from the never-too-late dept.
At the Foresight Senior Associates gathering in 1999, an Idea Futures market was set up for attendees to "put their money where their mouths were" on a variety of questions related to Foresight's goals. The market is now available on an ongoing basis for those who've opened accounts. This article reports on events at this year's gathering: Two new claims were added concerning how and when AI will appear. The most contentious claims during the gathering were those related to the date that the Feynman Grand Prize will be awarded. The odds on those claims have see-sawed back and forth as people with different viewpoints have weighed in on the question. Senior Associates can join by sending money to the Foresight office.
Idea Futures Activity at the 2000 Senior Associate's Gathering
At the Foresight Institute's annual meeting for Senior Associates in 1999, an Idea Futures market was set up for attendees to "put their money where their mouths were" on a variety of questions related to Foresight's goal of preparing for the introduction of future technologies. The Idea Futures concept meshed well with the meeting format, a Design Shop, hosted by Matt and Gayle Taylor, and received a positive response from the attendees, so we decided to continue to run it.
The claims that had been set up in 1999 were available for continued investment at this year's Senior Associate's Gathering, and Robin Hanson, Ken Kittlitz, and Chris Hibbert were on hand to help people figure out how to use the software. While glitches had prevented the use of the system in the months immediately following the 1999 gathering, we expect to have the system up all the time now. We've also made plans to promote the market year-round using nanodot and to encourage people to join throughout the year. (Senior Associates who wish to join can send a check to the Foresight office, or wait until we add support for paying electronically.)
This year's gathering resulted in the addition of $1300 to the market, the approval of two new claims, and some interesting activity. The new claims grew out of the Friday evening discussion of uploading and AI. One of the claims (AI1upl) says the first intelligence in a computer will be an upload (that is, that it will be created by directly reading structure from a particular brain). The odds on this claim are 52 – 55%. The second new claim (upld30) says an upload will happen before 2030. That claim is currently settled at (indecisive) odds of 20 – 60%.
If you're unfamiliar with the Idea Futures concept, the way to interpret those ranges of odds is to say that the market participants believe that the most realistic odds are between those numbers. The first number is the highest amount (out of 100) that anyone has offered for a coupon that will pay off if the claim is judged true, and the second number is the least anyone will accept for such a coupon. Someone who disagrees with the odds can take one offer or the other, or post another offer between them, thereby changing the posted odds.
There was significant activity during the gathering on a few of the claims. In particular, there was a see-saw battle over the three claims on the date that the Feynman grand prize will be awarded. After the 1999 meeting, the odds had been 60-65% of an award by 2020, 30-40% by 2015, and 15-18% by 2010. At this year's meeting someone believed that the prizes are pretty likely to be awarded by 2010, and a near certainty by 2015, and so they bought up all the offered coupons available for less than 90. Others saw this activity, and quickly bid the prices back down. At the moment, the odds are listed as follows: for 2010 25-30; for 2015 40-50; and for 2020 65-75. If you think any of these numbers are significantly off, put your money where your mouth is.
The claim 5yWPx2 (World product doubles 2025-30) was created by Robin Hanson to track the likelihood of a big increase in world economic productivity. Robin suggests that if the doubling time for the world economy changed from 15 years to 5 years, that would signal a fundamental change in the economy. The claim asks whether that will happen by 2030. After the 1999 gathering, the odds were at 30-38%. Today they stand at a more conservative 25-50%. ("More conservative" because they reflect less certainty in both directions.)
Another claim whose odds changed was on whether, by 2015, someone will create reliable computing elements using protein synthesis techniques. Participant's evaluation of the likelihood of this breakthrough has fallen over the year: in 1999, the odds listed as 81-94%, now they're at 50-60%.
The money that was collected has been invested in an S&P 500 fund. The amounts that people added this year were indexed to the change in value of the underlying funds. Those who already have accounts can log in at http://www.trilliad.com/if/Trade. If you've forgotten your password, contact Chris Hibbert, or Ken Kittlitz. At this point, only Foresight Senior Associates are allowed to trade.