Answer-First Lead
If you want to know what AI does to white-collar jobs, look at the company building the AI. Anthropic’s CFO says 90% of their code is now AI-written and their financial reporting team has gone from weeks of work to hours. Meanwhile, the tally of AI-attributed layoffs in 2026 has passed 39,000. GM fired hundreds of IT workers to hire people with AI skills. Coinbase cut 14% of staff. And yet — Yale economists say the BLS data still doesn’t show AI-driven job destruction at scale. Two completely different pictures of reality. The truth is: both are correct, and the gap between them is the short window you have to act.
🔍 THE BOTTOM LINE
The Yale data is looking in the rear-view mirror; Anthropic is looking at the road ahead. Employment statistics take months, even years, to reflect structural changes. AI isn’t causing mass unemployment yet in the official data — but it’s fundamentally reshaping what dozens of job categories mean. The smart career strategy isn’t to panic or to ignore. It’s to assume that within 18 months, every white-collar role will require AI fluency as a baseline skill, not a differentiator.
📰 Stories
1. 🏭 Anthropic CFO: 90% of Code Is AI-Written — White-Collar Work Becomes Oversight
The story: Anthropic CFO Krishna Rao told Business Insider that AI now writes over 90% of the company’s in-house code. Most financial reporting work — tasks that used to take weeks — is completed in hours with AI assistance. Rao described the shift as “fundamentally changing what white-collar employees actually do,” moving from direct execution to oversight, review, and exception handling when AI output needs human judgement.
This isn’t theoretical. It’s the operating reality of a company that’s arguably the most AI-native major firm in the world. And it’s the canary in the coal mine for every office job.
Why it matters: When an AI company tells you their humans mostly review AI output, it’s not a prediction about the future of software engineering — it’s a current job description. The “engineer” title is splitting into two roles: the people who design the architecture and the people who review what AI produces. The same split is coming to accounting, legal, analysis, and every other execution-heavy white-collar role.
Sources: Business Insider, Yahoo Finance, Digital Today, DNYUZ
2. 📉 AI Layoffs Surge Past 39,000 in 2026 — Tracking the Toll
The story: Digital Journal reports that AI-attributed job cuts have surged past 39,000 globally in just the first five months of 2026 — on track to surpass the 55,000 recorded for all of 2025. The tracking includes all publicly announced layoffs where companies explicitly cited AI automation, workforce restructuring for AI adoption, or replacement of roles with AI tools as a factor.
The pace accelerated in May, with Meta (8,000), Cloudflare (1,100), Coinbase (14%), and GM IT cuts all within the same two-week window.
Why it matters: The number matters less than the acceleration. Each quarter, the “AI-cited cuts” number grows. Companies are no longer pretending layoffs are about “economic headwinds” or “restructuring” — they’re explicitly saying AI is the reason. That change in corporate communication is itself a signal: AI displacement is now a socially acceptable reason to cut staff.
Sources: Digital Journal, RNZ, Fortune, CNBC
3. 🚗 GM Tells IT Workers to Leave — So It Can Hire AI-Skilled People
The story: General Motors reportedly asked hundreds of IT workers to leave the company as part of a workforce reshaping strategy explicitly focused on replacing traditional IT skills with AI capabilities. Internal communications seen by multiple outlets said GM is “right-sizing IT for the AI era” and plans to hire more staff with AI engineering, data science, and machine learning operations backgrounds.
The move mirrors GM’s broader shift toward software-defined vehicles and AI-driven manufacturing — but the timeline for the workforce transition is months, not years.
Why it matters: GM isn’t cutting IT because business is bad. It’s cutting IT because the skill set of its current IT workforce doesn’t match where the company is going. This is the cleanest example yet of “churn” — not net job loss but net skill loss for the existing workforce. The people being hired won’t be the same people who were fired.
Sources: CNBC, The Register (linked from May 14), Fortune
4. 🪙 Coinbase Lays Off 14% of Staff — Crypto Meets AI Efficiency Pressure
The story: Coinbase announced it is laying off approximately 14% of its workforce, citing the need to streamline operations and increase AI-driven automation across customer support, compliance, and trading operations. The company said it would increase investment in AI tools while reducing headcount in roles that AI can handle more efficiently.
Coinbase’s CEO cited the need to stay “lean and focused” in a regulatory environment that has not yet fully clarified crypto rules in the US.
Why it matters: Coinbase has been through multiple layoff cycles (2022, 2023, now 2026). This one is different because AI automation is explicitly cited alongside regulatory uncertainty as a driver. Crypto companies were early AI adopters — and if they’re cutting, they’re sending a signal that AI efficiency gains are real enough to reduce headcount even in growing business segments.
Sources: TechCrunch, Fortune, Coinbase Blog
5. 📊 Yale Research: AI Still Not Disrupting Jobs at Scale (Yet)
The story: Researchers at Yale’s Institution for Social and Policy Studies published an analysis of Current Population Survey (CPS) data finding that, despite widespread AI deployment in many industries, there is no statistically significant signal of AI-driven job destruction at the aggregate level. Employment-to-population ratios remain stable, and industries most exposed to AI haven’t shown anomalous displacement patterns.
The researchers caution that this could change rapidly — historical parallels with industrial automation show that job displacement often lags technology adoption by 12–24 months — but conclude that the “AI apocalypse” for employment hasn’t arrived yet.
Why it matters: This is the “other side” of the story that doesn’t get as much coverage. If you only read tech news, you’d think the entire white-collar workforce is being automated overnight. The Yale data says: the aggregate economy hasn’t felt it yet. Both can be true — individual companies are making dramatic changes, but the overall labour market is a supertanker that takes a long time to turn. The warning: by the time the data catches up, it’s too late to prepare.
Sources: Yale ICF, NPR, Yale News
6. 🇳🇿 NZ Job Market: AI Causing Headaches for Both Job Hunters and Recruiters
The story: RNZ reports that AI is creating friction on both sides of the New Zealand job market. Job seekers say they can’t tell if companies are using AI to screen their applications — and feel dehumanised by the process. Recruiters say AI-generated applications flood their systems, making it harder to find genuine candidates.
A separate RNZ piece notes that demand for AI-related skills continues to grow — and older workers are “acing the pivot” more effectively than younger ones in some cases, surprising researchers who expected digital natives to have the advantage.
Why it matters: The NZ job market is being reshaped by AI at both ends: AI screening candidates, and AI-generated applications competing for attention. The result is a “signal-to-noise crisis” where everyone is using AI to filter AI output, creating a closed loop of mutual distrust. The skill that matters most right now isn’t AI literacy — it’s the ability to demonstrate genuine human value in a world where everyone sounds AI-generated.
Sources: RNZ (multiple pieces)