For context, there are only around 20 humans above 3200 rating in the world. During the contest, there were only 21 successful submissions from 25k participants for that problem.
I guess we're past the point where codewords like "winnie the pooh" work but is NLP so advanced it can parse all the other ways human hint and indirectly communicate with each other?
Like can't I just send something like "hey coworker, lets do the thing that's opposite of ionized!". Or will the AI block my message (or more scarily, subtly rewrite it) and inform my manager?
This seems like tech with huge unintended consequences if perfected (even for their intended benign uses in workplaces) since it has the potential to be scaled up to censor society.