In some cases it may indeed be worth automating even when producing at low volumes: * data mining or AI can analyze data sets better than all the humans you co

In some cases it may indeed be worth automating even when producing at low volumes: * data mining or AI can analyze data sets better than all the humans you could possibly muster, data sets that can illuminate low volume but high impact decision making (new product need, safety critical alerts, incoming asteroids) * robot surgical devices can perform more consistently (possibly higher efficacy) which is highly valuable even for low volume surgeries * automated cars will be demonstrably safer even if overall demand for transportation declines * demand for more specialized and lifelong education will increase even if the number of students decreases (career changes, working retirements, self-improvement, etc) * automated solutions to climate change remediation (carbon sequestration, planting trees, etc) will likely scale well beyond manual labor options in an aging shrinking population and will be needed generations after the damage levels off * space exploration requires automation even for just a single mission, and even a smaller population will likely continue to explore the solar system I think imagination and adaptation will manage ok. In fact I would rather see capital deployed to less consumption by volume and more high value consumption. Would you rather eat a farm-to-table organic microgreens salad for dinner, or a bag of Wonder Bread?