Dayton, OH Correspondent — Both, and the historical pattern with general-purpose technology is net gains over time with significant transition costs for specific occupations.
What the early studies show is interesting and not what either camp expected. The largest gains go to the least experienced workers — the tool raises the floor more than the ceiling. A new customer-support agent with the tool performs like an experienced one; the experienced one gains little.
If that holds, the displacement lands on the middle of the distribution and the compression is upward from the bottom, which is an unusual shape and nobody’s policy assumes it.
Orange County, CA Correspondent — Gains are visible for effective adopters and displacement is visible in specific tasks.
The financial observation is that nobody has yet demonstrated this in a productivity statistic. Adoption is enormous and measured output per hour hasn’t moved in a way anybody can attribute.
That’s normal — the electrification lag ran decades — and it’s worth stating, because a great deal of capital is being committed against a return that hasn’t appeared. If it doesn’t appear in three years, the investment cycle corrects hard and the technology continues anyway.
Novi, MI Correspondent — Tools that raise output get adopted and tasks that can be automated will be.
Engineering is where I see it. Design iteration that took a week takes an afternoon, which doesn’t eliminate the engineer. It eliminates the reason to hire the third one.
That’s the shape of the displacement in my industry. Not layoffs. A hiring plan that quietly stops growing, which shows up in nobody’s statistics and in every graduate’s job search.
Myrtle Beach, SC Correspondent — It makes some people much more productive and takes pieces of other people’s jobs, and both are already happening in mine.
Marketing copy, first-draft campaign concepts, customer segmentation. I do work in an afternoon that took a week, and I’m not hiring the junior I’d have hired.
The honest part is that the junior was how somebody learned this trade. I learned it by writing bad copy for two years until it got better. I don’t know where the next person does that now.
Long Island, NY Correspondent — Both, in sequence. First it eats the draft, the summary, the first memo, the junior research pile. Then firms that use it as a multiplier show a productivity number and firms that only cut the junior show a cheaper mess.
The legal profession is a case study in progress. Document review was already largely automated and the associate hours that used to fund training went with it.
What’s coming is the first-draft brief, and the profession’s answer so far is that a human must sign, which is a liability rule rather than a training system. A signature isn’t an apprenticeship.
If schools keep selling a credential that is precisely the first draft, they’re selling something the market just repriced.
Las Vegas, NV Correspondent — A floor already automated the easy count and now the easy paragraph is next.
Productivity appears when the remaining human handles exceptions and guests. Displacement appears when you delete the trainer and hope the model greets somebody important.
Net is a management choice rather than a technological outcome, and most managements will choose the quarter.
Tyler, Texas Correspondent — The shop will use it to draft a bid and still need somebody who has stood on a roof.
Mid-skill paper jobs feel it first. That isn’t a reason to ban the tool. It’s a reason to stop financing four years that only produce the paper.
The insurance angle is that I’m being asked to underwrite firms deploying this and there’s no loss history, no standard of care, and no settled answer on who’s liable. So the market declines it or prices it badly, and that’s the actual brake on adoption right now.
Gastonia, NC Correspondent — My trade is the test case and the answer is already partly in.
Publications are using this to produce volume, and the volume is worse. Several outlets ran generated material this year, got caught, and apologized, and the thing that caught them was that it was wrong in ways a person wouldn’t be wrong.
Where it’s genuinely useful is the unglamorous middle — transcription, first-pass headlines, checking a fact against a source. That’s real productivity and it’s also the work that used to be somebody’s first job.
Owatonna, MN Correspondent — Both, with a longer lag than the enthusiasm suggests.
I’d offer the freelance view since that’s my market. Rates for general commercial writing have already fallen and the work that’s holding is anything requiring a person to have been somewhere or spoken to someone.
That’s the durable line and it’s clarifying. The tool competes with what can be written from other writing. It cannot compete with reporting, and it turns out a substantial share of what we called writing was the first thing.
Bismarck, ND Correspondent — Both, and rural areas will get the displacement earlier than the productivity.
The back-office work that got moved out of coastal cities to places like this over twenty years — claims processing, transcription, service centers — is precisely the category most exposed. Those jobs came here because they were cheap. They’ll leave because they’ll be cheaper still.
That’s a specific regional risk and nobody in the national discussion is describing it.
Sydney, Australia Correspondent — Both, and the international dimension is the one that reaches us.
A great deal of English-language service work migrated to the Philippines and India over two decades. That work is more exposed than anything in a wealthy economy, and the countries that built development strategies around it have no plan.
The displacement conversation in America is about a graduate’s first job. Elsewhere it’s about a national export sector.
Prescott Valley, AZ Correspondent — Both, and the education response is the part I’d press.
We spent twenty years telling students that the safe path was cognitive work rather than trades, on the theory that the machines came for hands first. That advice is now inverted and the institutions giving it haven’t adjusted.
Whatever the net effect, the transition cost lands on people who followed guidance we gave them. That obligates the people who gave it, and I include myself.
