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
25 Tech Giants just drew a line on open AI
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Referenced Links:
Ollama — Run Open-Weight AI Models Locally
LM Studio — Beginner-Friendly Local AI Tool
Mistral AI — Open-Weight Model Developer and Signatory
AI Hammock — Applied AI Certification for Non-Technical Professionals
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. 25 tech companies, some of them bitter rivals, just agreed on something, and what they agreed on could determine whether you ever get to run AI on your own computer again. I'm Chuck Getchell. This is AI in 10. What happened? Why it matters, what you can do with it. Let's go. Yesterday, a group of 25 major technology companies published a joint open letter to US policy makers. The signatories include NVIDIA, Microsoft, Meta, Mistral, and a roster of other AI labs and cloud platforms. Together, they're drawing a line in the sand on one specific issue: open weight AI models. And if you've never heard that term before, don't worry, we're gonna break it down completely because this story touches everyone who uses AI tools or plans to. So first, what exactly is an open weight AI model? Think of an AI system as having two main parts. There's the interface, the chatbot or app you interact with, and then there's the actual brain underneath the mathematical structure the system learned from training on billions of data points. That brain is called the model weights. It's essentially a massive file of numbers that defines how the AI thinks and responds. A closed model keeps those weights locked up. You can only access the AI through a company's app or API. You use it their way, on their servers, under their terms. Think of it like renting an apartment. You can live there, but you don't own the walls. An open weight model publishes those weights publicly. Anyone can download the brain, run it on their own computer, customize it, inspect it, improve it. That's more like owning a house. You can knock down a wall if you want to. Now here's why this matters right now. Washington is debating whether to restrict open weight models, especially the most powerful ones. The argument goes something like this: if a foreign adversary can download the same powerful AI that American Labs built, that's a national security problem, especially if they fine-tune it for cyber attacks or worse. That's a real concern. It's not invented, and it deserves a serious conversation. But here's where the coalition of 25 companies pushes back. They're saying the cure might be worse than the disease. Their letter argues that blanket restrictions on publishing model weights would do several damaging things at once. It would crush independent security researchers who use open models to find flaws before bad actors do. It would gut academic study of AI systems. It would force businesses, including hospitals, law firms, local governments, and small startups, back into a world where all their AI runs through a handful of massive cloud providers. And it would ironically make American AI more fragile, not less. By concentrating everything into a few choke points. That last point is worth sitting with for a second. Putting all of American AI infrastructure behind three or four corporate gates and calling it a security strategy is a little like locking every door in town but giving one person all the keys. It feels safe until that one person drops the keychain. The companies aren't saying there should be zero rules. That's important. They're not asking for a free-for-all. What they're asking for is a risk-based approach. Regulate the dangerous uses. Regulate specific high-risk applications like autonomous cyberweapons or tools designed to assist with biological threats. Don't regulate the act of publishing weights itself. There's a historical parallel here that the letter leans on, and I think it's a good one. Back in the 1990s, the US government treated strong encryption like a weapon. They tried to restrict its export. They called it a national security risk. And over time they lost that battle because encryption was too useful, too embedded in commerce and communication to contain. Today the open internet runs on the very encryption tools the government once tried to restrict. The letter essentially says open weight AI is the encryption fight of our era. Get the policy right early or spend years walking it back. Now let's bring this home because this might sound like a Washington policy debate that has nothing to do with your Tuesday, but it actually has a lot to do with your Tuesday. Here's the first way this touches you directly. Cost and competition. Open weight models create market pressure. They give developers an alternative to the big closed platforms. That competition is a big reason AI tools have gotten cheaper and better at a remarkable pace. If open weights get restricted, the innovation engine slows down and your options narrow. Fewer tools, fewer providers, and probably higher prices. Competition is great for consumers, always has been. Second, data privacy. This one is underappreciated. A lot of individuals and small organizations think small clinics, solo attorneys, local nonprofits have started using open weight models because they can run AI locally on their own hardware without sending sensitive client or patient data to some cloud server. That's not paranoia, that's reasonable professional practice. If those local options disappear, those same organizations get pushed back to big cloud APIs, whether they like it or not. Their data goes with them. Third, small business and entrepreneurship. There's a whole ecosystem of freelancers, startups, and solo developers who build products on top of open weight models because they're dramatically cheaper than licensing from the big players. These are the people building your local neighborhood app, your small business's customer service bot, your church's volunteer coordination tool. That's not a Silicon Valley thing, that's Main Street. Heavy restrictions make it much harder for the little guy to compete. And here's the bigger philosophical point because I think it's worth naming. For the last couple of years, AI has been getting more democratized. More people with less technical background have gained access to genuinely powerful tools. That's a remarkable thing. It's not finished. There's still a huge gap between what a casual user can do and what a well-resourced enterprise can do, but that gap has been closing. Open weight models are part of why. They put capability in more hands, they allow more independent eyes on AI systems, they distribute power rather than concentrating it. The question being debated in Washington right now is whether that trend continues or reverses. I'll say this: I don't think waiting for institutions to figure this out is a strategy. The people who will thrive in whatever regulatory environment emerges are the ones who understand AI well enough to adapt fast. That's always the move. So here's your one actionable thing for today. It's simple but genuinely useful. Try a local or open weight AI model this week. Even just once. You don't need to be a developer to do this. Tools like LM Studio and ALAM let you download and run open weight models on a regular laptop with no coding required. These are free tools. You can have a capable AI assistant running entirely on your own machine within about 30 minutes. Why does this matter? There's a few reasons. One, you'll instantly understand what the whole debate is about. Because you'll feel the difference between talking to a cloud-based AI and running one locally. Two, you'll know what you'd lose if open weight models became restricted. Three, you'll have a new skill and a new tool that a surprisingly small percentage of people have bothered to explore yet. When you try it, start with a small open weight model. Something in the 7 to 13 billion parameter range runs well on most modern laptops. Download Olima, it's free, it's beginner friendly, and they have a clean interface. Pick a model called Lama 3 or Mistral. Both are excellent starting points. Then just have a conversation with it. Ask it to help you draft an email, summarize a document, brainstorm ideas for a project. It's slower than ChatGPT. The interface is simpler, but it's running entirely on your device. Nothing leaves your machine. And if you're someone who wants to go beyond experiments like this and actually build a real foundation in AI, the applied AI certification at AI Hammock is designed specifically for non-technical people who want to earn a credential that means something. It's not just theory, it's practical skills you actually use. Worth knowing that's there. Here's the big picture on this story. Twenty-five companies, many of them fierce competitors, agreed on something. That alone should tell you how high the stakes feel to the industry right now. They're not asking for zero regulation, they're asking for smart regulation, targeted rules that go after actual dangerous uses, rather than broad restrictions that would reshape who gets access to AI and on whose terms. This debate will play out in hearings, in legislation, in executive branch reports over the next year, and its outcome will shape something pretty fundamental, whether AI stays distributed and open or becomes infrastructure that runs through a few corporate gates. That's a question worth paying attention to, and now you know what it means. 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.
