Descriptions:
Matthew Berman examines the growing gap between AI hype and measurable economic reality, anchored by Sam Altman and Dario Amodei both walking back earlier predictions about AI-driven white-collar job displacement — a reversal Berman connects to both companies preparing for IPOs. Altman admitted being “pretty wrong” about entry-level job elimination, while Amodei reversed his claim that AI could eliminate 50% of white-collar roles. Goldman Sachs CEO David Solomon echoed similar skepticism.
Berman argues that the high-profile layoffs at Duolingo, Pinterest, Amazon, and Block were driven by post-zero-interest-rate-era overhiring rather than genuine AI-driven replacement. He cites the AWS CEO directly pushing back on the idea of replacing junior employees with AI, and notes that data on enterprise AI spend, hiring, and productivity all point upward simultaneously — an “absolute narrative violation” of the bloodbath thesis.
The video also tackles the ROI problem: Uber burned its entire 2026 AI budget in four months with unclear consumer-facing payoff, and one unnamed company accidentally spent $500 million on tokens in a single month. Berman contrasts frontier model pricing — Claude Opus 4.8 at $25 per million output tokens and Sonnet 4.6 at $15, versus DeepSeek at $0.87 — arguing that most enterprise use cases don’t require frontier-level intelligence and that companies competent enough to pick the right model tier will see costs fall significantly as infrastructure matures.
📺 Source: Matthew Berman · Published June 02, 2026
🏷️ Format: Opinion Editorial
