyone can relate to, healthcare.

Remember Apple's big push into this area? Oh, yeah. They were starting to make some serious waves. Well, on March 31st, 2025, news broke about their AI-powered health care assistant. Unofficially called Health Plus A, right? I remember that. Yes. It was supposed to offer personalized medical guidance, diagnostics, even preventive advice.

all integrated with your Apple devices. So like imagine your Apple Watch not just tracking your heart rate, but actually giving you insights based on that data. Exactly. And it would leverage data from all your Apple devices, wearables, phone, everything, to give you a holistic view of your health. I remember them talking about workout analysis too. Pretty advanced stuff for the time. Yeah, they envisioned analyzing your exercise form,

making sure you were doing things correctly to prevent injuries. And this wasn't just some side project. Oh, no. Apple was taking it seriously. They had a dedicated team of physicians, led by Dr. Sumbul Desai, working on it. And if you recall, Tim Cook had been talking about Apple's health initiatives being their greatest contribution to mankind. Big words. But you can see how Health Plus had that kind of potential to really shake things up in health care. It was a bold vision, no doubt.

Early detection, personalized advice, all at your fingertips. But it also raises some questions that are even more relevant today. Oh, absolutely. Like, how much data are we comfortable sharing? What does it mean for the role of doctors? Those are still debates we're having today, but they were already brewing back then. Okay, so moving from our personal health to, well, the entire internet. We saw some pretty significant news from Amazon that day. They unveiled NovaAct. NovaAct. Ring

Ringing a bell. Wasn't that their AI agent for like controlling your web browser? You got it. It was designed to automate complex online tasks. So finally, an AI to handle those never ending online forms. Potentially, yeah. Think ordering groceries, booking appointments, things like that, all without lifting a finger. Sounds like a dream. How did it stack up against the competition? There were a lot of AI agents popping up around that time. Right. Well, Amazon claimed it outperformed both OpenAI's and Anthropix offerings on their internal tests.

But those tests weren't standardized, so it's hard to say for sure. Ah, yeah. Always got to be careful with those internal benchmarks. Still, the idea of an AI doing our online chores was pretty revolutionary back then. Oh, for sure. It hinted at a future where a lot of our online interactions could be automated. And that raises some big questions for today. Like what happens to the jobs that involve those tasks? How do we design these systems ethically? Exactly. The impact of automation is something we're still grappling with.

Okay, ready for another big development for March 31st, 2025? Hit me with it. It was a big day for AI in drug discovery. Google poured a massive $600 million into their AI pharma venture.

That's a serious investment. What were they focused on? Accelerating drug development using AI. They were leveraging technologies like AlphaFold. AlphaFold. That was the game changer for predicting protein structures. It was. Remember, it even contributed to a Nobel Prize in chemistry. I do. So Google was betting big on AI to transform how we find new medicines. Exactly.

Dimas Asavis, a key figure in the project, even talked about using AI to solve all disease. Ambitious goal, but I guess you need that kind of vision to push the boundaries. And that investment really signaled a shift in the pharmaceutical industry. AI was becoming a central player. What are the implications of that for us, the patients?

Well, potentially faster drug discovery, more targeted therapies, maybe even personalized medicine tailored to your specific needs. OK, let's switch gears again. Remember the merger of X and XAI? That was big news. Huge.

Elon Musk merging his social media platform with his AI company valued at $113 billion. That's combining a lot of power and data under one roof. It was. The idea was to build an everything app, integrating advanced AI into X.

And they were already integrating XAI's Grok chatbot into X using all that user data for training. Right. Musk talked about how the company's futures were deeply intertwined. What kind of innovations were they aiming for? Personalized content, AI-generated content, maybe even entirely new AI-powered services. But it also raised concerns about data privacy competition and the concentration of power. Those are still major issues we're debating today. Okay. Ready for some boardroom drama?

The Chronicle also mentions a new book that came out revealing some turmoil at OpenAI. Oh, that was quite a story. It detailed internal conflicts and power struggles at the very top. What were the main points of contention? It seems the then CTO, Meera Moradi, and co-founder, Ilya Setskiver, clashed with CEO, Sam Altman. Serious clash, right?

They reportedly gathered evidence of what they saw as toxic behavior and dishonesty from Altman. Right. One issue was Altman's ownership of OpenAI's startup fund, which wasn't publicly known. Ah, so potential conflict of interest there. What happened when they brought this evidence forward? Well, Sutzkever presented it to the board, and they actually removed Altman as CEO. Marotti stepped in as interim CEO. But it didn't end there, did it? No.

There was a huge backlash. Employees threatened mass resignations. Wow. So much for board decisions. Right. The board caved under pressure and reinstated Altman. Mirati and Sutskiver ended up leaving OpenAI. Quite the shakeup. What does it tell us about the state of AI back then? It shows that even at the cutting edge, there were these tensions between ethics, profit, and innovation. And those tensions haven't gone away, have they? Not at all.

OK, one more story that caught my eye from March 31st, 2025. It involves Elon Musk again, and it sounds almost unbelievable. Lay it on me. There was a report that Musk was secretly working on rewriting the U.S. Social Security Administration's code base. Hold on. The entire Social Security system using AI? That's what the report claimed. The goal was to improve efficiency and reduce errors. That's either brilliant or terrifying or maybe both. Right.

Modernizing a critical system like that could be a huge win, but it also raises questions about transparency and security. What if something goes wrong? Who's accountable? These are issues we still need to address as AI becomes more integrated into our essential systems. Before we wrap up, I want to tell you about something cool created by the person who makes this deep dive possible. Oh, yeah. Etienne's AI-powered JamGat app.

Exactly. It can help you master and ace over 50 in-demand certifications, cloud computing, cybersecurity, healthcare, you name it. Like having a personal AI tutor to guide you. Links to the app are in the show notes. Check it out after the deep dive. Okay, back to March 31st.

2025. We also saw Google release their Gemini 2.5 Pro experimental model. That was a big one. It topped the El Marina leaderboard, if I remember correctly. It did. And there were whispers of OpenAI working on their own cute computer. Cute. That's not a word you usually hear in tech circles. Right. But that's what Altman hinted at.

And ChatGPT users reported a new thinking slider, giving them more control over the AI's processing. Interesting. Giving users more options is always good. Absolutely. Oh, and Google's Gemini model also scored a whopping 130 on an IQ test. Above average human intelligence, AI was catching up fast. And it wasn't just in the U.S.,

Baidu's Ernie model reportedly defeated OpenAI's GPT-4.5 in Chinese chess. Global competition heating up. Absolutely. And finally, Xtropic AI revealed details about their probabilistic computer chips, claiming massive efficiency games. Potentially revolutionizing AI hardware.

A lot happening in a single day, right? It's mind boggling. Healthcare, web automation, drug discovery, social media, internal AI drama, government systems, new hardware. It really highlights how rapidly AI was and still is transforming our world. So looking back at March 31st, 2025, what's the biggest takeaway for you? I think it's the interconnectedness of it all. Advances in one area fuel progress in others.

And the pace is only accelerating. Makes you wonder what unexpected consequences or synergies might emerge down the line. And how society will adapt to this rapid pace of change. That was our deep dive into March 31st, 2025. A fascinating glimpse into a pivotal moment in AI history. We hope you found it insightful. Don't forget to check out Etienne's AI-powered Jamcac app. Links are in the show notes. It can help you stay ahead of the curve. Until our next deep dive.