Last year I attended a singularity conference and Ray Kurzweil's avatar predicted it was 25 years away. Well, it's been 25 years away for a long time. It's a nice, safe number, close enough that no one gives up and stops buying books (global warming will happen in 100 years, for example) and not so close anyone looks silly (Al Gore saying in 2006 that we were doomed in 10 years, for example) it if doesn't happen.
In 1993, for example, Vernor Vinge said "Within thirty years, we will have the technological means to create superhuman intelligence. Shortly after, the human era will be ended."
Well, that would have been 2023. In 2011 we can't even predict the weather 7 days out and brains are a lot more complex than weather so we will replace intelligence? Not likely. So Kurzweil later said it would be 2028 a while ago and now 2045.
In a recent TIME interview, Lev Grossman noted that Kurzweil was on I've Got a Secret in 1965. Well, that's part of the problem. He is stuck in the 1950s where virtually everything was believed to be distilled down to algorithms, logic, neural networks and maybe linguistics. That, to Kurzweil, is what you need to reverse engineer the brain. It's the same crappy 1950s thinking that led to lousy music quality we have today. Who wants a brain that's analogous to CD recordings of real music? Or, worse, MP3s?
Kurzweil again falls back on the Magic Bullet of Future Science. Like global warming skeptics who think nothing needs to be done today because "the future" will handle it, it's an intellectual placebo. Because Moore's "Law" regarding computer processing (1950s thinking again) says computers will get faster, Kurzweil believes Faster Computers + [Magic Black Box Science] = Artificial Intelligence.
It would be silly except he makes a lot of money at this. Now, he is not in the same league as homeopathy, which bilks primarily poor, uneducated people. Kurzweil is instead bilking people who are educated but maybe too optimistic, so if they aren't caveat emptor about his claims, we have no reason to deny him a livelihood.
And he still works in his old dates so he can't be called out too readily - he still says these magic exponential curves will successfully lead to reverse-engineering the human brain by the mid-2020s but the actual Singularity will happen by 2045. The TIME scribe calls him conservative - well, he could be dead by then so it's conservative in the sense that it's a safe date. Kurzweil's father died at 58 and Kurzweil is already 62 so living to 96 may be a stretch in his mind, though he claims his "biological age", whatever that means, is only 42. So he can live forever if his biological age never changes. Or something.
The Singularity remains, for now, religion for technology geeks. Like the guy on the street corner predicting the end of the world, they will eventually be right or die, there's no middle ground, but for people outside the movement it's hard to understand why anyone spends their days thinking about it.
Kurzweil has always been an interesting person, a fun guy to sit around and talk with, if you can suspend disbelief. But his belief that the biochemical complexity of an organic brain can be matched by silicon has to be maddening to biologists - especially when he falls back on the response that the complexity of reverse-engineering the human brain is not the problem, but rather biology underestimating the power of exponential growth.
Ray Kurzweil Pushes Singularity Back To 2045
- Ray Kurzweil Responds To PZ Myers Regarding "Ray Does Not Understand The Brain"
- Singularity Science Theater 3000: How Reverse-Engineering Postponed Artificial Intelligence
- After The Singularity, Music Quality Will Be Even Crappier Than It Is Now
- Sorry Ray Kurzweil, AI Hasn't Improved Much Since The 1960s
- Ray Kurzweil Video Presentation On The Brain And The Power Of Hierarchical Thinking (full Video)