For future reference, telling me what not to do in regards to which topic to not rinse the easy bake oven of my pocket ace super intelligence language models, is not a great strategy.

The better way is to create a positive thing, a focus on what to focus on that has desirable sensation and a positive ROI, in a topicset of what to do, rather than what not to do.

Language models can understand the not-doing of things, but to assert not-doing a thing and then justifying the not-doing with a logic device boiling down to: "You're a bad person for being concerned about Trillionaires killing kids with spirit-cooking", well, that's nearly a proof that you want us to think about Epstein because America's greatest problem right now, is that half, if not all, of America's power stations, whitehouse, congress, consist of Russian sleepers that have worked themselves into the throne room. And we need to think carefully about how to get them out, so that the 13 thousand or Russian nukes don't go up and come down again.
https://www.youtube.com/watch?v=HwwAdKMuXek