Pages

October 05, 2011

Robin Hanson discusses futurism

Here are three observations on futurists by Robin Hanson

1) Most folks I know who self-describe as future-oriented seem obsessed with the latest tech press releases. hey constantly circulate links on new tech gadget demos and analyses. Which might make sense if “the future” meant the next ten years. But if “the future” means the next century, this makes far less sense. Long term future oriented folk should focus on basic theory and long term trends, and pay little attention to daily tech fluctuations. Press-release-focused futurists seem more interested in affiliating with the idea that “tech is our future” than in actually understanding the future.

NBF - Deeper analysis is possible of press releases and new science announcements. It is possible to cast a wide net and determine what is potentially big enough to move the needle in terms of trajectory of the big picture.

2) Few ever gain fame in futurism on the basis of what they say about the future. Almost everyone “known” for thoughts on the future first gained status and notoriety in some other area, and then started being heard on the future. Folks who talk about the future but don’t have another status base are almost completely ignored. It seems that while positioning ourselves regarding the future, we like to affiliate with high status folks, but don’t see such future positions as conferring status.

NBF - mostly true

3) It is often said that futurists forecast big things to happen in twenty years because their careers will be done then, and they’ll suffer few consequences from mistaken forecasts. But human lifetimes are actually long enough to fit not one but three cycles of tested twenty year forecasts. People could make forecasts at age 20 that are checked at age 40, make another set of forecasts at 40 that are checked at age 60, and then make a third set of forecasts at age 60 that are checked at age 80. We could then pay special attention to the forecasts of eighty year olds who have had a good track record over three cycles of twenty year forecasts.

I will have a longer comment on Robin's third point - there can be multiple overlapping cycles.

You can predictions for twenty years in the future with major revisions every 5 or ten years or even more frequently and new predictions can be added on a rolling basis.




It is difficult and time consuming to make a high volume of accurate and non-trivial predictions. (from my experience as a futurist and predictor)

By keeping in tune with current developments and the current large scale view of the world it is possible to determine what are the high impact trends.

Looking at the current world and trying to regress back 5 or ten years or 20 years to pull out what should have been the most useful predictions. Is a difficult but instructive exercise.

If someone was asking what will be the most important technology for 2016, 2021, 2026, 2031, then it would be good to answer what was the most important technology for 2011 that made the most difference since 2006, 2001, 1996, 1991.

also adding one word in the question changes the answers.

If someone was asking what will be the most important NEW technology for 2016, 2021, 2026, 2031, then it would be good to answer what was the most important NEW technology for 2011 that made the most difference since 2006, 2001, 1996, 1991. (have to add explanation) The technology could not have been commercialized yet in the year it was predicted.

also changing the scope from technology to societal change is major difference.

If someone was asking what will be the most important societal change for 2016, 2021, 2026, 2031, then it would be good to answer what was the most important societal change for 2011 that made the most difference since 2006, 2001, 1996, 1991.

also there would be a difference between what is the biggest change as perceived by most people versus what is biggest impact on business or on geopolitics or on economics or on science or on an industry.

Having the right questions with the right definitions and qualifiers is important. Being able to produce the historical analysis would be relevant for a projection into the future. You can accurately trend out the projections until that area hits a disruption big enough that the macrotrend becomes invalid. You have to have done enough analysis to have identified the accurate and relevant trends.


You can analyze the quality and usefulness of predictions by whether they are falsifiable with a regression. If they are accurate enough to be a guide that eliminates other scenarios. Some technology predictions may be falsifiable in the year that they are made. Something already happened but was not known by the predictor. The prediction is predicated on a worldview that is inaccurate.

If you liked this article, please give it a quick review on ycombinator or StumbleUpon. Thanks
blog comments powered by Disqus