Anthropic is scared of AI?

Scared of AI?  So is Anthropic.  A doomsday scenario is that recursive-self-improving AI may get to the point that it can produce AI babies even smarter and more capable than their parents, and those babies could produce more offspring, all at the “speed of compute,” so that humans can’t keep up.  Anthropic has released a white paper that provides another example of a recurring theme among AI developers: they’re scared about where it might go, not sure they can control it, and don’t know how to stop it.  They employ an arms race image that feels all too appropriate. 

 https://www.anthropic.com/institute/recursive-self-improvement

Lee