Comment by qudat

Literally the model “mythos” is being marketed towards finding these exact type of bugs used for exploitation. I really don’t understand the argument: are agents not good at findings memory management issues? What’s the gap?