I'm writing to you from inside Memorial Auditorium on the beautiful campus of Stanford University in California. It's taken me about a half hour to find the right spot for a good wireless connection, but now we're up and running.
So, are you ready for the Singularity Summit? Here goes!
The organizers say they are expecting more than two thousand attendees. Right now the auditorium is only about half full (capacity is 2,060), but we're still about 20 minutes away from the planned start time of 9:00 AM.
Among the familiar faces I've seen so far are J. Storrs Hall, Robert A. Freitas Jr., Christine Peterson, Ray Kurzweil, Neil Jacobstein, Michael Anissimov -- and, of course, Tyler Emerson, who is the tireless but largely unsung major organizing force behind this event.
We've just finished the opening introductions and now Ray Kurzweil is beginning his talk. He began with a demonstration of a handheld text to speech translator; quite impressive. Ray says that just five years ago the software was not yet "feasible" to support such a powerful capability.
Can a computer tell the difference between a dog and a cat? Yes! Seven years ago this was not possible; today the software is better, but also today there is a huge repository of online images of cats and dogs (and other things), from which computers can learn.
He's discussing intuitive linear extrapolation [CRN PPT] versus exponential acceleration, with examples of the human genome project, serch engine adoption, etc. This first half of his talk, as expected, is pretty much Ray's standard presentation. We'll see if he introduces any new concepts or arguments in the second half, which, according to him, is supposed to address skeptics.
Even though I've heard Ray speak several times, and I've read his books, he is so excited about his what he says -- and so well informed -- that's it's always inspiring, enjoyable, and educational to attend his talks.
He's talking about expanding the "AI tool kit," with AI research, reverse engineering the human brain, cognitive science, etc.
Is the human brain too complex for us to understand?
"Models often get simpler at a higher level, not more complex."
Hm, it looks like he's running late, because he's skipping over several slides.
"The bulk of human intelligence is based on pattern recognition."
Now he's doing a demonstration of a proposed real-time verbal language translation system. We're not quite there yet, but getting much closer.
He's reviewing and adressing the standard objections. This is based on chapter 9 -- "Response to critics" -- from his book, The Singularity is Near.
He's issuing a challenge to Bill McKibben, author of Enough, as to whether we should establish some arbitrary limits to technological growth and human progress.
And that wraps us Ray's opening keynote.