Response to Gary Stix on the Secrets of the Brain
Gary Stix, Senior Editor at Scientific American, is ragging on our extremely successful recent event, the Singularity Summit 2009 in New York, for which all the videos are online. He attended a meeting of neuroscientists and found that everyone agrees the brain is extremely complex and subtle. To quote Kurzweil (something I do rarely), "They say, Kurzweil underestimates the complexity of — fill in the blank. I agree with critics about the challenge. I disagree about the power of the tools we’ll have to solve the problems at hand." The complexity of the brain could be duplicated in simulation very quickly if we had the appropriate scanning tools.
Yet, even some of the most prominent advocates of brain emulation believe it could take until 2070. That's a while. AI is less certain. Kurzweil is considered an optimist with his 2029 date, but I think he could be right -- if we develop MNT by 2022 or so, which we probably won't. Still, Stix seems to be mystery-worshipping to some extent -- "the secrets of the brain"? Maybe the brain's only big secret is annoying complexity, and that's not so special. Anyway, here's my response to Stix's blog post:
The point is not to achieve cryonic revival tomorrow, but eventually. Even if it takes hundreds of years, the practice of preserving neural connections after death to retain one's memories and personality makes sense. The alternative -- letting yourself become worm food -- destroys decades-worth of knowledge and happy memories.
Your analysis of the difficulty of building a "brain in a box" ignores that 1) it could be possible to emulate the brain by fine-level scanning and software implementation without actually "understanding" it, and 2) Artificial Intelligence need not be a carbon copy of the brain any more than the Wright Flyer (or a 747 for that matter) needs to be a carbon copy of a bird.
Stix, like so many others, assumes that intelligence-in-general is the same thing as a copy of the human brain. Why..? Everything else has basic principles that can be abstracted out and instantiated elsewhere, so why doesn't this apply to the brain? Even if the principles of intelligence are highly complex, it seems virtually certain that the human brain contains even more unnecessary complexity. Evolution is a well-known klutz when it comes to details.