This paper analyzes how output and wages behave under different scenarios for technological progress that may culminate in Artificial General Intelligence (AGI). The authors assume that human work can be decomposed into atomistic tasks that differ in complexity. Advances in technology make more complex tasks amenable to automation. The effects on wages depend on a race between automation and capital accumulation. If the distribution of task complexity has a sufficiently thick infinite tail, wages may rise forever. However, if task complexity is bounded and full automation is reached, wages collapse. Automation may lead to broad-based gains in returns to all factors, but bottlenecks from scarce factors may worsen wage declines. The paper introduces an economic framework to analyze these scenarios. It considers two cases: one where task complexity is unbounded, and another where it is bounded. In the unbounded case, automation leads to more tasks being automated, but some remain. In the bounded case, full automation is reached in finite time, leading to wage collapse. The paper also considers dynamic settings where automation and capital accumulation compete. It shows that wages depend on the balance between productivity gains and labor displacement. If automation outpaces capital accumulation, wages may decline. However, if capital accumulation catches up, wages may rise again. The paper also explores the effects of automation on the factor price frontier (FPF), which reflects all possible combinations of factor prices. For a given level of automation, wages lie within a bounded interval that expands as automation increases. Once all tasks are automated, the FPF collapses to a single point where the returns to labor and capital are equalized. The paper also considers the role of fixed factors, such as minerals, which may limit economic growth and worsen wage outcomes. It also analyzes the potential for automating technological progress, which may lift returns to all factors, including wages. The paper considers different scenarios for the advent of AGI, including a business-as-usual scenario, a baseline AGI scenario, and an aggressive AGI scenario. It shows that automation can lead to faster growth in output and wages, but also to wage collapses as automation approaches full scale. The paper also considers the role of compute as a specific capital that is tailored to automating specific tasks. It shows that compute may earn high returns in the short term but may become just another form of capital in the long run. The paper also analyzes the impact of automation on workers with heterogeneous skills and susceptibility to being automated. It shows that automation may lead to an ever-declining fraction of superstar workers' earnings, while the majority of the labor force is devalued. The paper also considers the role of societal choices to retain certain jobs as exclusively human, which may help keep labor scarce and wages growing even when full automation is technically possible. It also analyzes the wage-maximizing rate of automation and shows that slowing down automation in an AGI scenario may deliver significant gains to workers, albeit at the cost of forgoing a growing fraction of outputThis paper analyzes how output and wages behave under different scenarios for technological progress that may culminate in Artificial General Intelligence (AGI). The authors assume that human work can be decomposed into atomistic tasks that differ in complexity. Advances in technology make more complex tasks amenable to automation. The effects on wages depend on a race between automation and capital accumulation. If the distribution of task complexity has a sufficiently thick infinite tail, wages may rise forever. However, if task complexity is bounded and full automation is reached, wages collapse. Automation may lead to broad-based gains in returns to all factors, but bottlenecks from scarce factors may worsen wage declines. The paper introduces an economic framework to analyze these scenarios. It considers two cases: one where task complexity is unbounded, and another where it is bounded. In the unbounded case, automation leads to more tasks being automated, but some remain. In the bounded case, full automation is reached in finite time, leading to wage collapse. The paper also considers dynamic settings where automation and capital accumulation compete. It shows that wages depend on the balance between productivity gains and labor displacement. If automation outpaces capital accumulation, wages may decline. However, if capital accumulation catches up, wages may rise again. The paper also explores the effects of automation on the factor price frontier (FPF), which reflects all possible combinations of factor prices. For a given level of automation, wages lie within a bounded interval that expands as automation increases. Once all tasks are automated, the FPF collapses to a single point where the returns to labor and capital are equalized. The paper also considers the role of fixed factors, such as minerals, which may limit economic growth and worsen wage outcomes. It also analyzes the potential for automating technological progress, which may lift returns to all factors, including wages. The paper considers different scenarios for the advent of AGI, including a business-as-usual scenario, a baseline AGI scenario, and an aggressive AGI scenario. It shows that automation can lead to faster growth in output and wages, but also to wage collapses as automation approaches full scale. The paper also considers the role of compute as a specific capital that is tailored to automating specific tasks. It shows that compute may earn high returns in the short term but may become just another form of capital in the long run. The paper also analyzes the impact of automation on workers with heterogeneous skills and susceptibility to being automated. It shows that automation may lead to an ever-declining fraction of superstar workers' earnings, while the majority of the labor force is devalued. The paper also considers the role of societal choices to retain certain jobs as exclusively human, which may help keep labor scarce and wages growing even when full automation is technically possible. It also analyzes the wage-maximizing rate of automation and shows that slowing down automation in an AGI scenario may deliver significant gains to workers, albeit at the cost of forgoing a growing fraction of output