The fixed-location version of Starlink consumes around 60 to 100 W constantly which is problem if you want to e.g. use solar panels on a sail boat to supply the device.
I'm guessing a more affordable version more like the RV product will become available for cruising sailors in time. I think the mammoth price delta over RV is because it'll be usable offshore (starting Q4 this year) and presumably that requires some more complex satellite-to-satellite data exchange (which at a guess they want to limit usage of until it works well.
Sure, just like Tesla Autopilot is right around the corner, Starship will be flying this year, the Cybertruck has been released 2 years ago etc.
None of that is true for the self-driving car problem. That problem still contains a multitude of unknowns, including unknown unknowns.
Cybertruck is yet another kind of problem. I don't know what the issue with that is but I'd guess it's about manufacturing capacity.
Those communications are still an unsolved and hugely difficult engineering problem. It will be awesome if SpaceX has actually achieved this: getting the kind of precision required to communicate over direct laser links between specks of dust hundreds of km apart traveling at thousands of km per hour is no easy feat.
The additional load of the Starlink, relative to the chillers, water makers, and other onboard systems would be nothing.