Summary
David Shapiro revisits his “automation cliff” thesis, the idea that AI-driven labor displacement arrives suddenly rather than gradually. Drawing on his background as an infrastructure automation engineer, he argues that perfectly automated work scales almost without limit, and he updates his earlier forecast: the cliff does not hit the whole economy at once, but arrives one value stream at a time.
He supports the argument with examples including Klarna’s AI assistant handling 2.3 million conversations in its first month, Salesforce cutting support staff from 9,000 to about 5,000, Duolingo launching 148 courses in roughly a year, and Goldman Sachs drafting 95 percent of an IPO prospectus with AI. He also covers cases where automation backfired, such as Commonwealth Bank reversing 45 call center cuts and Klarna rehiring humans, and uses Andrej Karpathy’s “march of nines” to explain why reliability takes time.
The video closes with labor data: the US information sector down from 3.12 million to 2.74 million jobs since ChatGPT launched, call center employment at its lowest since 1994, and 120,136 job cuts attributed to AI so far this year. It also weighs Sam Altman’s remark about “AI washing” of layoffs. Viewers interested in how AI is reshaping work will find a clear, example-driven framework.
📺 Source: David Shapiro · Published October 29, 2026
🏷️ Format: Opinion Editorial
