Artificial Intuition

by Carlos Perez

Book Reviews

  • Around a year ago, we discovered reversible induction machines. We know these now as diffusion models. The first kinds of these were discovered 7 years ago, I mention them in my book. to Tweet
  • The reason why System 2 is "slow" is because it is emulated using System 1. I've said this for years, yet people don't seem to grok it! to Tweet
  • I've decided to refactor my books to create a trilogy. I wrote 'Artificial Intuition' in 2017. This book will need to be updated to align with recent developments. to Tweet
  • The cart and the horse is an apt analogy. The cart is like representation and the horse is the intuition that drives the interpretation of symbols. Artificial intuition drives complex thought, but they employ symbols as a means to scale attractor states. to Tweet
  • IMHO, if you don't understand intuition and its relationship to deep learning, then you aren't on the right path. to Tweet
  • My journey into my explorations into general intelligence began with my book 'Artificial Intuition'. to Tweet
  • Glad to be one of the few people who kept saying that #deeplearning was an Intuition machine (i.e. System 1) several years ago. .Link to Tweet
  • @yudapearl A key building block towards empathy is intuition. Some (like Yoshua Bengio) would call this 'system 1' in the dual process theory framework: to Tweet
  • @GaryMarcus @FelixHill84 @nutanc @geoffreyhinton But a word of caution. There are many problems in the domain of modern human problems that can't be solved using an S2 system. But solving AGI isn't one of those problems. AGI is completely an S1 problem and we have biological evidence that this is true. to Tweet
  • Perhaps 95% of these books I haven't read when I wrote my first books and to Tweet