Ask HN: What would you put as the last parameter of this AI equation, and why?

10^22 estimated number of stars in our universe

10^20 estimated number of stars having planets

10^18 estimated number of star systems with planets having at least one planet in the habitable zone

10^16 estimated number of planets in habitable zones having life

10^14 estimated number of these planets having multi-celled life

10^12 estimated number of planets having evolved a life form analogous to the abilities of early pre-language primates

10^10 estimated number of those planets with life forms that use tools and language to a level analogous to early humans

10^8 estimated number of those planets with modern societies and technological capabilities analogous to late 20th-Century humans

10^6 number of those planets hosting civilizations that develop AI

10^?? number of those planets that developed AI, where the AI experiment went well both for the AIs and also for the civilizations that created the AIs.

All the other steps are using (for purposes of discussion, but it seems like a reasonable very rough estimate in aggregate) a factor of 100 at each step. Note there is no source for this factor, it is just a seat-of-the-pants estimate. Yes, I pulled it from there.

What factor do you think should be used in the last step, and why?

Or maybe you might think it's a flawed question: the "experiment" (for lack of a better word) may go badly for the creators, but may end up creating a fantastic and worthwhile/meaningful (not just paperclips all around) world for the AIs. Yes there's room for quibbling over "meaningful" but let's go with an admittedly unfounded assumption that human standards would be a decent starting point for judging what is meaningful and what is not... given that, again what should the estimated factor be for the last step, and why?