EFTA01019562.pdf

From: John Brockman To: Jeffrey Epstein <jeevacation@gmail.com> Subject: "Enlightening, entertaining, and exciting" Date: Tue, 16 Oct 2018 01:18:10 +0000 How are you? Heading to mideast?? First review of POSSIBLE MINDS in PUBLISHERS WEEKLY. Pull quote: "While the [POSSIBLE MINDS] authors disagree on the answers, they agree on the major question: what dangers might AI present to humankind? Within that framework, the essays offer a host of novel ideas. . . . Enlightening, entertaining, and exciting reading." Instead of Carnegie Hall how about an evening at the Brattle Theatre in Harvard Square week of Feb 19th. (The publisher may even pay for it.) PUBLISHERS WEEKLY "Possible Minds: 25 Ways of Looking at AI" Edited by John Brockman. Penguin Press, $28 (320p) ISBN 978-0-525-55799-9 https://www.publishersweekly.com/9780525557999 Brockman (THIS IDEA IS BRILLIANT, editor), founder of the literary agency Brockman Inc., devotes this entry into his series of science-oriented essay anthologies to exploring the frontiers of artificial intelligence. The 25 contributors come from a wide range of disciplines and include philosopher Daniel Dennett, psychology professor Alison Gopnik, and Skype cofounder Jaan Tallinn. While the authors disagree on the answers, they agree on the major question: what dangers might AI present to humankind? Within that framework, the essays offer a host of novel ideas. Several argue that Al has already become a hallmark of human culture, with genetics researcher George M. Church provocatively suggesting that modern, technology-using humans, when compared with Stone Age cultures, are already "transhumans." Other essays underscore the necessity for ensuring that advanced AI acts in alignment with human values, while science historian George Dyson explores the difficulties inherent in controlling the technology, ending with the unsettling observation that "provably `good' AI is a myth." Readers will appreciate that the discussion is accompanied by intriguing explanations of AI development strategies, among them "deep learning," generative adversarial networks, and inverse- reinforcement learning. The combination of theory and practice makes for enlightening, entertaining, and exciting reading. (Feb.) EFTA01019562