the problem is that autonomous driving and services like the high class busses will still take a lot of money out of the public transportation system, and if then they need to raise their prices, some people would loose their only means of transportation
It's hard to legitimately say which is cheaper. Buses don't need their own infrastructure, but they do need roads and these are often funded by various car taxes. Still, they often need subsidies to run. Many also have subsidized rates for poorer people on top of that.
Transport is generally expensive. I don't think it can be taken as plain fact that driving cars is the most expensive method, especially outside of dense urban areas. If you gave those same subsidies to poor drivers, many would be able to afford to drive.
Excluding taxes, driving a low cost vehicle has a fairly decent cost per-km. A litter of petrol can take you 15-20 km in an efficient, small car. Untaxed, that's <$1. If you travel 500km per month, that's about $30. Add $110 for purchase and repairs (cheap car), $40 for insurance and we are at $170 per month. $5.50 per day. Pretty close to the price of public transport in many European cities, maybe less.
This is driving relatively little, but most car owners just consume more transport (travel more/farther) that public transport users. Also, more than one person can ride.
It also doesn't take into account for infrastructure costs though, but public transport's purchase price often doesn't' either.
I'm not saying cars are better/cheaper, just that it's not a clear win for either mode. Transport is expensive. If you have no money, you can't afford much of it. We subsidize public transport for poor people and could do the same with cars.
Also most cities would not have the road-space to replace subways and busses by individual cars
I'm not qualified, but I'm guessing that the money will mainly be moved from private car transportation to these new transportation services.
and the other argument is that new "high class" bus services take money out of public transportation, which is badly needed there, but as you said, this may not add up to much. (However I could see myself switching from public transport to self driving cars, so they can also be included in this argument)
But both arguments go in the same direction - that public transport is definitely necessary and will remain necessary in the future but is often neglected and in a really bad condition.
Engineering cultures can be revolutionary even if the core ideas are the revolution. The startup community knows this well as it is repeated ad naseum - ideas are worthless (mostly), and its largely about the execution. Calling a cab from a phone is not new or original. Uber didn't think of it nor did they execute it first or probably even technically best. But their combination of technical execution and business acumen allowed them to raise shit tons of money, aggressive loss lead into users and grow internationally rapidly. And Taxis drivers in France are lighting cars on fire. Engineering itself can be revolutionary.
Second, and more pedestrian I suppose - It seems odd in the time of intelligence being move to the cloud, Waze and autonomous cars, to predict that the 101 will be a parking lot and investing in yesteryears mass transit is the right move. It would seem that autonomous fleets will be optimized to move people with less resources including energy, roadways, and vehicles. If anything one could argue that we should be pumping money and engineering into vehicle autonomy as its likely a better long term (something he should be able to get behind) investment than more lanes and asphalt. Sure we need to maintain roadways, but I would rather plan for more efficient and safer transport.