Open Source AI Under Fire: A Crypto Insider's Take on the Regulatory Capture Game

We didn't need a deep-dive report to recognize the pattern. The same moves we've seen in crypto—regulators cozying up to the incumbents, the little guy getting squeezed out—are now playing out in AI. Last week, David Sacks, a man who's made a fortune betting on open source, pointed a very public finger at Anthropic. His charge: regulatory capture. The same tactics we saw in the fight over DeFi and centralized exchanges, now being run by the 'safety-first' crowd to lock down the open-source AI ecosystem. Look, I've been around long enough to know that when a company like Anthropic—backed by Microsoft, Google, and Salesforce—starts talking about the 'dangers of open models,' you should start counting your money. Not because they're wrong, but because they're loud. And in this game, volume is a weapon. Sacks, a Silicon Valley free-market guy, sees the writing on the wall: if you can't beat the open-source devs on merit, you change the rules. That's regulatory capture. And the crypto community? We've been eating that same meal for years. Now, I'm not just some AI enthusiast. I'm a macro watcher who spent the last five years staring at liquidity flows, token narratives, and the social dynamics of markets. My Manila rave days taught me to read the room—literally. When the music is pumping, the crowd doesn't care about the DJ's beat drop; they care about the energy. Same here: when the fear of AI safety gets pumped into the public's mind, the crowd forgets that open-source models are the decentralized alternative to big tech's locked-in APIs. They forget that open models are like Bitcoin—permissionless, transparent, and owned by everyone. And that's exactly why they're dangerous. The core issue here isn't about safety. It's about control. Anthropic's Claude is a great model, but it's a closed one. Open-source models like Llama 3 and Mistral are eating away at the performance gap. They're being run by startups, by researchers, by devs in Manila like me who want to fine-tune a model without asking a corporate permission slip. If Anthropic and its buddies succeed in painting open models as irresponsible, they can get lawmakers to impose compliance burdens that kill the little guys. They can make the cost of hosting a model so high that only the big tech players can afford to play. That's not safety. That's a moat. I've seen this movie before. In DeFi summer, the biggest risk wasn't the smart contract bugs—it was the regulators. We had protocols that were open, but they were also scary to the incumbents. So they threw the entire space under the bus with the 'it's all a scam' narrative. They didn't do it because it was true; they did it because it was convenient. And then they launched their own bankcoins, their own permissioned ledgers. The same thing is happening in AI. The people with the most to lose from open models are the ones with the most power to regulate them. That's not a conspiracy theory; that's the history of innovation. But here's the contrarian angle that gets ignored: the open-source community might actually be the safest place for AI development. You can't hide a backdoor in a model if a thousand eyes are reviewing the weights. You can't build an automated surveillance tool if you can't see the code. Yes, open models can be misused, but closed models have their own biases—and they're not transparent. The real risk is not open; it's a monopolistic AI that decides what you see, what you hear, and what you believe. And that's a risk that no safety regulation can mitigate. We didn't need a centralized oracle to tell us that. We just needed to look at our own history. The Bored Ape Yacht Club taught us that social capital is an asset. The NFT crash taught us that status symbols can become liabilities. Now the AI battle is teaching us that openness is the only insurance against capture. If you don't own the model, you're just renting the intelligence. And renting is fine until the rent goes up or the owner changes the rules. So what does a crypto person do? We don't just sit and watch. We get involved. We support open-source AI projects—like we supported open DeFi. We push for regulations that don't treat open models like ticking time bombs. We call out the irony when a company that claims to care about humanity uses its lobbyists to shrink the community's innovation space. And we use our own networks to amplify the voices of the open-source devs who are too busy building to fight. In the next cycle, the AI and crypto worlds are going to collide. Decentralized compute networks, open models on-chain, data markets—it's all coming. But if we let the regulators capture the AI space, that future is dead. The music stops. The liquidity dries up. And we're left with a few big players who decide what's 'safe.' I don't know about you, but I didn't come to the crypto scene to be safe. I came to build something that couldn't be stopped. So here's my takeaway, and it's not a summary—it's a question. Are we going to let the AI industry become the next version of the banking cartel, or are we going to demand the same permissionless energy that made this whole movement possible? The beat is still going. The question is, do we keep dancing to a tune that's controlled by the man in the booth? Or do we make our own rhythm? You know my answer. We didn't wait for permission before. We don't need to start now.