important followups on "simulators"
https://www.lesswrong.com/posts/523mueiug9RapHtWb/language-models-can-be-utility-maximising-agents
https://www.lesswrong.com/posts/dYnHLWMXCYdm9xu5j/simulator-framing-and-confusions-about-llms
https://www.lesswrong.com/posts/FLMyTjuTiGytE6sP2/inner-misalignment-in-simulator-llms