I’m very near to finishing the final book of Nassim Taleb’s Incerto series, and what you are saying there is straight out of Taleb’s playbook. In a nutshell, Taleb advises that to stay in the game, be sure to survive at every turn! As one example, he suggests that if airplanes landed successfully at 98% confidence levels, all professional pilots would be dead. When the outcome of tail risk is ultimately catastrophic, the first order of business is to eliminate any exposure to such tail risk to the greatest possible extent. For that reason, for example, I’d be perfectly willing to pause AI frontier model research for a couple hundred years or so. What’s the big hurry? Why are we accepting any tail risk at all in this area? If the downside is some species-level extinction event, what conceivable upside could justify that? To me this is pretty good evidence for the delusional impact of the Sant Mat five passions, (or equivalent expressions from other religious and philosophical traditions).