1. A fallacy in predicting tech progress is the faith that going from a 0% to 90% solution means that a 100% one is possible. That's often not true. The hardest use-cases are often saved for last. Which means the last 10% will take far more time or turn out to be impossible.
2. Projects like autonomous cars, voice recognition and "AI" can do very impressive things. But we're prone to the fallacy of believing progress will be consistent/linear. Even though the challenges to reach 90/95/100% get harder and harder.
3. Developing new ideas is exploring the unknown - u don't know if at 91% of a solution a new, previously unknown limitation arises, that could not have been predicted, that creates a dead end or untenable costs. This is the common case, not the exception, in innovation history.
4. I'm skeptical about autonomous cars, voice recognition and "AI" (quoted since it's a sloppy term) not because of the tech precisely, but because success for them hinges on both understanding and predicting human behavior at 95-100% levels, which may very well be impossible.
5. People working on ambitious projects are often the least objective. First, they have the sunk cost of years invested. Second, they need prodigious optimism to risk their careers on these projects. Lastly, they have faith that there is no limit to what technology can do.
6. That's why asking tech pioneers "how long until X is mainstream?" is useless - they often say "5 to 10 yrs" because it's a satisfying sounding answer, despite its lack of basis in knowable reality. "Never" is a probable answer they will never offer.
7. Progress is good :) I admire people who risk so much to make an idea real! However, innovation has a long history - the patterns in today's developing tech echo the old - and we should learn from it to help us to think about what's happening now.
