So... you put in a flexible air tight seal in the tube. It will have to be modular anyway. Doesn't seem that challenging or cost prohibitive. A quick google turns up NASA projects that were looking into flexible airlock seals.

> near vacuum inside the tubes

This isn't true at all. The tube will still have air inside of it, you just need higher pressure behind the car than the air in front.

The cars themselves work a lot in the same way a cylindrical air hockey table would work.

The current hyperloop company has deviated substantially from Musk's design - now it's just narrow train in a semi-evacuated tunnel.

It won't be a complete vacuum though? I mean it doesn't have to be. Even removing 70% 80% of the air would greatly reduce drag. Plus don't they also need safety systems in case of issues so the tube segments can be locked and re-pressurized without damaging the train?

Drag reduction is, I think as important as decreasing the shock wave generated by breaking the local speed of sound. Partial vacuum does both. Not sure how to optimize for cost and reliability of construction vs max sped