How our timelines forecasts have shifted since AI 2027 and our last update
1. These are our all-things-considered forecasts, meaning that they are our forecast for what will happen in the world conditional on going as fast as technically feasible (e.g. no government-imposed slowdown). This includes adjustments on top of the outputs of our quantitative timelines and takeoff models.
2. We operationalize the term "AGI" as TED-AI, which is defined as: AIs at least as good as top human experts at virtually all cognitive tasks.
3. Superintelligence is defined as: the gap between an ASI and the best humans is 2x greater than the gap between the best humans and the median professional, at virtually all cognitive tasks.
4. This wasn't recorded at the time, so we aren't sure what Daniel's exact median was.
5. Superhuman Coder is defined as: AIs providing equivalent software engineering productivity as if there were 30x as many agents as an AI company has engineers, accomplishing tasks 30x faster than the best engineer.
2. We operationalize the term "AGI" as TED-AI, which is defined as: AIs at least as good as top human experts at virtually all cognitive tasks.
3. Superintelligence is defined as: the gap between an ASI and the best humans is 2x greater than the gap between the best humans and the median professional, at virtually all cognitive tasks.
4. This wasn't recorded at the time, so we aren't sure what Daniel's exact median was.
5. Superhuman Coder is defined as: AIs providing equivalent software engineering productivity as if there were 30x as many agents as an AI company has engineers, accomplishing tasks 30x faster than the best engineer.