Why did he start a research direction that 50 years later "might turn out" to not be a good idea?

That's a relatively silly question, don't you think?