When General Motors cut 600 IT roles last week, the company reached for a careful phrase. It called the move a “deliberate skills swap,” and said the plan is to replace those 600 people with a smaller number of AI-native workers.

Meta ran a similar strategy the same week. It cut as many as 8,000 roles and froze another 6,000 it had been hiring for, under the internal banner of an “AI efficiency push.”

If you have spent 20 or more years inside a large company, two words in there strike harder than the rest. Skills. Swap.

The instinct, reading a story like this, is to make it about yourself in the wrong way. You hear cost. You hear age. You hear a market that suddenly prefers cheap and fast over expensive and proven. A 2026 survey found that 71% of working people now expect AI to eliminate more jobs than it creates over the next three years, up 34 points since 2023. So the fear is widely shared.

But that reading of the story is the one that keeps capable people stuck. It treats the swap as a verdict on the people who were let go. Read the phrase more carefully and it describes something narrower. It tells you which tasks moved, and says nothing about anyone's worth.

What companies are swapping out is narrower than experience. It is the part of every role that a competent junior person, working with an AI agent, can now do.

That distinction carries more weight than it first looks. Inside almost any role with two decades behind it, two kinds of work sit mixed together. There is work that looks impressive but is really pattern execution: the reporting, the first-draft analysis, the standard recommendation, the meeting prep. And there is work that is actual judgment: the call you make when the data is incomplete, the read on a difficult client, the decision nobody can reduce to a procedure.

For 20 years those two were bundled into one job and one salary. AI agents are now pulling them apart. The first kind is being repriced toward zero. The second kind is becoming scarce, and scarce means more useful, not less.

Ken Griffin, the CEO of Citadel, called AI “garbage” a few months ago. More recently he said AI agents are doing work that used to take his PhD researchers months and finishing it in hours. His opinion of his researchers held. What changed was his view of which part of their work was the rare part.

So put the same lens on your own role this week. Take a blank page and make two lists.

List one is every task in your current job that a sharp person ten years your junior, working with an AI agent, could now do at 80% of your quality.

List two is everything that still needs the specific judgment your two decades built. The reads. The calls. The relationships. The decisions that would go wrong in less experienced hands.

List one is what your employer is repricing. There is little point defending it.

List two is your practice. It is the thing you can package, position, and sell as an independent expert. It is also the thing AI agents make you faster at rather than replace. The same agents hollowing out list one can run the outreach, the delivery, and the operations of a business built on list two.

A “skills swap” memo, read this way, is doing something useful for you. It labels in public which of your skills the market will pay a premium for and which it will not. Most people spend years guessing at that line. GM printed it.

The executives who struggle over the next few years will be the ones who keep proving they can still do list one. The ones who build something will be the ones who saw list two clearly and moved while it was still theirs to claim.

You do not need a restructuring to tell you which list your value sits on. You need an afternoon and an honest page.

Salama

If you are planning your move out of corporate in the next 3 to 6 months and want help drawing that line for your own role, I keep a few 15-minute strategy sessions open each week. No pitch. We map your list two and what a practice built on it could look like. Apply here.