More great stuff on the second day of the event.
The morning opened with Google’s Peter Norvig, who discussed the question of whether innovation has stopped. He looked at some trends that don’t appear to be accelerating — life extension and economic growth.
Next came J. Storrs Hall, who discussed the need for a revised Three Laws of Robotics. He came up with four, actually:
1. Robots shall be be built according to evolutionarily stable strategies
2. Robots shall be open source
3. A robot shall be economically sentient (that is, consider the value placed on things by others)
4. Robots shall be trustworthy, loyal, helpful, friendly, courteous, kind, obedient, cheerful, thrifty, brave, clean and reverent…and shall do a good turn daily.
Josh Hall was on a panel with Peter Thiel, who took us through the recent history of economic booms and busts — pointing out their amplitude has gotten bigger. Thiel’s take is that the booms can’t all be fake. He explains it this way:
One of them is going to be real, or the world is going to end.
Given those choices, our investment choices are somewhat different than they have been. To say the least!
[Wired is reporting this morning on Peter Thiel's speech. - Stephen]
The third member of the panel was Charles Harper form the John Templeton Foundation. He took us through three basic questions that have to be addressed:
1. What does a slug know of Mozart? And, by extension, what if AIs quickly become as fard removed from us as we are from slugs?
2. How serious is the dilemma of power? The problem is that the acquisition of power seems to outpace the ability to use power safely an responsibly.
3. How important is the transformation of Desire? Here Harper sites a book by Leon Kass (someone we have had our share of disagreements with on this site) entitled The Hungry Soul. Might have to check that one out.
Over Lunch, we saw a presentation from Michael Lindsay of the X Prize Foundation. The Foundation is considering doing a prize for education. They’re looking at starting out by measuring Algebra, Reading Comprehension, and Second Language acquisition. Lindsay’s intent was to gather feedback from the Singularity Summit crowd. There was a good deal of push-back, particularly concerning the fact that the Foundation plans to use standardized tests to measure the results. It will be interesting to see whether an X-Prize for education ever materializes.
Next came Steve Jurvetson, talking about the dichotomy of design and evolution paths to AI futures. There are strengths and weaknesses to both potential paths, with some possibility that the two approach may converge. Take a look at this site to get an idea of how effective the evolutionary approach may prove to be.
The final panel included Eliezer Yudkowsky, Christine Peterson, and James Hughes, Executive Director or the Institute for Ethics and Emerging Technologies. I made a video recording of both Eliezer and Christine’s presentations in their entirety, so look for those after I edit them and get them posted. Hughes talked about the potential dangers of AGI, particularly the dangers arising from the fact that AGI’s will be completely alien. He pointed out that inaddition to Jurvetson’s two categories, there is a third category — emergent AI, which may be the most alien (and dangerous) of all. He noted that an emergent AI rising out of (for example) Google wouldn’t have to have human level intelligence. Rats and cockroaches have far less than human intelligence, but they can be pretty annoying.
The day wrapped with a Q&A session with Ray Kurzweil, otherwise occupied with Aubrey de Grey’s SENS conference, but not too busy to show up at the Singularity Summit in virtual form.
All told, it was a fantastic two days. Got to meet Speculist reader and sometimes commenter D. Vision in person, although I missed out on seeing the small Colorado contingent who were here. I’ve suggested we have a Colorado singularity get-together in the near future. I did video interviews with a number of attendees, and have video clips from virtually all of the sessions. I also have full audio transcripts of most of the sessions, but I doubt these will have much use other than personal — SIAI will be publishing all this content anyway, in much higher quality. However, I will be putting my video clips together into something not unlike what I did at the library conference last May.
So stay tuned.