A not-so-amusing followup. I'm not feeling great today so instead of working on my startup (which requires real concentration) I decided to generate some realistic cost estimates for a comprehensive gene drive based mosquito control program. I would have then turned this into a post with some estimates on the feasibility of such a thing. Nothing I was discussing was sensitive or could have been used for anything like creating biological weapons. There wasn't even a discussion of the genetic mechanisms involved; it was all basically just population modeling, which is about as innocent as you can get. However, after cooperating for a while, ChatGPT shut down hard, as can be seen below. This is the sort of obvious and beneficial research question the AI safetyists have denied us being able to ask. Knowing whether realistic control and elimination of dangerous disease bearing insects is economically feasible is not harmful to anyone, it's a positive good, but I'm not allowed to ask because of safetyism pushed by the Doomer community. Suggestions on what open source model I should try working with instead are solicited. I'm increasingly uninterested in jumping through hoops to deal with the twisted idiocy of the "safety" community and just want to do work without being harassed or moralized at.