With this argument you can dissolve anything in "why even bother?". In the real world it very much makes a difference, if the breach of trust is secretly implementing a side channel, or secretly using a documented one. The latter e.g. can maybe by activated "by accident" plausibly, but having a backdoor at all is hard to justify. And for a company like Apple, the size of a breach of trust matters for financial incentives.
If you can't trust that, then you can't trust the whole OS and all this speculation has been as valid as any given time.