"The most lucrative field will be robotics, where AI control systems facilitate physical manufacturing and all aspects of physical labour." -- why? Since you are talking about AI alignment this statement seems to imply that you think people will "let AI loose" on robots, i.e., let robots decide what to do and, by virtue of having alignment, hope that it will do something useful for humans. That by itself is absolutely murderous and insane for the foreseeable future, because even 99.99% alignment may still result in lethal events if you actually let that AI freely control the robots behavior. So I actually tend to completely disagree with your above statement, not even though I'm working in robotics, but BECAUSE I'm working in robotics. I understand the reasoning, but I think what you've got wrong is the timing: AI alignment will not be beneficial in robotics for a much longer time than it will be useful in much less dangerous applications.
