Early Years, Predictions, and Unlocking AGI

In 2009, before the rise of deep learning, the person accurately predicted that AI would reach their level by 2025. They based this estimate on the belief that computational power and data would grow exponentially, driving the development of more scalable algorithms. They also anticipated that by the 2020s, models could be trained on much more data than a human can experience in a lifetime, leading to major advancements and the eventual unlocking of AGI. They believe there's a 50% chance this will happen by 2028, but acknowledge that unexpected problems and delays are always possible in research and science.