“AI” skepticism went mainstream this week
After years of non-stop hype and bloviation, a flood of recent news stories suggest the tide may be turning on “AI”. Headlines this week announced AI is a “mass delusion event,” “a big nothing,” and even warned readers to “get ready for the crash.”
New research published in July and just making the rounds this week found that an extraordinary 95% of generative AI pilots at companies that rushed to implement them are bringing in zero returns, according to a new report from MIT.
But that hasn’t stopped the hypers. Sam Altman of OpenAI this week declared that although we’re in an AI “bubble” that investors are “overexcited” about, he’s still going to spend trillions on new data centers. “Smart people get overexcited about a kernel of truth,” he offered.
Looking for a fresh take?
At the Mystery Hype AI Theater 3000 podcast, we’ve been dismantling AI hype for a while now.
If you’re looking to get caught up, here are some of our best introductory episodes to help you separate “AI” fact from fiction and navigate bullshit mountain:
The ChatGPT Awakens: our take on the 2023 launch of the chatbot (spoiler alert: LLMs are “mathy-maths” and “AI” is always humans!) [Livestream, Podcast, Transcript]
AGI Funny Business (Model), with Brian Merchant: the author of Blood in the Machine: The Origins of the Rebellion Against Big Tech joined us to break down the history of the current “AI” hype wave and the business model behind it [Livestream, Podcast, Transcript]
Deflating Zoom's 'Digital Twin’: Remember when Zoom CEO Eric Yuan said in the future you’d be able to send an LLM-powered “digital twin” to work meetings instead of going yourself? We do! [Livestream, Podcast, Transcript]
The Anti-Bookclub Tackles 'Superagency': We read LinkedIn co-founder and venture capitalist Reid Hoffman's book of magical thinking, Superagency: What Could Possibly go Right with our AI Future, so you don’t have to [Livestream, Podcast, Transcript]
Our book, The AI Con, is now available wherever fine books are sold!
