ETS Calls It Adaptability Paralysis, and Tech Workers Have It

ETS put out a report this week with a name for something tech workers have felt for about eighteen months: adaptability paralysis. American workers mostly understand that their jobs are changing, the report found, and they still can't act on it.

The reason is mechanical. People have historically adapted by switching employers, relocating, or going back to school, and all of those have become more expensive or harder to reach at the same moment.

We think that's the story underneath the daily run of cut announcements. The 2026 tech labor market still has jobs in it. What's scarce is movement. And when movement freezes, every individual bad bet costs more, whether it's a shrinking specialty, a declining metro or a skill stack that peaked in 2023, because exits that used to be routine now take a deliberate campaign.

Layoff survivors don't have the hours to retrain

The lazy reading of the ETS finding is that workers have got complacent. The other data this week doesn't back that up. HR Dive's roundup led with employees struggling to find time for upskilling, which is the more honest version: people are trying, and the hours aren't there. Layoff survivors pick up the work of colleagues who left, and then they're asked to learn a new toolchain on top of it.

Employers have more or less conceded the point. HCA Mag reported on a growing argument that employers owe their people AI skills even when those people leave, which throws out the old rule that training only pays if you keep the trainee. Cisco's AI Workforce Consortium is doing the same thing at industry scale for cybersecurity roles, treating the talent pipeline as something companies share rather than own.

Both are reactions to a market where keeping people and helping them adapt have come apart. Firms that need AI-capable staff can't hire them fast enough at the salaries LinkedIn is measuring, so they're being pushed to build the skills in-house and accept that some of those people will walk.

Moving to a better metro got harder

Relocating used to be the cleanest fix for a bad local market, and that option narrowed at the worst possible time. Return to office, as TheBanker put it this week, is no longer a soft issue. Attendance is enforced now, which means a job offer comes with an implied commute radius. Fortune covered a study of 7,700 employees finding that fully remote workers report the highest well-being, and ScienceAlert flagged new evidence that complicates the usual productivity case against remote work.

That research debate matters less than what's happening on the ground. Fewer roles are genuinely location-flexible, so where you live is once again a bet on one metro's direction, and those directions are pulling apart. New York passed San Francisco at the top of the closely watched annual tech talent ranking. Seattle held second with explicit warning signs, and a local economist on KING5 was left asking where the bottom is for the region's shrinking job market.

Seattle's recent run shows why. Starbucks cut another 224 jobs in the city, tech roles included, as its restructuring winds down. TikTok cut 75 in Bellevue as part of a wider 250-person reduction that also closed its Nashville office. Apple is reportedly trimming hundreds from its Siri and Vision Pro teams. None of that is catastrophic on its own. Together, though, it thins out the local demand that lets you job hunt in a city without having to leave it.

The jobs worth retraining for keep changing

Even if you find the time and money to retrain, you're aiming at something that moves. The New York Times documented where AI is creating jobs for humans and noted plainly that many of those roles are transitional. The Guardian's reporting on Hollywood creatives training the models that will replace them is the starkest example. Some of the best-paid adaptation work around right now eliminates itself once it's finished.

The destination jobs are narrower than the headline numbers make them sound, too. LinkedIn research puts average pay for AI jobs at roughly $177,000, but women make up just 26 percent of new hires into them, a gap HR Dive covered separately this week. A high-paying lane that recruits from a thin slice of the workforce can't work as an escape hatch for everyone else.

The argument we think holds up best against the data comes from Revelio Labs' Ben Zweig, speaking on CNBC. AI is changing the mix of tasks inside jobs faster than it's deleting job titles. That's good news if you can shift within your role, and bad news if your plan depends on a clean jump into a new one.

Looking out to roughly 2031, we'd bet on three things lasting, with the caveat that five-year calls are where we're most likely to be wrong. Training costs keep moving toward employers and consortia, because the public pipeline can't retool fast enough and the skills in demand now turn over on an eighteen-month cycle instead of a four-year one. Geography keeps concentrating, with New York gaining, along with (to a lesser degree) the metros taking in industrial and data center investment, while single-industry tech towns lose out. And the entry ramp into the best-paid AI work stays narrow unless someone widens it on purpose.

The policy response is scattered. Consortium models like Cisco's, employer obligations to fund portable skills and state-level workforce programs each do a piece of what a national retraining system would do. We'd plan on that system arriving late, which means your own timing will matter more than any institution's.

The way out of adaptability paralysis is to make the move smaller. The people getting unstuck in this market aren't making dramatic leaps. They're stacking small, checkable changes inside their current role until the next job looks like an obvious step instead of a gamble.

Start by working out which constraint is blocking you before you worry about skills. Most people roughly know what to learn. Fewer have mapped whether it's time, geography, employer policy or a missing credential holding them back, and constraints are often easier to negotiate than they look.

If it's time, ask for training in hours rather than budget. A funded course you can't attend is worth nothing, and two protected hours a week beat a $3,000 stipend. Take the employer-funded AI training even if you plan to leave, since the emerging norm treats that training as portable and employers have largely stopped pretending otherwise. If you're offered transitional AI work, like training a system, treat it as paid tuition and negotiate for the skills and references that will outlast it.

If it's geography, look at your metro's direction rather than its size. Seattle is still ranked second and still shrinking, and a market's trend line tells you more about your next search than its current headcount does. Expect location to matter more as office mandates harden into hiring filters.

Whatever the constraint, build evidence in public. Portfolios and certifications keep gaining ground on degrees for mid-career pivots, and in a low-mobility market hiring managers have fewer slots and less appetite for risk, so they screen on output they can see.