The two root problems in your post are that you treat self-driving cars as cost-free instant teleportation devices and that you don't understand which costs drive the particular forms that businesses take.
Diversification of vehicle types
Somewhat, but much less than you expect because contemporary car design is driven by law-mandated safety requirements. The same requirements put a floor on the cars' weight.
There is also the fact that a general-purpose car will spend less time sitting in a parking lot doing nothing while waiting for someone to require it. Renting specialized equipment is expensive partially because of this -- there is a lot of idle time.
Services at home
Nope. It's not the case that the doctor doesn't come to your house because she can't afford a driver. The doctor doesn't come to your house first because her time is more valuable than yours and second because it's hard (=expensive) to bring along all the nurses and assistants and the medical equipment that she has around her office. And, by the way, the doctor doesn't fill out the insurance paperwork -- she has a much cheaper assistant who does.
Of course you can get a doctor (and a hairdresser, and a tatoo artist, etc.) to come to your house, even without self-driving cars. It's just going to be very very expensive. I don't expect this to change.
Rent anything
The cost of a driver is a minor component of the cost of renting large, expensive, luxury things. Taking it out will not make them suddenly affordable. And, by the way, who will unload, set up, dismantle and load back into the self-driving truck all these jacuzzis and huge sound systems?
Also, about the "stuff that previously only millionaires or billionaires would afford" that your median-income person would be able to rent if only you take the truck driver out of the equation -- literally nothing comes to my mind.
Self-driving hotel rooms
They are called caravans or camper vans or RVs. They exist. Have you tried renting them? They are quite expensive to rent, much more so than hotel rooms.
Rise of alcoholism and drug abuse
" for a large number of people, driving is their only reason not to drink or do drugs" -- that's, um, wrong. I have no idea how you came up with such nonsense.
Autonomous boats and yachts
There is no such thing as a "sailing license" (in most countries that I know of) and renting sizeable boats is quite expensive. It will not become less expensive if the boat has an autopilot. Recreational boating doesn't want fully autonomous boats, anyway, and commercial shipping already uses autopilots and still finds out that it needs people to run ships.
Mobile storage
I think this already has been mentioned -- storing things in trucks is much, MUCH more expensive than storing them in warehouses. Also, just-in-time (JIT) manufacturing was all the rage a couple of decades ago. The enthusiasm has cooled down considerably since then, mostly because people have found out that JIT production is not robust to disruptions and does not degrade gracefully if something goes wrong. Hint: something always goes wrong.
Renting a self-driving truck is not going to be cheaper than renting a regular truck (the rental company does not supply a driver in any case). Go ahead, look up how much renting trucks costs and then see if you want to store stuff in them.
contemporary car design is driven by law-mandated safety requirements
I don't know about your country but in mine (Germany) the car industry has so much influence they basically write their own laws. (That's how we got those safety requirements: They're defense against cheaper cars from abroad.) If their business model stops being focused on general use cars, the laws will change very quickly.
a general-purpose car will spend less time sitting in a parking lot doing nothing while waiting for someone to require it
Sure! Not a problem if its TCO is prett...
This is a slightly expanded version of a talk presented at the Less Wrong European Community Weekend 2017.
Predictions about self-driving cars in the popular press are pretty boring. Truck drivers are losing their jobs, self-driving cars will be more rented than owned, transport becomes cheaper, so what. The interesting thing is how these things change the culture and economy and what they make possible.
I have no idea about most of this. I don't know if self-driving cars accelerate or decelerate urbanization, I don't know how public transport responds, I don't even care which of the old companies survive. What I do think is somewhat predictable is some of the business opportunities become economical that previously weren't. I disregard retail, which would continue moving to online retail at the expense of brick and mortar stores even if the FedEx trucks would continue to be driven by people.
Diversification of vehicle types
A family car that you own has to be somewhat good at many different jobs. It has to get you places fast. It has to be a thing that can transport lots of groceries. It has to take your kid to school.
With self-driving cars that you rent for each seperate job, you want very different cars. A very fast one to take you places. A roomy one with easy access for your groceries. And a tiny, cute, unicorn-themed one that takes your kid to school.
At the same time, the price of autonomy is dropping faster than the price of batteries, so you want the lowest mass car that can do the job. So a car that is very fast and roomy and unicorn-themed at the same time isn't economical.
So if you're an engineer or a designer, consider going into vehicle design. There's an explosion of creativity about to happen in that field that will make it very different from the subtle iterations in car design of the past couple of decades.
Who wins: Those who design useful new types of autonomous vehicles for needs that are not, or badly, met by general purpose cars.
Who loses: Owners of general purpose cars, which lose value rapidly.
Services at home
If you have a job where customers come to visit you, say you're a doctor or a hairdresser or a tattoo artist, your field of work is about to change completely. This is because services that go visit the customer outcompete ones that the customer has to go visit. They're more convenient and they can also easily service less mobile customers. This already exists for rich people: If you have a lot of money, you pay for your doctor's cab and have her come to your mansion. But with transport prices dropping sharply, this reaches the mass market.
This creates an interesting dynamic. In this kind of job, you have some vague territory - your customers are mostly from your surrounding area and your number of competitors inside this area is relatively small. With services coming to the home, everyone's territories become larger, so more of them overlap, creating competition and discomfort. I believe the typical solution, which reinstates a more stable business situation and requires no explicit coordination, is increased specialization within your profession. So a doctor might be less of her district's general practitioner and more of her city's leading specialist in one particular illness within one particular demographic. A hairdresser might be the city's expert for one particular type of haircut for one particular type of hair. And so on.
Who wins: Those who adapt quickly and steal customers from stationary services.
Who loses: Stationary services and their landlords.
Rent anything
You will not just rent cars, you will rent anything that a car can bring to your home and take away again. You don't go to the gym, you have a mobile gym visit you twice a week. You don't own a drill that sits unused 99,9% of the time, you have a little drone bring you one for an hour for like two dollars. You don't buy a huge sound system for your occasional party, you rent one that's even huger and on wheels.
Best of all, you can suddenly have all sort of absurd luxuries, stuff that previously only millionaires or billionaires would afford, provided you only need it for an hour and it fits in a truck. The possibilities for business here are dizzying.
Who wins: People who come up with clever business models and the vehicles to implement them.
Who loses: Owners and producers of infrequently used equipment.
Self-driving hotel rooms
This is a special case of the former but deserves its own category. Self-driving hotel rooms replace not just hotel rooms, but also tour guides and your holiday rental car. They drive you to all the tourist sites, they stop at affiliated restaurants, they occasionally stop at room service stations. And on the side, they do overnight trips from city to faraway city, competing with airlines.
Who wins: The first few companies who perfect this.
Who loses: Stationary hotels and motels.
Rise of alcoholism and drug abuse
Lots of people lack intrinsic motivation to be sober. They basically can't decide against taking something. Many of them currently make do with extrinsic motivation: They manage to at least not drink while driving. In other words, for a large number of people, driving is their only reason not to drink or do drugs. That reason is going away and consumption is sure to rise accordingly.
Hey I didn't say all the business opportunities were particularly ethical. But if you're a nurse or doctor, if you go into addiction treatment you're probably good.
Who wins: Suppliers of mind-altering substances and rehab clinics.
Who loses: The people who lack intrinsic motivation to be sober, and their family and friends.
Autonomous boats and yachts
While there's a big cost advantage to vehicle autonomy in cars, it is arguably even bigger in boats. You don't need a sailing license, you don't need to hire skilled sailors, you don't need to carry all the room and food those sailors require. So the cost of going by boat drops a lot, and there's probably a lot more traffic in (mostly coastal) waters. Again very diverse vehicles, from the little skiff that transports a few divers or anglers to the personal yacht that you rent for your honeymoon. This blends into the self-driving hotel room, just on water.
Who wins: Shipyards, especially the ones that adapt early.
Who loses: Cruise ships and marine wildlife.
Mobile storage
The only reason we put goods in warehouses is that it is too expensive to just leave them in the truck all the way from the factory to the buyer. That goes away as well, although with the huge amounts of moved mass involved this transition is probably slower than the others. Shipping containers on wheels already exist.
Who wins: Manufacturers, and logistics companies that can provide even better just in time delivery.
Who loses: Intermediate traders, warehouses and warehouse workers.
That's all I got for now. And I'm surely missing the most important innovation that self-driving vehicles will permit. But until that one becomes clear, maybe work with the above. All of these are original ideas that I haven't seen written down anywhere. So if like one of these and would like to turn it into a business, you're a step ahead of nearly everybody right now and I hope it makes you rich. If it does, you can buy me a beer. :-)