We could legislate around this very irrational/human fear. Personally I’d feel much safer if the roads were primarily filled with drivers that don’t get emotional and are objectively safer. Even if a few accidents confuse me and seem avoidable. I think what your analysis is missing is that most human accidents are also 100% avoidable. Why aren’t we looking at the incredibly dumb things humans do and asking the same hard questions? Why doesn’t it spook us when a human doesn't see a red light and t-bones cross traffic? Or when a multi-car pileup happens because a large pickup truck with a big car complex is tailgating someone at 85 mph and a sudden stop is required?
Easy to say if you live somewhere without winter. AI models are a LONG way off from the kind of driving you need in near-whiteout conditions.
> should you complete 70% of your goals to 100%? Or should you complete 100% of your goals at 70%?
explicitly addressed in various articles and books. you average your scores across objectives and KRs to get a score. So the answer is both, looking for average score of 0.7
> if your key result is "migrate 100% of users to the new system" and you only migrate 70%, guess what? Now you're stuck maintaining two systems in perpetuity
so you're going to migrate all of your users in a big bang moment?? hell no. it'll be bit by bit no matter what. of course you're maintaining two systems, that's a sign of a good migration, not a bad one. the KR just helps you get done with it faster
of course there are valid criticisms of okrs, and many (most?) teams get them wrong.
but saying "okrs are bullshit" has become a common trope for new-age companies who use that as a way to differentiate themselves. do you really think Andy Grove and John Doerr are bullshit artists?