Only if you can completely isolate the balls so their states don't decohere. That is not practically possible to achieve, particularly since in your scenario you reach into the bag and touch the balls. As soon as you interact with the balls in any way, you become entangled with them and the behavior of the system becomes classical.

You didn't add anything to his example. This is just purely to be pedantic.

No. The only way you can actually observe entanglement is in an isolated entangled system (this is the reason quantum computers are hard to build). It is true that at a philosophical level there is no difference, but from the point of view of physics, which is to say, what is observable, isolation is crucial. Non-isolated systems behave classically, notwithstanding that they are actually quantum systems.

Would you claim that when Einstein developed his theories of relativity, they were invalid (from the point of view of physics) because their consequences were not yet observable? For example, Einstein used thought experiments to develop special relativity in 1905, but since kinematic time dilation was only experimentally confirmed in 1971, his work was not a contribution to physics until then?

The difference is that the limits on observing relativistic effects in 1905 were technological, whereas in QM you cannot observe the effects of entanglement in a non-isolated system even in principle. This is a fundamental constraint imposed by the theory itself. You can't get around it even with arbitrarily advanced technology.