1) Regardless of whether or not Uber is making an actual profit, which is not the point of the article, it is quite easy to show that Uber is withholding more revenue from drivers than theoretically necessary - on all rides, Uber takes a percentage cut of the overall fare, along with ~$1 or so in fees. You could argue that Uber's cut goes to running Uber itself, but the whole point of the article is that there is no need for a monolithic company running the platform in the first place - the software running it is fairly easy to create (given that there are hundreds of ridesharing apps worldwide) and the server costs are minimal - if the average Uber fare is $10 and it takes a 20% cut, it is ludicrous to suggest each ride costs $2 in server time. $0.02c might be more accurate (or even $0.20, if you add in other operational expenses), but that's a delta of $1.98 or $1.80 not going to the drivers for every fare.

2) Yes, to a company that is owned by the drivers and gives them back the maximal share of the revenue, taking into account the (minimal) operating costs.

3) To call Uber a "mediator" when both the customer and the driver have no say in the cost of the fare is laughable. Uber unilaterally sets the price, and if you don't like it, you're shit out of luck.

4) Like the point I made in 1), Uber takes a much larger cut than what is truly necessary. Even if the hypothetical ridesharing co-op had the exact same fares as Uber, a much larger proportion of the fare goes to the driver. And I would disagree that riders are the only constrained variable - there is almost zero switching cost between different ridesharing services for not only the rider but the driver as well. If only Uber and the ridesharing co-op are the only options for ridesharing in a given city, and they have the exact same fares for customers, but the co-op pays the drivers better than Uber, any driver would logically switch to the co-op since they would be paid more, and the customers would quickly switch to using the co-op's app as well, since the waiting times would be shorter compared to Uber.

I think your analysis it on point except for one part: the barrier to entry for becoming a driver is actually non-trivial and perhaps more importantly there is not a large supply of drivers in a given region. The pool of people who are looking for a new job, do not mind being a taxi and have a suitable car is somewhat small, and this is why uber and lyft must campaign heavily to recruit these folks. If this pool of people took collective action it would be very hard for Uber to come in and replace them - they could likely flip a single region to a new coop if they could self-organize.

That said the fact it could happen it not really related to if it should happen - as you say it's unclear how much extra money would actually flow to drivers as potentially Uber is operating on thin margins.