thoughts on The Eureka Machine

Wow. Just wrapped up Richard Socher’s book The Eureka Machine. What a ride. I had the pleasure of spending time with Richard recently, and I love his framing of “there are plenty of AI doomers out there, what about the mind-blowing positive stuff?”

The book clearly articulated something that has bugged me since my teenage years: the reductionary drive of western approaches to science and technology loses so much. I hated engineering pedagogy so much in college that my friends and I created something we called The Academy of Unstructured Thought. Hung banners from our house, tried to foster absurd discussions through the night mashing up all of our respective disciplines.

Richard talks about AI’s potential to “weave the world back together,” scouring millions of publications across disciplines to interweave patterns, as well as create “alien hypotheses,” i.e. hypotheses mined in the data that show zero connections in human-authored articles. I find this deeply exciting.

Luckily artists have been doing this work for millennia.