The Nyquist figure is also tossed around in digital sound processing. Your sample rate must be, at minimum, twice the frequency of the highest frequency you intend to process.
As a guideline it's prominent today because it emerged as a usable concept as the technology was in development.
Not that it's a be-all and end-all guideline. Since it's possible for your sampling window to catch a waveform at zero crossings. Then you get nothing.
Or suppose the waveform is slightly different from half the sample rate. Then you get a subsonic rumble, or other such incongruous result.