Squaring the AI circle

Literally every day I’m trying to square the circle of using an LLM. I’m not going to bullshit/lie and say I don’t use them, because I do. I avoid using it for actually important stuff (work or life stuff that I need to understand, learn, know, repeat, and care about), but there are admittedly scenarios where something I’m researching benefits from having more specific context. I try to use “responsible” toolsets for this like Kagi Assistant, but still feels wrong.