Wednesday, October 14, 2026

What Needs Your Attention Today

Green means watching — we are tracking it and it looks normal. Amber means warning — it needs a closer look. Red means danger — time to act. The color never carries the meaning alone: every card says it in words.

So what? If you run a team, this is the week the "AI made us do it" memo went mainstream — and the week the paperwork to prove it started being written into law. If you are building a career, the market is still not firing people. It is quietly deciding which rungs of the ladder get rebuilt and which get removed.

The Picture in 60 Seconds

Think of AI as a tide that lifts the harbor while swallowing the beach steps. The harbor is the economy: investment, growth, and output rise. The steps are the entry-level and back-office rungs people used to climb into careers. Where the analogy breaks: a tide comes back in. A job ladder does not rebuild itself — someone has to decide to rebuild it.

The Big Picture: Displacement Leaves the Tech Campus

The claim: AI-driven job cuts are no longer a tech-sector story. They are now a corporate-America story, and Kroger is the clearest example yet.

The mechanism: a 143-year-old grocer does not cut more than a third of its corporate staff on a whim. CEO Greg Foran — the former Walmart executive who took over in February — announced the cuts after a companywide call on Monday, framing them as simplification amid slowing growth, heightened competition, and persistent inflation worries. His memo, viewed by Bloomberg News, says Kroger carries more layers and overhead than it needs, and that it will cut duplicative work and adapt as technology evolves, including rising AI use to make workflows more efficient. The savings go to lower prices, a better shopping experience, and investment in frontline employees. That is the pattern in one sentence: automate the office, spend the savings where customers can see it. Foran put it this way in the memo: "Simplifying means making faster decisions and deploying more resources where they matter most."

The evidence: roughly 3,000 corporate roles over three years, about 37% of Kroger's 8,200 corporate employees — one of the biggest cuts in the company's 143-year history — with details on affected roles to follow in the coming months. Bloomberg first reported the cuts from the memo it viewed; Grocery Dive later obtained the memo from Kroger, and Reuters also reported its contents. Kroger gave no timeline for the layoffs and did not say which departments will be hit. Kroger had already cut nearly 1,000 corporate roles last year — so this is an acceleration, not a first step. The competitive pressure behind it is measurable: Grocery Dive reports Kroger's market share has slipped while Walmart, Costco, and Aldi gained. Foran's response so far has been to cut prices — Kroger told Bloomberg in May it planned storewide price cuts — and to rebuild the top team, including a first e-commerce chief, a new chief store operations officer, a new chief people officer, and a chief data and AI officer. Shares fell about 3% on the news, to $60.74, per Bloomberg.

What's already baked into the numbers: the 37% figure. Nearly four in ten corporate roles disappearing within three years is not a pilot program — it is a restructuring plan with a number attached. If you work in corporate retail, procurement, or back-office coordination, that number is your planning assumption now.

What would prove it wrong: if Kroger's corporate headcount is roughly flat in its 2027 filings, or if the same roles quietly reappear under new titles. Watch the filings, not the memo. Two more tests: whether the promised price cuts show up on shelves and in earnings calls, and whether the coming months bring real role-by-role details rather than a slow drift of quiet exits.

Correction (Oct 14): an earlier version of this issue said the Kroger memo was seen by Reuters. The memo was viewed by Bloomberg News, which first reported the cuts; Grocery Dive later obtained the memo from Kroger.

The Global Economy: Growth Propped Up, Ladder Thinned

The IMF's October outlook — titled "Resilience Under Strain, Urgent Choices" and released Tuesday in Bangkok — puts global growth at 3.1% in 2026 and 3.4% in 2027, with inflation running hotter than expected at 4.7% this year before easing to 4.1% next year. The 3.1% is a small upgrade from the 3.0% the Fund forecast in July, but still below the 3.5% the world managed in 2025. The Fund is explicit that AI investment is a load-bearing wall in those numbers: the technology upturn is one of the forces that kept the world resilient through geopolitical shocks. Growth is also concentrating — 32 economies now produce 95% of global growth, down from 54 in April, as AI-linked and energy-diversified countries pull away from the rest. The concentration runs through the supply chain: the US and China account for more than three-quarters of net demand for imports of chips and other AI-related equipment. The US itself is expected to grow 2.3% in both 2026 and 2027; advanced economies as a group slow from 2.1% this year to 1.9% next year. And the forecast rests on a stressed energy baseline — oil prices up nearly 41% this year, averaging $96 a barrel.

Managing Director Kristalina Georgieva named the other side of that same coin on Monday. AI, she said, is a "slow-moving tsunami" for jobs: about 40% of jobs globally, and roughly 60% in advanced economies, will eventually be augmented, downgraded, or eliminated. Young workers entering the workforce face the most immediate danger, because the routine, repetitive roles that used to be first rungs are the ones AI absorbs first. A preliminary IMF study found about 10% of US jobs already require AI skills, while lower-paying work grows in construction and data centers. Georgieva added a fiscal warning that deserves more attention than it got: as gains shift from workers toward owners of capital, the tax base built on paychecks erodes with it.

Note what the Fund itself flags as the risk to its own forecast: a reassessment of expectations around AI sits on its list of downside risks, alongside escalating conflicts, renewed trade tensions, and fiscal vulnerabilities. If the AI investment boom cools before productivity gains arrive, the same force holding growth at 3.1% becomes the hole in it. That is the falsifier for the whole outlook: watch AI capital spending and the Fund's next revision, not just the headline.

Put the two halves together and the labor market's weird shape makes sense. The economy can grow at 3.1% while hiring freezes, because the growth is coming from investment in machines and buildings — not from adding headcount. That is why claims sit at 197,000 while payrolls crawl: the machine is running, and it needs fewer new operators.

The Entry-Level Split — and Two Governments Answer

The most argued-over corner of this market is the bottom rung, and the data still refuses to settle the argument. A paper from CESifo, a Munich-based economics research network, by Robert Fairlie and Jane Wu, using microdata from the Census Bureau's Current Population Survey — the monthly survey behind the official US jobs numbers — found unemployment among bachelor's degree holders aged 22 to 25 at 7.3% in summer 2026 — within the range of the previous four years. No apocalypse in the official numbers. The year-by-year path makes the point: 7.1% in summer 2022, 6.3% in 2023, 7.8% in 2024, 7.2% in 2025, 7.3% in 2026. Summer is the right stress test — this cohort's unemployment typically runs nearly two points higher in summer than the rest of the year — and summer 2026 still looks ordinary. The authors also tested whether graduates in AI-exposed occupations fared worse: point estimates were slightly higher, but the differences were small and not statistically significant. Nor did recent graduates lose significant ground against older college graduates (ages 30–49) or against 22-to-25-year-olds without a degree. Their broader "sidelined" measure — adding graduates who want a job but have not looked in the past four weeks, so they fall outside the official rate — hit 10.4%, the highest of the five summers studied, though again the increase was not statistically significant. Yet a survey of 1,000 hiring managers found 48% would rather invest in AI tools than hire and train a recent graduate, 55% have shifted at least part of their entry-level hiring budget to AI, and 45% have restructured so one senior worker plus AI does the work of multiple entry-level hires. Both can be true at once: graduates are still finding work, but the jobs they find are narrower, more selective, and more likely to demand proof of skill on day one. And the broad average hides a sharper split underneath: separate studies from the Census Bureau and the Federal Reserve Bank of Dallas, both published in September 2026, find graduates in the majors most exposed to AI — computer science in particular — slower to land a first job, and more likely, when they do find work, to land in lower-paying industries such as retail and restaurants, earning less. Fairlie and Wu cannot test by major in their data, and neither of the September studies includes the class of 2026 — so the major-level story is real but still one cohort behind. One more confound: unemployment was relatively higher among recent graduates in occupations that can be done remotely — but remote-capable jobs also tend to be the most AI-exposed, so the two explanations cannot yet be separated.

Two governments answered this week in opposite styles. California signed SB 951 on September 30: from January 1, 2027, any covered mass layoff driven in whole or substantial part by AI must say so at the top of the notice — "This notice is for a technology displacement" — and name the jobs, the functions being automated, and the category of technology. The state employment department will publish summaries and quarterly statewide reports, and must send the Legislature a one-time report by January 1, 2028 on AI's effects on hiring. For the first time, AI displacement will be countable in official data instead of arguable in anecdotes. Note the scope: SB 951 adds disclosure on top of the existing Cal/WARN machinery — it does not make every AI-related job cut a Cal/WARN event. Cal/WARN already requires 60 days' notice for mass layoffs of 50 or more employees in a 30-day period at establishments employing at least 75 people, and violations can draw back pay and benefits plus civil penalties of up to $500 per day. Two companion laws complete the package. SB 947, the "No Robo Bosses Act" — a narrower version of a bill Newsom vetoed last year — bars employers from making a disciplinary or termination decision by relying solely on an automated decision system; where an employer primarily relies on one, a human must corroborate the output, and if the human cannot corroborate it, the output cannot be used. Workers get a written post-use notice and, on request, a description of the data the system used. AB 1883 bars AI-enabled workplace surveillance that infers or predicts an employee's emotional state or collects neural data, and AB 1331 bars monitoring in workplace bathrooms.

Singapore chose carrots over paperwork. On October 12, the National Trades Union Congress and PwC Singapore partnered to help companies adopt AI, redesign jobs, and reskill workers, with grants covering up to 70% of qualifying project costs and up to 90% salary support for three months as workers transition into redesigned roles. The structure matters as much as the money: PwC joins as NTUC's newest Lead Multiplier for its AI-Ready SG initiative, bringing off-the-shelf AI solutions for procurement, inventory management, document processing, and resource planning; the reskilling runs through the Job Redesign Reskilling–Career Conversion Programme (JRR-CCP), administered by NTUC's Employment and Employability Institute (e2i); and NTUC's Queen Bee networks carry the literacy work — FairPrice Group, the Queen Bee for wholesale trade and retail, had already brought 27 suppliers together on September 22 for an agentic-AI workshop run by PwC. PwC's workforce survey gives the deal its context: 53% of Singapore workers see AI as a moderate or major threat to their job security, above the 44% global average — yet 72% say AI has improved the quality of their work and 63% say it has made them more creative. Adoption is already broad: 74% of Singapore workers used AI at work in the past year, up from 69% in 2025. The gap the deal has to close is support, not enthusiasm — PwC's survey found only 65% of workers said they had access to the learning and development resources they needed. NTUC Deputy Secretary-General Desmond Tan framed the bargain this way: "AI adoption must go hand in hand with job redesign, so that workers can take on higher-value work and gain better wages and work prospects." People are not afraid of the tool. They are afraid of being left alone with it.

What to Watch: The Narrow Moves That Still Pay

If you are building a career: the training-time gap from last week still stands — 58% of big-employer workers are offered AI training, 56% say they get no time at work to use it. The winning move is not collecting more course logins. It is one applied project inside your actual job, finished and visible, before the end of the month. Concretely: pick one recurring chore — say, the weekly status report or the first draft of a vendor summary — let an AI tool produce the first pass, check every number yourself, and show your manager the hours saved. That is what a hiring manager who would otherwise buy the tool instead of the graduate can actually see. If you are between jobs, the same project works as a portfolio piece: redo a real report for a company you admire, publish the before-and-after, and bring it to the interview.

If you run a team: California just changed what your layoff memo has to prove. If any part of a future reduction involves automation, start documenting now which functions the technology performs and which humans still own — that record becomes a legal filing on January 1. And SB 947 changes discipline, too: if a decision primarily relies on an automated system, a human must corroborate it before it can stand, and the worker gets a written notice. The companies that treat this as a paperwork exercise will write worse memos than the ones that treat it as a design question: what work do we actually want humans doing?

If you invest: training vendors are selling the cure for the freeze and staffing firms are priced for a thaw that has not arrived. Coursera's October 29 report is still the first number that separates those two stories. Watch whether enterprise learning revenue grows because employers are buying outcomes — Singapore-style, tied to redesigned roles — or because they are buying course catalogs workers never open. And watch the second-order effect Kroger exposes: corporate AI cuts shrink the office headcount that seat-based HR software is priced on. Every 3,000-role cut at a customer is seats that do not renew. When HR-tech vendors report, listen for seat counts and renewal rates, not just revenue — that is where a 37% corporate cut shows up on someone else's income statement. The policy split is a third thing to price: California's disclosure rules raise compliance costs and create a public record of displacement, while Singapore's subsidies pay companies to retrain — two different answers to the same problem, and training vendors exposed to one regime will not behave like those exposed to the other. Checklist for the next three months: (1) Coursera, Oct 29 — does enterprise revenue grow on outcome-based contracts, or on seat licenses workers never open? The former confirms the Singapore model is monetizing; the latter says the training boom is still hollow. (2) HR-tech earnings — do seat counts and renewal rates hold while customers cut corporate headcount? A seat decline at a vendor with heavy retail exposure confirms the Kroger spillover. (3) Thursday's claims — anything near 197,000 keeps the freeze thesis intact; a sustained move above roughly 250,000 would break it and change every staffing call. (4) California, Jan 1 — the first EDD quarterly displacement summaries will give this story its first official count. How to position around it: do not pay a growth multiple for seat-based HR software until seat counts stabilize — the Kroger spillover hits renewals before it hits headlines. Favor training vendors that can show outcome-based contracts tied to redesigned roles over vendors selling course catalogs by the seat. Treat staffing as a show-me story: no thaw, no premium — wait for claims to break sustainably above roughly 250,000 or for payrolls to re-accelerate before assuming the freeze ends.

Calendar: US jobless claims land Thursday, October 15 — the next read on whether the freeze holds. Kroger says details on affected roles will follow in the coming months. Coursera reports October 29. California's disclosure regime starts January 1, 2027, with the first quarterly state displacement summaries to follow and an EDD report to the Legislature on AI and hiring due by January 1, 2028.