5 comments on “AGI Foom

  1. Pingback: » Superintelligence Bayesian Investor Blog

  2. “The difference between the evidence from software and the evidence from biological intelligence suggests an important one-time transition to faster growth when software reaches human-level intelligence.” Could you explain your reasoning there a bit more?

  3. Pingback: Overcoming Bias : Irreducible Detail

  4. Robin,
    Some of the evidence about software suggests it’s improving faster than I expected. There are enough domains experiencing close to Moore’s Law type improvements on top of hardware improvements that I’d expect an AI that’s a collection of many domain-specific modules to improve faster than Moore’s Law in many domains even before accounting for increased ability to copy the intelligence.

    I’m being vague about whether that implies something faster or slower than what you’ve suggested for the next transition, since I’m rather uncertain about that.

  5. Pingback: Assorted links

Comments are closed.