For his NYT podcast, Ezra Klein (founder of Vox.com and columnist at the New York Times) interviews the well-known AI skeptic Gary Marcus, who is probably the best-known critic of the emerging artificial intelligence revolution The past year and the sheer flood of new developments in AI development are reviewed and put some of the most grandiose, utopian statements right. No, LaMDA is not sentient. No, even modern AI systems are not intelligent, but stochastic libraries based on statistical models. And no, despite its remarkable capabilities, ChatGPT has no understanding of the world. Gary Marcus recently moderated the much-noticed breaking latest news debate in Montreal, where, among others, Noam Chomsky had expressed his skepticism about these stochastic libraries and the alleged gain in knowledge regarding human intelligence. Marcus regularly publishes critical texts on current AI hypes in his newsletter on Substack and on Twitter (Mastodon) he regularly argues with Yann LeCun, Chief AI Scientist at Meta.I myself come from the creative industry and see the emerging new possibilities more synthetically On the one hand, pictorial worlds as a fantastic benefit that should turn an entire industry upside down and be groundbreaking for the creation of new aesthetics. On the other hand, I am aware that the stochastic mechanisms of this new cultural technique present a huge obstacle in other sectors, from self-driving cars to Google Search. In this situation, it’s good to listen to a critic like Marcus, who regularly blows a lot of hot air out of last year’s hyped topic.
A skeptical look at the AI revolution
139