When looking at robotics is also good to consider not just economics but safety. There are situations you simply don't want to put a human into where robotics shine, usually in some form of emergency response in hazardous situations. Not only is it acceptable to be more expensive than its human counterpart, it's often acceptable to be a lot worse at the task. In cases where a robot is saving human suffering or doing something incredibly dangerous, robotics can be a rewarding field. I've done some limited robotics work in emergency response and you can tell people up front "listen, this stuff isn't great, but it may save some lives" and be completely honest.
The Fukushima reactor exploration for example is only possible because of robotics. Those things are failing like flies due to radiation but it's clearly worth it.
I love what you're preaching, but I actually worked on something similar to what you're describing (just not for nuclear reactors, but something else famously dangerous). Ultimately the company footing the bill decided to axe the project because they decided they'd rather send people into a deadly environment rather than continue to pay for the R&D to develop the robot.
> When looking at robotics is also good to consider not just economics but safety.
The problem is the companies paying for the robots only care about economics, at least in my experience. I'm glad your experience differs from mine.
The Fukushima reactor exploration for example is only possible because of robotics. Those things are failing like flies due to radiation but it's clearly worth it.