Who said it needed to be well-funded? Creating a sentient AI is likely to become an easier undertaking as general computing technology and AI research advances. Creating such an AI in 20 or 30 years might require a large deal of funding, but what about in 50, 60 or even 100 years?
I don't think anyone can predict more than 20-30 years into the future with any real degree of accuracy.
So yeah, I agree that X might very well be possible (or in widespread usage) in 50 or 100 years, where X is basically anything that seems magical today.
