Ray Kurzweil Wants to be a Robot
newsweek.com
newsweek.com
But as I've learned more about the limitations of our understanding, our biology, and computers I've realized that this isn't something that will be happening any time soon without some amazing breakthroughs. I don't expect the Singularity to happen my lifetime, and I'm still in my 20s.
We really understand very little about the brain. We don't know exactly how neurons work and the most information we can get from the brain is on the order of 1-7 BITS per second when sampling from implanted electrodes. It's also not a lack of data. We don't even know how to fully decode the data we already have.
Our understanding of AI really is pretty poor. Knowledge representation and machine learning has become a hard enough task that many researchers are back to developing algorithms for specific tasks, which embeds the knowledge of the researcher in the algorithm, precluding the system from learning anything truly novel.
They say we can't see past the Singularity because things will change too much and too fast. Right now we can't even see the Singularity because too many things would have to change too much and quite fast.
It's too much work to do it that way. We need a biologically inspired computer architecture... one that's non von Neumann; where the stored data and computation aren't separate.
Keep in mind that he's been making predictions like these for decades and he's got a pretty good track record.
What specifically did you learn that stands against the singularity theory?
One thing that makes me confident that there can be machines that are as smart as the human brain is that there is already a proof of concept: the human brain.
Of course that does not prove that there can be machines that could be smarter than the human brain, but it does not seem too much of a leap to imagine it.
Did you miss the whole accelerating change thing? Assuming the models are correct, it's going to look like we won't make it until just before we do.
Did you see the burst in development of comms and personal devices in the past 20 years?
Do you see that someone (a westerner, anyway) who died 50 years ago and someone who was born 50 years ago would have very different lives?
Do you see that we don't need AI, we could go with human-computer hybrids (and I mean very good personal assistants acting as external memory and computers, not even brain implants) or effective collaboration instead.
We already have one form of effective collaboration - the company. Groups of people working together to do more than any individual. View that from the perspective of two thousand years ago. See how fast change seems.
If you've had someone you love get old and sick and die, you know what I'm talking about. Get to work. It's not just about you.
Don't know, should I laugh or should I cry?