Tech Visionary Says the Big AI Labs Don’t Get What People Want

The irony is delicious. Tim O’Reilly, the man who helped canonize open source in the 1990s, is watching AI cannibalize his publishing empire—and he’s still smiling. But his smile is pointed. He’s telling the big labs something they refuse to hear: you’re building a monoculture, not a utility. Here’s my take: The frontier labs are obsessed with scaling intelligence into a sentient oracle. They measure success in benchmark scores and compute budgets. Meanwhile, regular people want tools that fix their taxes, write a decent email, or help them learn a hobby. The labs are solving a problem nobody actually has—AGI—while ignoring the messy, practical software that defines daily life. This connects directly to the Linux moment of the 1990s. Closed, centralized forces always underestimate the power of fragmented, user-driven innovation. O’Reilly isn’t naive about AI’s risks; he’s betting that open-source won’t just survive—it’ll be what saves AI from becoming a corporate walled garden. If he’s right, the real signal isn’t about models. It’s about who controls the stack beneath them. The big labs are selling you a silver bullet. O’Reilly is reminding you the future is a shotgun of a thousand small, weird, interconnected tools. {"key_insight":"The real AI revolution isn't AGI—it's empowering a thousand niche, user-centric applications via open weights, not corporate oracles.","confidence":0.82}