cool, so the latest thing in digital audio recording is apparently the same idea i had recently:
32-bit float recording that can't clip.
with modern delta-sigma converters that we've been using for decades now, there isn't really a clipping point. they're just measuring voltage changes. the clipping only happens when decoding to PCM.
so there's no reason to limit the dynamic range to 16 or 24 bits. you can just use floating point numbers, and those have an insane numeric range and precision.
all DAWs use 32-bit floats internally anyway, so it makes perfect sense to also record in that format.