81,272 tech jobs gone and no plan in sight

Meta Cuts 8,000, Microsoft Offers Historic Buyouts — Both Cite AI Spend

• Meta will cut ~8,000 employees — 10% of its global workforce — starting May 20, while closing 6,000 open roles; CPO Janelle Gale's internal memo cited the need to "offset other investments we're making" in AI infrastructure, per Reuters and CNN.
• Microsoft simultaneously offered voluntary buyouts to ~8,750 U.S. workers — the largest such program in the company's history — targeting employees where "age + years of service ≥ 70," per Mint; both companies report Q1 earnings April 29 and have explicitly tied the reductions to AI data center investment.
• Forbes documents the 2026 AI-attributed layoff roster to date: Oracle (20K–30K), Block (4,000+), WiseTech (2,000), Atlassian (1,600), Snap (1,000), Meta (8,000) — each citing AI efficiency as the primary stated driver.
• Barron's reports 81,272 total tech layoffs in 2026 as of mid-April; a Harvard Business School study finds postings for repetitive/structured roles have dropped 13% since ChatGPT launched, while demand for roles integrating AI into strategy grew 20%.
• The Conversation notes that Meta grew headcount 14% in 2025 (reported last issue) — making its current 10% cut partly a reversal of recent expansion, not purely an AI displacement event.

---

Goldman Maps Substitution vs. Augmentation; Anthropic Tracks Job Anxiety by Exposure

• A new Goldman Sachs analysis formally distinguishes "AI substitution" (outright replacement) from "AI augmentation" (cost reduction that may expand demand): telephone operators, insurance claims clerks, and bill collectors face the highest substitution risk; education workers, judges, and construction managers show the highest augmentation potential.
• Interior designers and customer service reps share similar raw AI exposure scores — but interior design's unstructured tasks and on-site requirements give it far higher augmentation potential; Goldman identifies demand elasticity — not AI capability — as the decisive variable in whether a role contracts or grows.
• Anthropic's analysis of 81,000 free-form Claude user surveys finds a mean productivity gain of 5.1 out of 7 ("substantially more productive"), with scope expansion — doing tasks previously beyond reach — the most common gain (48%), ahead of simple speedup (40%).
• In the same survey, job-threat concern rises 1.3 percentage points for every 10-point increase in AI exposure — users in the top-quartile exposure bracket mention job-loss fears 3× more often than those in the bottom quartile; early-career workers report lower personal benefit (60%) vs. senior professionals (80%).
• Anthropic simultaneously launched a monthly Economic Index Survey designed to track how workers experience disruption in real time — on a cadence faster than traditional BLS labor statistics.

---

AI Anxiety Spreads to Hourly Workers, Indian IT Staff, and Chinese Employees

• A new PYMNTS Wage to Wallet Index (32,464 U.S. adults) finds more than 1 in 3 hourly workers say their employer introduced automation or AI in the past 12 months — with most reporting they received no training; lower-income hourly workers also have fewer emergency savings to buffer a disruption, making their financial exposure more acute than white-collar workers'.
• Bernstein's open letter to PM Modi warns India's 10–15 million IT services and BPO workers face "growing disruption from AI," while manufacturing stuck at 16–17% of GDP cannot absorb displaced labor at scale; Bernstein adds that data center investment will not save these workers because value accrues to global AI players, not India.
• Chinese employees are being directed by managers to document their full workflows in order to train AI agents via software like OpenClaw; some have built counter-tools that rewrite workflow manuals into "non-actionable language" to resist AI ingestion, per MIT Technology Review.
• BlackRock CEO Larry Fink warned that AI could leave the class of 2026 facing the highest unemployment in years — even without a broader recession; CNBC/ZipRecruiter finds recent grad unemployment at 9.7%, though 77% find jobs within 3 months by adopting "bridge job" strategies — accepting first offers regardless of fit.

---

Congress Has No Coherent Plan; Economists Debate Whether "Relational" Jobs Buffer the Blow

• A New York Magazine deep-dive (April 23) finds Senators Warren, Hawley, Brown, and Reps. Khanna and Raskin all acknowledge AI disruption could be "5× more transformative than the China shock" and could define the 2028 election — yet none has a coherent legislative package; cited barriers include pro-AI PAC donations and reluctance to expand the safety net under Republican control.
• Proposals circulating but unadvanced: OpenAI's 32-hour workweek + public wealth fund, Anthropic CEO Dario Amodei's suggestion of a special duty on AI companies, universal healthcare decoupled from employment, and free post-secondary retraining — none has cleared the proposal stage, per the same report.
• University of Chicago economist Alex Imas — whose research Morgan Stanley cites — now argues the demand-collapse scenario may not materialize: as AI commodifies production, value will shift to the "relational sector" — nurses, teachers, therapists, hospitality staff — where human presence carries a scarcity premium AI-generated goods cannot replicate.
• Imas cites Starbucks' reversal of some automation and re-emphasis on baristas and in-person atmosphere as an early market signal — though he warns the optimistic scenario hinges on AI adoption remaining gradual; a rapid acceleration beyond institutions' absorptive capacity would still produce the feared demand collapse.
• Carnegie Endowment researchers conclude AI may simultaneously hollow out jobs, reshape them gradually, and create new ones — and that "the case for acting now doesn't depend on resolving the debate."