The 14% probability economists assign to rapid AI progress by 2030 feels low when you look at what's already happening in physical AI deployment. The disconnect is that economists are modeling AI as a productivity multiplier for existing workflows, but the real disruption comes from AI enabling entirely new operational architectures -- autonomous robot fleets that don't just replace human labor but eliminate the need for the facility designs, shift schedules, and management layers built around human constraints. The labor force participation drop in the rapid scenario understates the structural shift because it doesn't capture the infrastructure transformation: every warehouse, factory, and farm that deploys autonomous systems also needs new connectivity stacks, edge compute, OTA update governance, and fleet telemetry -- creating new job categories while eliminating old ones faster than the models capture.
It would also be helpful to ask this surveyed group how a labor force transition would happen. Presumably, a job retraining program would have two positive effects:
1. A workforce trained to collaborate with AI tools such that the productivity gains are more widespread, thereby further increasing the economic diffusion
2. Stabilized employment rate so the spending power of the population doesn’t drop, as the the US economy is largely a measure of consumption levels of the population
It would be good to actually depict what these training programs would look like. Some industries may have much larger effects of unemployment than others. Humans are also not well suited for learning exponentially changing technologies, especially in roles that have the most AI exposure (what is a translator who lost his/her job supposed to transition to, when any relevant language-related AI capabilities are improving at exponential rates?)
The 14% probability economists assign to rapid AI progress by 2030 feels low when you look at what's already happening in physical AI deployment. The disconnect is that economists are modeling AI as a productivity multiplier for existing workflows, but the real disruption comes from AI enabling entirely new operational architectures -- autonomous robot fleets that don't just replace human labor but eliminate the need for the facility designs, shift schedules, and management layers built around human constraints. The labor force participation drop in the rapid scenario understates the structural shift because it doesn't capture the infrastructure transformation: every warehouse, factory, and farm that deploys autonomous systems also needs new connectivity stacks, edge compute, OTA update governance, and fleet telemetry -- creating new job categories while eliminating old ones faster than the models capture.
What about actual people and their: energy bills, access to water or education and opportunities etc
It would also be helpful to ask this surveyed group how a labor force transition would happen. Presumably, a job retraining program would have two positive effects:
1. A workforce trained to collaborate with AI tools such that the productivity gains are more widespread, thereby further increasing the economic diffusion
2. Stabilized employment rate so the spending power of the population doesn’t drop, as the the US economy is largely a measure of consumption levels of the population
It would be good to actually depict what these training programs would look like. Some industries may have much larger effects of unemployment than others. Humans are also not well suited for learning exponentially changing technologies, especially in roles that have the most AI exposure (what is a translator who lost his/her job supposed to transition to, when any relevant language-related AI capabilities are improving at exponential rates?)