I should make more 5 and 10 year out predictions, because people are very recency-biased in their view of the world and "Hey let's revisit this list from 5 years ago" helps to combat that.
A great example: blog.samaltman.com/bubble-talk
A great example: blog.samaltman.com/bubble-talk
There is, in November of 2019, a narrative around "Oh goodness startups have been brought to earth, the music has stopped, whatever will happen in the near future", and it's useful remembering "People were confidently calling the top in 2015. Here was an optimist's prediction."
And not to scoop an update post, but at least two of those three predictions are extremely uncontroversially right.
In addition to recency bias I think people are *far* too likely to be negative in general outlook and will be able to find at least one negative thing which happened and say "See? I was right all along", so putting an actual numerical threshhold on observable outcomes is good.
