Tech Visionary Says the Big AI Labs Don’t Get What People Want
Tim O’Reilly built a publishing empire that AI is helping to destroy. Yet he loves AI—as long as it’s open source.
Steven O'Reilly, a tech visionary, argues that the big AI labs are missing the mark on what people genuinely desire from artificial intelligence. He contends that the focus on open-weight models, while important, is merely a part of a larger vision for AI. O'Reilly emphasizes the need for a clean separation between the model, harness, and application, advocating for an architecture of freedom and participation rather than control.
He believes that big models like Claude are optimized for specific use cases, but these may not align with what ordinary people want. O'Reilly goes on to criticize the current architecture of control employed by major AI labs, which enables easier tracking of users. He argues that giving up control is not in the best strategic interest of these companies, despite appearances.
O'Reilly predicts that the real breakthroughs in AI will come from open-source approaches, allowing individuals and communities to innovate freely and diffuse cutting-edge technology throughout society. He warns that focusing on higher-level frontier models could push the US behind China, which has more widely distributed lower-level models.
Instead, O'Reilly suggests that the future of AI lies in open-source solutions, akin to the rise of the web in the 1990s, which emerged independently of corporate control.
Written by urgent.news from Wired Business's reporting — not their text. Machine-written — it may contain errors, so check the original before relying on it.
