My impression was that, when botnet sinkholing was first practiced, a lot of care was taken over avoiding collateral damage, but if so, then the standards seem

My impression was that, when botnet sinkholing was first practiced, a lot of care was taken over avoiding collateral damage, but if so, then the standards seem to have decayed. This might have been a consequence of frustration - effective action seeming to be just out of reach - but I have also seen the fallacy that goes "we are the good guys, so therefore what we do is for the good." See also "normalization of deviance." This is why I suspect that, if the AI Apocalypse is technically possible, it is inevitable.