I think we’re ‘20 years away’ from it in the same way we were ‘20 years away’ from practical fusion power in the 50’s. It feels close, but we don’t even know what we’re actually trying to achieve. People don’t even agree or understand what human intelligence is, what creativity is. You can’t progress down this path like we do with newer, faster processors. It’ll take a new epiphany, a whole new approach to get to the kind of AI people have been dreaming of and writing stories about. You could give the machine a thousand times the processing power and all the training data possible, but it will never really progress past a the current shallow mimicking of intelligence. There’s no mechanism for it to grow or be corrected on the facts. That will take something new.