AI in 10
The most important AI story—explained in 10 minutes.
Every day, I break down the biggest AI story in just 10 minutes - what it is, why it matters, and how you can actually use it. No tech jargon, just AI made simple.
AI in 10
China Drops Trillion-Parameter AI Without Nvidia Chips
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Referenced Links:
DeepSeek Official Website
DeepSeek Models on Hugging Face
LM Studio - Run AI Models Locally
Anthropic Official Website
LocalLLaMA Community
Want to go deeper with AI? A community of professionals is learning AI together right now at aihammock.com — show notes, links, tools, and real conversations about how to actually use AI in your life.
Welcome to AI in 10. I'm Chuck Getchell, and every day I break down the biggest AI story in just 10 minutes. What it is, why it matters, and how you can actually use it. Here's something wild happening in the AI world right now. A Chinese company just announced they're about to drop a trillion parameter AI model. And here's the kicker. They built it without using a single Nvidia chip. This is like building a Formula One car without buying parts from Ferrari. It shouldn't be possible, but somehow they pulled it off. The company is called DeepSeek, and they just teased their V4 model on March 6th. Now, when I say trillion parameters, think of that as the AI's brain cells. For context, that's bigger than most of what OpenAI and Google are running. But here's where it gets interesting. The US has been blocking China from buying our best AI chips for years. So DeepSeek said, fine, we'll build our own. They use Chinese chips from companies like Huawei and CamberCon instead. This is huge for several reasons. First, it proves China doesn't need our hardware to compete. Second, it's about to make AI way cheaper for everyone. And third, it's causing some serious drama in the tech world. Let me break down what actually happened here. DeepSeq didn't just announce this model. They're claiming it can handle text, images, and potentially video. All while running at a fraction of the cost of Western competitors. Think about what you pay for ChatGPT. 20 bucks a month, right? DeepSeek is talking about running queries for literally pennies. We're talking about AI that costs a hundred times less to operate. Now here's where the accusations start flying. Dario Amode, the CEO of Anthropic, that's the company behind Claude, he's basically calling DeepSeek cheaters. He says they're using something called industrial scale data distillation. That's TechSpeak for training your AI by feeding it outputs from other companies' models. It's like copying someone's homework but with billion-dollar AI systems. DeepSeq says they're just being innovative. Anthropic says they're stealing intellectual property. Both sides have valid points, honestly. But let's step back and look at the bigger picture here. This isn't just about one AI model. This is about a fundamental shift in how AI gets built and who controls it. For years, Nvidia has been the king of AI chips. Their hardware powers everything from Chat GPT to Google's Bard. They've been printing money because everyone needs their chips to build competitive AI. Now, DeepSeek is saying you don't need NVIDIA at all. You can build world-class AI on completely different hardware. That's like someone proving you can win the Indianapolis 500 with an engine Toyota built in their garage. So what does this mean for your daily life? Well, buckle up because things are about to get interesting. First, AI is about to get ridiculously cheap. When DeepSeek releases this model and they're planning to make it open source, anyone can run it. That means the apps on your phone could soon have AI capabilities that rival Chat GPT, but for free. Imagine editing photos with AI, getting real-time translation, or having a personal tutor for your kids, all without paying monthly subscription fees. We're talking about potentially saving your family$200 a year on AI tools. Second, your devices are about to get way smarter. Because this AI is designed to run efficiently, it could actually work directly on your laptop or phone. No internet required. That means faster responses and better privacy. But here's the flip side: this escalates the tech cold war between the US and China. And that could mean some uncomfortable choices ahead. If Chinese AI becomes dominant because it's cheaper and more accessible, where does that leave your data privacy? Do you trust Beijing with your personal information the same way you might trust a US company? That's a question every family will need to answer. There's also the job impact to consider. When AI becomes this cheap and powerful, it accelerates automation everywhere. The grocery store checkout, customer service calls, even some office work, it all becomes more economical to automate. But here's my take on this, and it ties directly into everything I write about in the comeback code. This isn't a reason to panic, it's a reason to get ahead of the curve. The people who are going to thrive in this new world are the ones who start learning how to work with AI now, not in five years when everyone else figures it out. Right now. So let me give you something concrete you can do today. You don't have to wait for DeepSeek V4 to drop. They already have models you can try for free. First go to deepseek.com and sign up for their current V3 model. It's free and you can test it against Chat GPT. Try asking it to write a simple budget spreadsheet or help plan your next vacation. See how it compares. Here's something even cooler. If you have a decent laptop, you can actually download and run some of their models locally. No internet required. Look up something called LM Studio. It's free software that lets you run AI models on your own computer. Download Deep Seat Coder and ask it to write a simple program for something you need. Maybe a tool to track your family's chores or organize your photo collection. This isn't about becoming a programmer, it's about understanding what's possible when AI costs almost nothing to run. The actionable step here is diversification. Right now, most people who use AI only use Chat GPT or maybe Claude. Start spreading your usage around. Try the Chinese models, try the European ones, get comfortable with different AI tools. Why? Because this market is about to fragment. The US might restrict access to Chinese AI. China might restrict access to American AI. The more tools you're comfortable with, the better prepared you'll be. Set up alerts for when DeepSeek V4 actually releases. Follow them on Twitter or join communities like the local El Lama subreddit. When it drops, you want to be among the first to try it. Now, let me address the elephant in the room. Is this data distillation thing actually cheating? Honestly, it's complicated. In the AI world, everyone learns from everyone else's work. That's how progress happens. But there's a line between inspiration and copying, and reasonable people disagree on where that line sits. What I find interesting is that this accusation is coming now, right when Chinese AI is becoming genuinely competitive. For years, when China was behind, nobody cared much about their methods. Now that they're winning, suddenly everyone's concerned about fair play. That's not to say the concerns aren't valid, they might be. But it's worth noting the timing from your perspective as someone trying to navigate this changing world. Here's what matters the technology exists. The genie is out of the bottle. Whether DeepSeek cheated or not doesn't change the fact that trillion parameter AI models are about to become commoditized. Your job is to figure out how to benefit from that reality. Not to worry about the politics behind it. As I always say, I'm not a lawyer or a policy expert. Always think through the privacy and legal implications for your specific situation. But from a practical standpoint, this technology is coming whether we like the methods or not. The broader trend here is clear. AI is moving from expensive centralized systems controlled by a few big companies to cheap distributed systems that anyone can run. That's incredibly empowering if you're ready for it. It means you won't be dependent on subscription services or corporate policies. You'll be able to run your own AI on your own terms, but it's also potentially destabilizing. When technology this powerful becomes this accessible, it changes everything faster than institutions can adapt. The winners will be the individuals and families who start building AI literacy now. The ones who understand how to use these tools to enhance their work, education, and daily lives. The losers will be the ones who wait for their employers, schools, or governments to figure it out for them. This Deep Seek announcement is just the beginning. We're about to see an explosion of competing AI models from countries and companies all over the world, each one trying to prove they don't need Silicon Valley's permission to build the future. That competition is great news for users like you and me. More options, lower prices, better capabilities. But it also means more complexity to navigate. The key insight here isn't technical, it's strategic. We're moving into a world where AI capability is becoming divorced from AI control. The companies that build the models won't necessarily be the ones that decide how you use them. That puts more power in your hands, but it also puts more responsibility there too. You'll need to make smarter choices about which tools to use, how to protect your privacy, and how to stay ahead of rapid changes. The Deep Seek V4 story is really about something bigger than one AI model or even the US China Tech rivalry. It's about the democratization of artificial intelligence. And that's either the best thing that could happen to individual empowerment or the most chaotic. Probably both. That's today's AI Inten. If you want to go deeper and learn AI with a community of people just like you, join us at aihammock.com. I'll see you tomorrow, my friends.
