14-bit A/D converters and Exposing to the Right

The new Canons and Nikon released this fall are based on 14-bit A/D converters, meaning they now are capable of producing ≈16,000 discrete tonal values. Can someone who understands more about this technology fill us in on how this affects past discussions on this forum and Luminous Landscape about the strategy of "Exposing to the Right" (here, here, and here).

Tell me if I am wrong, but in a normal 5-stop Dynamic Range a 14-bit chip would produce 512 tonal values in the darkest F-stop and 256 levels in the stop below that. This is (literally) exponentially more information than was available with a 12-bit chip (128 and 64 values respectively). And blows away the paltry 8-bit chip in the shadow areas.

I assume this 14-bit chip is the reason Canon decided to include the "Highlight Tone Priority" function. This function (as revealed by others) simply under exposes the scene one stop by increasing the ISO, then bumps back up in post (In the RAW converter for RAW or in camera for JPGs). Isn't this effectively "Shooting to the Left"?

With a 14-bit A/D the camera can shoot a stop darker but still have twice the tonal range of a 12-bit chip, right?

Does all this mean that "Shooting to the Right" is now overkill? It would still give a broader range of tonal values to work with but is there a perceptible advantage to use the 8k stop when the next stop down still has 4k?

An ancillary question is what happens to the tonal ranges below 6 stops? are there still 64 tonal values 8 stops darker than "paper white"? Are those values salvageable?

Can a 14-bit chip get close to the High Dynamic Range images created by bracketing?

If someone understands this stuff please enlighten us!

The theoretical dynamic range of a 12 bit A/D converter is 12 stops, a 14 bit A/D converter 14 stops, and so on. That assumes, ofcourse, there isn't noise.
In reality, the noise (foton + readout) is limiting the DR to about 11 stops at best. The reason we speak in terms of 5 stops is because the monitors and papers have only this range.

But since the 40D presents no perceptible noise improvement over the 12 bit 30D, the 14 bit data shall be handled just like the old 12 bit data.

So if I have an effective 11 stops of information available from a 40D RAW image (even if my monitor or prints don't show it yet), couldn't that Dynamic Range enable me to play with the curve during RAW processing and compress the 11 stops of DR into a 5 or 6 stop range? (similar to HDR images that use multiple bracketed frames to achieve the same thing.)

A 14-bit chip and a 12-bit chip exhibit similar levels of noise? Does the extra tonal variations not provide more information to work with when trying to brighten shadow areas and therefore exhibit less noise? I know this must be false but I don't really know why? Is noise completely unrelated to dynamic range? I think of them as related but I guess they are not.