hydrargyrum,
I call BS on that article. Having heard the difference between the same song recorded at the same sample rate but different bit depths, I can say that 24 bit does make a big difference.
The author is correct that most systems can't reproduce ultrasonic frequencies, and that most people can't hear the entire 20Hz-20KHz range. The misinformation lies in that increasing the sample rate also decreases the sampling period (1/the sample rate). This allows for more samples per second and increases resolution.
The ability to playback in 24 bit is what has me most excited about Pono. The clarity in the bass range and lower noise floor are very noticeable. Increasing the bit depth, like increasing the sample rate, reduces rounding error (the system is only capable of encoding a set number of values - 16 or 24, respectively). Increasing the bit depth alone changes the quality enough that, if any user is paranoid about intermodular distortion ruining the playback system, the music can still be enjoyed at a better fidelity than an mp3.