Sure, but a NeuralHash system with a collision chance of 100% would obviously not be abused to say NeuralHash collision implies CSAM - the secondary validation system will definitely be the last bastion.
In contrast, a NH system believed to have a collision chance of 1 in a trillion trillion may well be considered infallible, and any detection be directly reported as CSAM, with the 'backend verification' amounting to nothing more than a rubber stamp.
Of course, if you implicitly trust Apple not to do the second, than you're right, the NH collision rate doesn't matter too much.
In contrast, a NH system believed to have a collision chance of 1 in a trillion trillion may well be considered infallible, and any detection be directly reported as CSAM, with the 'backend verification' amounting to nothing more than a rubber stamp.
Of course, if you implicitly trust Apple not to do the second, than you're right, the NH collision rate doesn't matter too much.