Noam Brown: AI won't plateau — if we give it time to think

AI won't plateau — if we give it time to think

To get smarter, traditional AI models rely on exponential increases in the scale of data and computing power. Noam Brown, a leading research scientist at OpenAI, presents a potentially transformative shift in this paradigm. He reveals his work on OpenAI's new o1 model, which focuses on slower, more deliberate reasoning — much like how humans think  — in order to solve complex problems.

AI progress has long depended on making models bigger and training them on more data, with costs soaring from thousands to hundreds of millions of dollars. Yet, dramatic leaps have come from teaching AI to pause and think, not just react instantly. In high-stakes poker, simply giving an AI 20 seconds to reason matched a 100,000-fold increase in brute-force training. This new approach, now powering emerging models, opens the door to smarter solutions without runaway costs.