And to what 'exact' threshold of max block size this will have an real effect to the security ? 2 MB ? 4 MB 10MB ?
This is the more serious Q&A to be formulated.
It starts even at 2 MB. Please look into the original Bitcoin Classic. Gavin knew this was a problem, that's why the original proposal in Classic had additional TX size and sigops limitations (if I remember correctly). The higher you go the easier it becomes to attack via this vector.
If you're going to appeal to authority, let's just remember that this is the guy that claimed Craig Wright is Satoshi.
