With all due respect, this is so far from the truth that it really takes away everything from the argument. How is computational photography any different than simply using other statistical methods to improve the signal-to-noise ratio? Is ISO and changing shutter speed any different? That also changes the actual thing.
Shutter and ISO are subtle and predictable variants that you know exactly what tweaking them will do. Computational photography is a black box.
It looks at the noisy pixels (those that aren't mapping to the rest of the area) and decides, based on AI modeling using "attractive photos" in their dataset, what those pixels should be so the image maps to those attractive images.
This can result in all kinds of suboptimal outcomes if you want a specific look other than that, whether that be as true to life as you can get it or something entirely different.