I came across this quote on Twitter: "In short, predicting the future of technology is easy: just look at what's already failed, and assume that somebody will eventually make a version that actually works." (attributed to a "departing Bernstein analyst"). Any thoughts? Is something already having been tried and failed an indication that it's a bad idea, a good idea, or neither? https://twitter.com/modestproposal1/status/1300483246699020294

Electric cars in the 1930s and Segway "failed", but now battery tech is forefront - failure is manifold, mostly due to a scale factor

Probably because it was on HN recently, nuclear power comes to mind. I think there are a class of technologies for which this is true. It's a case of incremental improvement vs. fundamental new understanding. The former is much more common because it's "easier to reach in the search space." I think this is another way to state "failure is manifold" from @Chris G's comment. I don't think all future technology can be predicted this way. There is a class of technological evolution that breaks out of such local optima. Using the nuclear example again, the shift from combustion to "understanding and harnessing of the atom", was such an event.

Marc Andreessen a few years back: "there are no bad ideas, only bad timing." https://a16z.com/2020/01/30/neighbor https://samypesse.gitbook.io/pmarca-notes/ideastiming