I regret that I work in a job which will, at some future point, require me to be one of maybe 2 or 3 people who have to think about this matter in order to confirm whether any damage has probably been done and maximize the chances of repairing the damage after the fact. No one who is not directly working on the exact code of a foomgoing AI has any legitimate reason to think about this, and from my perspective the thoughts involved are not even that interesting or complicated.

The existence of this class of basilisks was obvious to me in 2004-2005 or thereabouts. At the time I did not believe that anyone could possibly be smart enough to see the possibility of such a basilisk and stupid enough to talk about it publicly, or at all for that matter. As a result of this affair I have updated in the direction of "people are genuinely that stupid and that incapable of shutting up".

This is not a difficult research problem on which I require assistance. This is other people being stupid and me getting stuck cleaning up the mess, in what will be a fairly straightforward fashion if it can be done at all.