RE: Solar power

hal@finney.org
Mon, 12 Jul 1999 10:53:50 -0700

Mark writes:
> No, it's hot fusion that's been twenty years away since the 50s or thereabouts
> (can't remember when the first claim was made, so you may be right); cold
> fusion is just a fantasy. Oh, and usable solar power systems have been a
> decade away since the mid 70s.

It's interesting to consider how we should update our time estimates when earlier estimates are wrong. Should we take into consideration the earlier mistake and use it to increase our current estimates? Or should we just ignore the past and do our best with the information we have at hand?

As Mark says, there seem to be many cases where technologies recede almost as fast as time advances (or faster in some cases). I even think we will see the same thing happen with nanotech and the singularity. Probably it will still be 20-30 years away in 2010 and 2020.

(Vinge's story, "Bookworm, Run" was published in 1966 and predicted intelligence enhancements in 1984, according to the notes published with the story in True Names. Moravec's "Mind Children" was published in 1988 and predicted computer power adequate for human-level AI would be available in a supercomputer before 2010.)

This has come to mind as I have been trying to estimate times on http://crit.org/http://www.textropians.com/Timeline/home.htm. One question is when we will have a real Idea Futures market (i.e. using real money). Well, in 1995 the IF game set up a claim on that very topic, asking whether there would be a real IF market by 2000. At the time it was considered likely with 80% odds. Over the succeeding four years the likelihood has fallen steadily, and now stands at 8% since we are about out of time.

OK, we were wrong back in 1995. Now we must estimate again, when will there be a real IF market? Once again I find myself thinking that it will be 4-5 years away. It seems like a fun gambling game with a limited but highly interested clientele, we have internet gambling available from selected sites, so it is a matter of getting funding and putting the technology in place.

My 4-5 year estimate is based on these considerations. Now I need to take into account that the consensus in 1995 was also that things were 4-5 years away, and we were wrong back then. Maybe this means that this project is more difficult than it looks. On that basis I should increase my estimates. Instead of looking 4-5 years out, maybe 8-10 years is more appropriate.

Hofstadter's facetious Law says that everything takes longer than you expect, even when you take Hofstadter's Law into account. On this basis I should further increase my estimate to maybe 12 or 15 years.

Hal