We may as well think of the spammers as their own ML whos aim and weighting is to fool the spam prevention. So as good as you make the spam AI, the spammer AI evolves to be better than it since they keep iterating until they find whatever formula does not trip the spam protection.
>Any human being or half-decent algorithm would be able to catch these with no trouble at all
And this statement from OP is blatantly false because if it was true, no human would fall for the scams. But they do fall for them, in huge volumes. Especially if the scams appear to come from a known authority like Elon, the police, your boss, etc.
And if you did find a half decent algorithm that worked, the spammers would just iterate again to make spam that doesn't trip it.
>Any human being or half-decent algorithm would be able to catch these with no trouble at all
And this statement from OP is blatantly false because if it was true, no human would fall for the scams. But they do fall for them, in huge volumes. Especially if the scams appear to come from a known authority like Elon, the police, your boss, etc.
And if you did find a half decent algorithm that worked, the spammers would just iterate again to make spam that doesn't trip it.