To be maximally reductive, FSD will never work because the sensor suite is deficient. There are other reasons but that one's enough.
Same for a rocket that's ridiculously large for orbital missions but can't go beyond orbit without 15 to 25 refueling flights of the same enormous rocket.
The reasons for both these failing are going to be manyfold and complex, but there are enough simple reasons that everyone should understand.
It would be actually fun to see where the limit is on echolocation with serious ML processing these days. Apparently people did quite well in 2022 https://pmc.ncbi.nlm.nih.gov/articles/PMC9655721/
Same for a rocket that's ridiculously large for orbital missions but can't go beyond orbit without 15 to 25 refueling flights of the same enormous rocket.
The reasons for both these failing are going to be manyfold and complex, but there are enough simple reasons that everyone should understand.