Can anyone here comment on the kinds of things you can and can't learn accurately from a scaled prototype? I'd guess that for something that flies, changing the surface area/volume ratio makes a big difference.
Much research has gone into the applicability of scale models to real world conditions and the validity of a model is contingent on maintaining similarity coefficients. So for example you want your model to be flying at the same Reynolds Number as your full size aircraft will. If you can achieve those conditions then you can expect good predictive results from your model.
In general you can learn a great deal about your aircraft from a model including handling and stability characteristics but you're limited by the tunnel and the test apparatus. Post stall behavior comes to mind as something I would imagine to be very difficult to investigate with a scale model in a tunnel.
This changes the Mach number.
So no, in general, you can’t build and test a smaller model and transfer the results to a larger model using similarity, because your similarity model is incomplete.
You can either have similarity for the Reynolds number or the Mach number, but not both.
This is a tech demonstrator and not quite a "scaled prototype." Two occupants vs ~40 makes it pretty much a completely different aircraft. What it's probably demonstrating is some key design aspects like inlet/engine interactions, sonic boom mitigation, thermal resistance on various points, and overall validating simulation data. It won't show anything about extrinsic performance like range or takeoff distance of the final passenger jet, at least not directly.
The main problem with supersonic wind tunnels is ensuring there's no interference from shock waves bouncing off the sides of the tunnel, which is why the X-plane program used actual airplanes. So wind tunnels are rated to a certain Mach, and above that, like for hypersonic, actual airplanes are still used.
FYI: I believe the DC-3 was the first passenger airplane designed with a wind tunnel, and as a result was efficient enough to be the first twin-engine commercial passenger airplane to fly over the Rockies with a safety margin to fly single-engine, and adequate to be the first proposed Shuttle wing design.