Hacker News Reader

Yudkowsky isn't claiming "innovations always turn up quicker than expected", to which indeed fusion would be a counterexample, he's claiming "very soon before an innovation turns up, it often seems decades in the future even to most practitioners", and fusion is not a counterexample to that.

View on HN

Right but on the quadrants of predictions of timing vs imminence, all the examples are in one quadrant to justify the need to act now. Fair enough for reinforcing a narrative but a tad disingenuous.

All possible examples would seem to be in one quadrant, because what we remember -- if anything -- is the time just before it in fact was made possible.

The alternative would be technologies that were never developed at all, most of which never had this sort of discussion and therefore wouldn't work as examples.

Take a more historical view, though, and you'll notice there were people claiming flight was near even decades before the Wright brothers.

Right, exactly as Eliezer says. There are plenty of examples in all four quadrants, so as a way of working out how near a technology is, this works less well than you'd like.

Reply on news.ycombinator.com