Guns were pretty much immediately better. They were easy to use, cheap to reload, and you could shoot through armor.

I'm not sure this is true. Reloading muskets took a long time, no? You had to fumble with pouring the powder, tamping it, putting in the round, possibly also a sabot. "Reloading" a bow takes a second.

Bows could shoot faster and further but you would tire out. You also had to be a trained archer vs a random nobody with a gun. Becoming an archer was hard, and arrows were a lot more expensive.

1k longbowmen beat 1k musket bros, but that’s not the comparison that really mattered.

I think we should solve autonomous driving before jumping multiple steps to autonomous surgery. There are many low hanging fruit in the world of automation; automated surgery is not one of them.

Autonomous surgery is probably easier to solve than driving, because the patient is lying still. Though you'd still need at least a nurse observing and a surgeon standing by. But you could have only one surgeon as backup for multiple robots in the same hospital.

A surgery is a closed task. Driving is an open ended task interacting many actors and novel challenges. It’s probably slow enough that it can visually confirm assumptions with a human operator if need be.I would bet on the autonomous surgeon over the autonomous driver any day.

Human anatomy is actually surprisingly variable and there is, of course, trauma - so novel challenges.

The task is also not necessarily closed at the start, i.e., target of surgery is established during the procedure and might evolve.

So, I'd take the other side of that bet for surgery as a whole.