Say it takes 10 years to replace most of the fleet. That still means someone driving a truck at age 25 will be facing an extremely different job situation at age 35.
We could use the tech to enhance the worker instead of replacing the worker, though. Right now, drivers are limited in the number of hours they can do in a day, for safety reasons. If drivers become mobile managers, there is still a need for them. Still less of them, obviously, but it does mean we don't suddenly get rid of 80% of their job openings in 10 or 20 years.
Passenger cars have a short enough lifespan that I doubt those sort of retrofits will be especially common.
I think it comes down to what you mean by "soon". I would be surprised if in 10 years a huge number of trucking jobs hadn't already been replaced, and in 20 years they will almost gone outside of a few niches.
These sorts of things get waved off with tech enthusiasm as things that will just somehow get figured out; this leads to a straw man where the real costs of human drivers today are compared against the hypothetical, unknown, but believed to be effectively zero costs of the glittering future.
Not a very fair comparison I'd say, comparing fantasy to reality. We'll see how it shakes out in the years to come.
Much less than what they're paying now for insurance covering their human drivers.
> How much will the annual software updates cost?
Less than their human drivers cost, if anything at all.
The need to replace truck fleets would be one reason why it won't happen overnight. It would take time to replace millions of expensive vehicles.