Don't think about money, that only confuses things. Think about resources instead. Imagine that the worker burns oil to survive, and the robot also burns oil to function. Now replacing the worker with the robot makes the economy burn twice as much oil for the same amount of production (assuming the worker goes on welfare). If you think in terms of money and pockets, the picture will be more complicated, but the end result will be the same.

This logic is full of holes. A robot can, by not needing to rest, be vastly more productive than the same person. The robot can be designed to never make a mistake too. Indeed, automation of all aspects of bread production would increase the bread supply. And the person needed bread either way, so unless you're saying that their death is the alternative, what they do with their day is irrelevant. And about fuel, I guess "burns oil" tugs at more heartstrings than "burns solar energy."

    > This logic is full of holes.

I don't agree. The commenter said that now both the robot and the worker use resources. It's gotten better if you only look at the business and ignore the laid-off worker.

Note that I'm always for such improvements, the solution for a broken system should not be to let people do useless jobs (like the road workers whose sole job it is to hold up a "Stop" sign in a construction zone).

I think the argument is solid, but not that impactful. Sure, worker + robot = seems like more energy used than worker alone. But robots will be more efficient, probably impressively so if you factor in energy savings on human commute and maintaining "habitable" (for lack of better word) conditions in the workplace. Lights, HVAC, safety equipment, etc. Given that, it's not inconceivable that welfare humans + working robots could use less energy than working humans alone!

But that's all besides the point. Even if energy usage E(humans+robots) = 2 x E(humans) (and I suspect it's more like factor 1.1x), it's still worth the cost for the quality-of-life increase for those humans. Economy should not be about optimizing its random and fluctuating "productivity" function. It's about improving quality of life for humans.

The argument is talking about the marginal case [1] where the robot is about as cost-effective as the worker it's displacing.

[1] https://en.wikipedia.org/wiki/Marginal_concepts

But robots don't burn bread. Their ongoing input costs are only a tiny fraction of a human's, that's why people buy robots even though their up front cost is usually much higher than hiring a person.

The point is that the robot replaces the worker only on the business level - but in the economy the worker is still there! And in today's world either often gets no job or a worse job.

But doesn't this rest on the assumption that a working worker uses the exact same amount of oil as a worker not working? I don't think that assumption always true. Or am I missing something?

I agree that the factor might be 1.5 instead of 2, or even below 1 in some cases.

Or into the pockets of the owners of the robots who develop, maintain, and power these robots.