AI Daily News Jan 21 2025: 🤖DeepSeek’s Open-Source R1 Beats OpenAI o1 🐍AI-Designed Proteins Tackle Snake venom🤖Humanoid Robots Assemble iPhones in China 🧬UK’s Supercomputer Develops AI Vaccines 🖥️OpenAI’s ChatGPT Crawler Vulnerability Revealed
14:52 Share

Hey everyone, welcome back to the Deep Dive. Today we are going to be looking at a pretty wild day of AI news, January 21st, 2025. We've got a lot of news to get through from open source, AI, robots getting real jobs, AI designing vaccines, and even a presidential policy flip-flop. Oh, wow. On AI safety. So a lot to unpack, but that's why we're here. Yeah, it's really a fascinating snapshot of a field that's just

changing so rapidly. Absolutely. All right. So first up, let's talk about this bombshell out of China. Okay. DeepSeek, a company you might not have heard of. Yeah. Just drop DeepSeek R1. It's this open source AI model and get this. Okay. It's going toe to toe with OpenAI's O1. Really? And some seriously complex tasks because we're talking math, coding, even logical reasoning. It's pretty nuts.

That is impressive. And what's interesting to me is not just that it performs well, but that it's also doing all this at a fraction of the cost of what we're used to seeing. What kind of cost are we talking? We're talking like 90 to 95 percent cheaper to run than some of these other models. So this is a big deal potentially for researchers, startups, anyone who needs serious AI horsepower, but can't afford the big players like Google or Amazon.

Microsoft or whoever. Because this isn't just some tech upgrade. This could actually shift who gets to even use AI, right? Exactly. It could totally democratize access to these powerful tools. Open AI kind of ironically started with open source ideals, but now they've become very closed and proprietary. So this move by DeepSeat could really spark a renaissance of open development and lead to faster innovation, more diverse applications, just really shake up the industry. Yeah. Okay.

Definitely something to keep our eyes on. Yeah. But let's move from the digital world to the physical world for a second. Okay. Because it turns out robots are now officially off the demo stage and on the factory floor. What? We're talking like the iPhone assembly line. Oh, wow. UB Tech's Walker S1 robots are actually working alongside humans in Foxconn factories in China. I hadn't heard about that. Yeah. So imagine this, a robot about 5'6", 167 pounds.

And it's doing things like quality inspections, assembling components, all this stuff that used to be exclusively human domain. That's pretty incredible to think about. And it's not just iPhones either. We've seen figure robots working for BMW. Really? Apptronic robots at Mercedes plants. So it seems like we're at the very beginning of this rapid automation wave. Yeah, it really does feel like it's all happening so fast. And it makes you wonder, are we prepared for the changes that are coming?

that are coming. Exactly what happens to the global workforce. How do supply chains adapt? These are huge questions. Absolutely. Things we definitely need to be thinking about now and trying to get ahead of. Yeah. All right. Before we go too deep down that rabbit hole, let's switch gears to something a little bit more optimistic. A,

AI being used to will potentially save lives. Oh, okay. That sounds good. We've got some news out of the UK about an AI-powered supercomputer that's being used to actually design vaccines. So the Isambard AI supercomputer is a really fascinating example of AI being used for good.

It can virtually test millions of potential drug combinations to find the most promising candidates. And it significantly speeds up the process of developing new vaccines and treatments. That's amazing. So what kind of things can it target? I mean, think about emerging diseases, chronic illnesses, even things like Alzheimer's. Wow. Hold on. Millions of combinations. How does it even do that? I mean, drug development normally takes years.

Even decades sometimes. Yeah, so think of it this way. Traditional methods are like trying to find a needle in a haystack by hand.

Eisenbard AI is like using a powerful magnet to pull the needle straight out. Like it's just a complete game changer. That's a great analogy. And it's incredibly energy efficient, too. Oh, really? The waste heat from the supercomputer is actually being used to heat homes and businesses in the local area. That's amazing. Talk about a win-win. Wow. So sustainable and potentially life-saving. Absolutely. All right. So it's not all good news on the policy front. Former President Trump actually made a pretty controversial move.

He revoked an executive order from the Biden administration that was all about trying to address AI risks. Oh, wow. What does that mean practically? Well, it's definitely a setback in terms of establishing clear guidelines and regulations for AI development, especially.

when it's advancing so rapidly. Right. It's like building a super powerful engine with no safety features. Exactly. It's exciting, full of potential, but also carries a huge amount of risk. Yeah, I see what you mean. And this feels especially relevant given the open source news that we were just talking about, right? Does a lack of regulation help or hurt that movement? Yeah, it's a tough question. It's very complex. I mean, without those guardrails,

development could go in a lot of different directions. Right. Both positive and potentially really harmful. All right. We've got to keep moving. Okay. Because we're just getting started with the news. So there are actually some concerns about security vulnerabilities that are popping up. OpenAI's chat GP key crawler, you know, the one they use to gather data from the web. Well, it turns out it has a weakness that could be exploited for these things called DDoS attacks. What are those?

Basically, they can shut down websites. Oh, wow. That's not good. And on a similar note, AI analysis is showing us that some of the data used in scientific research

is actually really inaccurate. Like shockingly so. And this is a huge problem because AI is only as good as the data it's trained on. Right. Garbage in, garbage out, as they say. Exactly. So if the data is flawed, the results could be too. And that has implications for everything. Yeah. I mean, everything from new discoveries to actual product development. Wow. Okay. So we've got robots, we've got vaccines.

We've got vaccines, security concerns, political maneuvering. Yeah. And we're just getting started. And it's only part one. I know. It's crazy. It really feels like AI is really touching every aspect of our lives now. It is. And it's only going to accelerate from here. It's wild to think about. Absolutely. All right. We're going to take a quick break.

And come back with even more AI news. So don't go anywhere. And it's not just like the big splashy headlines, you know, there are all these smaller, almost under the radar things happening that I think are just as significant in the long run. Yeah. Give me an example.

I'm still trying to wrap my head around robots making iPhones. Well, take like developer tools, for instance. Sure. ByteDance, the company behind TikTok, just released this thing called Trey. It's an AI coding assistant. Oh, wow. Specifically for Mac OS. And then you've got Codium. They've got a new version of their Windsurf platform. It's got like web search built right in. Okay. And it can learn coding patterns automatically. So it's like AI is becoming like a developer's best friend.

helping them write code faster and better yeah exactly i mean does this mean human developers are going to be out of a job soon i don't know it's a good question i don't think it's about replacing them it's more like augmenting their abilities you know okay these ai tools can handle like the more tedious parts of coding right the repetitive stuff so the humans can focus on the more creative strategic stuff yeah that makes sense it's like having an ai apprentice

Who can handle the grunt work while the human master focuses on the big picture. Exactly. It's all about collaboration, not replacement. Okay. So we've got open source AI robots in factories, vaccines, being designed by supercomputers, AI assistance for developers. I mean, is there anything AI isn't involved in these days? Well, that's the thing, right? It really does feel like it's infiltrating every aspect of our lives. And one of the most exciting areas, I think,

is multimodal AI. Multimodal AI. Yeah, have you heard about this? I remember reading about Moonshot AI and their Kimi K1.5 model, which was supposed to be able to analyze both text and visual data. That's the one. So why is that a big deal? I mean, it sounds cool, but what are the implications? Well, imagine like self-driving cars.

That can not only understand the rules of the road, but also interpret the visual environment around them. You know, pedestrians, cyclists, traffic lights, all that stuff. Or what about medical imaging, where AI could analyze a patient's medical records. Right.

and their scans to give like a more accurate diagnosis. Yeah, that would be incredible. So multimodal AI is like the key to unlocking these kinds of possibilities. Exactly. It's like giving AI a whole new set of senses. Yeah. Allowing it to perceive the world in a much richer and more nuanced way than ever before. It feels like we're on the verge of some major breakthroughs. I think we are. And that's both exciting and a little bit scary.

don't you think? Yeah, for sure. I mean, it's amazing to think about all the potential benefits. Yeah. But it also raises a lot of questions about the future, right? Right. I mean, what happens to jobs, to privacy, even to our sense of

what it means to be human. Exactly. In a world where AI is increasingly capable. These are questions we all need to be asking ourselves. And more importantly, discussing as a society, you know, the future of AI isn't something that just happens to us. It's something we have a responsibility to shape. It's like we're all in the driver's seat now. And we need to be paying attention to the road ahead. You know, it's funny you brought up self-driving cars because it made me think about

That news about former President Trump revoked that executive order on AI safety. I mean, that decision could have huge ripple effects, especially in fields like autonomous vehicles. Right. Absolutely. The executive order was supposed to establish guidelines for responsible development, you know, promoting things like transparency and accountability. So by repealing it, it's basically saying.

That those issues aren't a priority. And that could lead to a more chaotic and potentially dangerous development landscape. Yeah, exactly. I mean, without those guardrails in place, it feels like companies could be tempted to cut corners, prioritize speed over safety, maybe even push AI into areas where it's not quite ready yet. And it's not just self-driving cars. Think about AI in health care.

In finance, in law enforcement. I mean, these are all areas where mistakes or biases in AI systems could have really serious consequences. It feels like we're at this crucial turning point. You know, we've got this incredible technology with the potential to revolutionize so many aspects of our lives. But we still haven't figured out the rules of the road. Yeah. And how do we make sure...

that this technology is used for good and not for harm? I mean, that's the big question. Well, I think it starts with having these conversations now. While we still have the opportunity to shape the future of AI, we need to be pushing for regulations that promote safety and accountability. And we need to be holding developers and policymakers accountable for the choices they make. I couldn't agree more. We can't just sit back

and let this technology unfold on its own right. We need to be proactive. We need to be engaged. We need to be demanding that AI is developed and used in a way that benefits all of humanity, not just a select few. Well said. All right, before we move on to the final part of our AI News Marathon, what's your biggest takeaway from all of this? What's really sticking with you? Honestly, it's the contrast between the incredible potential of AI

and the lack of clear guidelines for its development. I mean, we've got AI designing lifesaving vaccines, helping developers write better code, but at the same time, we're seeing vulnerabilities that could be exploited for cyber attacks and a lack of political will to really address the potential risks. It's like we're simultaneously witnessing the best and the worst of what AI has to offer. Exactly. And we need to find a way

to amplify the good while mitigating the bad. And that's going to require a much more thoughtful and nuanced approach to AI development than what we're seeing right now. I couldn't agree more. All right, we're about to enter the final stretch of this AI news journey. Stay tuned for part three where we'll wrap up our discussion and leave you with some food for thought. Okay, we are back for the final stretch of our AI news marathon and...

I've got to say my head is kind of spinning after all of that. Yeah, it's a lot to process. We covered so much, you know, from open source breakthroughs to robots on factory floors to

to the incredible potential of multimodal AI. It really highlights just the sheer breadth and depth of this field and how fast it's all moving. It does, and I think it brings us back to this underlying tension we've been kind of circling all episode, right? This incredible potential of AI to really improve our lives, but also the very real risks that it poses if we don't develop and use it responsibly. Right, like that's this...

tightrope walk trying to balance those two sides. Exactly. And we've seen examples of both today. You know, on one hand, we've got AI designing life-saving vaccines, helping developers write better code. But on the other hand, we're seeing these security vulnerabilities that could be exploited, you know, by bad actors. And there's this lack of clear regulations to guide all this, this rapid development. And it makes you wonder, like,

Is it just up to the tech experts and the policymakers to figure this out or do we all need to be part of this conversation? Yeah, I think it's definitely bigger than just them. I mean, we all have a stake in the future of AI, right? Absolutely. It's going to impact all of our lives in really profound ways, whether we realize it or not. So what can like everyday people actually do to ensure that AI is used for good? I think a lot of people just feel overwhelmed.

by how complicated this all is it is complex but there are definitely things we can all do you know stay informed about the latest developments um support organizations that are promoting ethical ai right advocate for policies that protect our rights and our privacy and i think most importantly just be mindful of how we're using ai in our own lives it's a good reminder that we're not powerless in all this like our choices actually do matter every click every purchase

Every interaction we have with an AI system, it's all shaping the direction of this field. So we need to be conscious of that power and use it wisely. So as we wrap up this deep dive, I want to leave you with this final thought. We've seen some incredible advancements today. Okay.

It's important to remember that we're still really in the early stages of this whole AI revolution. That's right. So the next few years, I think, are going to be absolutely critical in determining how AI shapes our world. I agree what we do now, the choices we make, the conversations we have, they're going to have a ripple effect.

for generations to come. So to our listeners, the question is, what kind of future do you want to see? And what role are you going to play in making it happen? Thanks for joining us on this whirlwind exploration of AI news. Until next time, stay curious, stay engaged, and stay informed.