Google invests in a lot of moonshot projects, and also shuts down projects regularly. To their credit, they are willing to invest in high risk/high reward stuff, so the fact that Google is investing in something doesn't actually mean that this thing will succeed, or not be cancelled in the future.
But remember that there is a lot of value between "no computer on board" and "fully self-driving". Cars already have things like lane assist technology, smarter cruising technology, collision avoidance, software to help maintain distance between the car ahead of you, smarter breaking, better mapping and routing technology.
These are valuable features that can generate revenue even if fully automated self-driving cars never happen, or happen in 100 years. These are also features that can help older drivers with slower reaction times stay behind the wheel and keep driving, thus increasing the number of human drivers.
No, that was closer to “more wood/fewer arrows” (but that was more about aimless toys than deliberate moonshots), which was several years before Alphabet.
Alphabet was to more cleanly separate (in branding and organizationally) the core Google business from other ventures, which moved the moonshots out of Google proper. It wasn't to kill moonshots, though it may be accurate to say it was in part about being better at maturing them.
I never said it was impossible, it's just not an absolute certainty or just around the corner as people make it out to be. It's at best a research project.
