The Split In The Job Numbers Is The Signal
The same news cycle carried two claims that cannot both be the headline. One set of reports says artificial intelligence is quietly cutting jobs. Another says it is creating them faster than companies can fill the roles. The split job claims ran alongside communities fighting new data centers and Washington drafting rules for a technology already loose in the economy.
Most people read that contradiction as noise, as if the analysts simply disagree and the truth sits somewhere in the comfortable middle. I read it as the actual shape of the change. AI is not adding or subtracting jobs at the level of the economy. It is doing both at once, in different rooms of the same building, and the aggregate number averages away the only thing a hiring leader needs to know, which is which room you are standing in.
The economy-wide job number is the one figure that tells you nothing about your own next hire.
This is what the split is really pricing. Roles built on routine production, the work that can be specified and handed to a model, are getting cheaper and thinner on the ground. Roles built on judgment, on deciding what the model should do and catching it when it is confidently wrong, are getting scarcer and dearer. Both trends are true at the same time. They simply land on different desks. A company that plans against the headline number will under hire the judgment it is about to need and over hire the production it is about to automate.
For hiring leaders, the practical read is simple. Stop planning against the national figure and audit your own functions one at a time, asking of each role whether it mostly produces or mostly decides. The producing roles you can staff lean and let the tools carry. The deciding roles are where the market is quietly moving against you, and the cheapest time to secure them is before the number inside your own building turns.
Planning a senior AI or data hire and want the market view first? We share it either way.
Send us the brief →