This claim by
@brian_armstrong misses a few fundamental realities of cybersecurity:
The Asymmetry: Defenders must secure 100% of the code, but attackers only need to find one single loophole.
AI is Probabilistic: AI isn't a mathematical proof; it can't guarantee zero vulnerabilities.
False Negatives: AI might be great at catching known bugs (Type 1 errors), but it will inevitably miss novel or complex exploits (Type 2 errors).
Because AI likely will always miss something, the dynamic definitely still favors the attackers.