AI jobs panic meets the labor data + Cal State's ChatGPT EDU renewal

An MIT Technology Review analysis published May 26 weighs predictions of imminent AI-driven white-collar collapse against actual labor market data and finds the case largely speculative. US Census figures show only one in five companies currently use AI in any business function. Stanford research found a 16 percent decline in entry-level jobs in AI-exposed occupations through 2025, but the effect concentrates on workers aged 22 to 25; overall coder employment still grows even as the growth rate has slowed about 3 percent since ChatGPT launched. About 40 percent of US workers use generative AI, per a separate Harvard survey cited in the piece.

California State University has renewed its OpenAI partnership at $13 million a year for three years, extending ChatGPT EDU access to roughly 470,000 students and 63,000 faculty across 23 campuses — still the largest deployment of ChatGPT inside any single organization. A 2025 systemwide survey found 95 percent of students using an AI tool and 84 percent using ChatGPT specifically, but 82 percent also said they worry AI will hurt their future job security. A faculty petition opposing the contract continues, with some campuses adopting AI tools without consistent guidance or training.

In this week's Stratechery analysis, Ben Thompson argues that Nvidia's revised reporting taxonomy — splitting hyperscaler revenue from everyone else — reveals the contours of an emerging commoditization fight at the top of the stack. Nvidia is "fighting commoditization" with its largest customers, where hyperscalers increasingly design their own silicon, while it "runs the whole stack" for AI clouds, sovereigns, and enterprises. The reframe lands one week after Nvidia's record $81.6 billion quarter on May 20, when hyperscaler revenue held at roughly half of the $75.2 billion data-center total.

The EAGLE Team, vLLM, and TorchSpec released EAGLE 3.1 on May 26, a speculative-decoding update that addresses "attention drift" by adding FC normalization after each target hidden state and feeding post-norm states into subsequent decoding steps. On Moonshot's Kimi K2.6 base model with vLLM, the team reports a 2.03-times per-user throughput speedup at concurrency one and a 1.66-times speedup at concurrency 16 on Nvidia GB200 hardware, plus up to twice the acceptance length in long-context scenarios versus EAGLE 3. The update maintains backward compatibility with existing EAGLE 3 checkpoints.

ClickUp CEO Zeb Evans cut 22 percent of the company's workforce on May 25, framing the move not as cost reduction but as a "radical embrace of AI" — replacing roles with about 3,000 internal AI agents that handle complex tasks under human direction. Evans pledged that most savings would flow into raises, promising "million-dollar salary bands" for the people who stay, and said "the people that automate their jobs with AI will always have a job." A Gartner survey cited in the same report finds roughly 80 percent of companies using autonomous technology have cut jobs, though workforce reductions have not necessarily produced meaningful financial returns.

The Vatican formally released Magnifica Humanitas on May 25, Pope Leo XIV's first encyclical and the Catholic Church's first major theological document devoted to artificial intelligence. The text argues technology "is never neutral, because it takes on the characteristics of those who devise, finance, regulate and use it," and calls for ending the AI arms race plus "adequate regulatory tools" to curb concentrated technological power. Anthropic co-founder Chris Olah presented alongside Cardinal Pietro Parolin in the Vatican's Synod Hall, signaling the Holy See's direct engagement with frontier AI labs.