The breakthroughs in AI today aren’t just coming from bigger datasets and more compute; Reinforcement Learning (RL) has quietly become one of the most powerful forces in modern AI development. RL is teaching models to reason and self-correct, enabling capabilities that make AGI feel less like science fiction and more like an inevitable future. In this fireside chat, Datadog Chief Scientist Ameet Talwalkar and Sholto Douglas of Anthropic will discuss the rise of RL-based post-training and its role in making the case for imminent AGI. They’ll look at the technical signals currently pointing toward general intelligence and the obstacles that could still derail it. The conversation will also examine the tension between specialized and generalized models, how they’re being used to build products today, and their implications on future intelligence. Join them as they explore what today's choices in AI research and product development mean for the intelligence of tomorrow.

 132      0