VisiCalc shipped in 1979. It was the first spreadsheet built for personal computers. It could do in seconds what took an accountant a full day, and its public debut drew little mainstream attention.
By 2015, US government jobs data cited by NPR showed around 400,000 fewer bookkeeping and accounting clerks than in 1980, and around 600,000 more accountants.
No press conference announced either number.
The launch was a footnote. What followed changed accounting in both directions at once.
That is the part that interests me. The launch is the least informative moment in a technology’s life. The information is in what happens after it lands.
Look at the work the spreadsheet removed. Change one assumption in a paper model and an accountant had to recalculate every dependent figure by hand. Once software did that in seconds, accountants could spend more time deciding what to do with the numbers.
Questions that had been too expensive to ask became a few keystrokes. What if we borrow more? What if we buy the competitor? What if we change the price?
The spreadsheet did not end number work. It made more questions cheap to ask. That helps explain why accounting jobs grew even as bookkeeping jobs declined.
The shift did not arrive as one event. It accumulated inside the work: less manual recalculation, more scenario modelling, fewer clerical roles, more demand for accountants. The job totals eventually recorded it. The news had no single event to report.
A similar kind of quiet signal may be appearing now.
Erik Brynjolfsson’s team at Stanford tracks payroll data for workers aged 22 to 25 in occupations with different levels of AI exposure. As of June 2026, employment among young workers in highly exposed occupations stood about 19 percent below where it would have been if it had kept pace with people of the same age in less exposed occupations.
Experienced workers in those occupations show no comparable gap.
The study is descriptive, not causal. It cannot yet say how much of that divergence generative AI created. The authors also flag education, pre-existing trends and differences between their sample and the wider economy.
What the data do show is that the adjustment appears to run mainly through reduced hiring of young workers rather than increased separations.
No mass layoff announcement. Just fewer people being brought in.
We watch for the spectacular version, mass layoffs, robots at the door, and can miss the quieter one: a hiring freeze here, a junior role that never gets posted there.
The comparison with spreadsheets is not clean enough to prove that the delay between launch and labour-market effects is shrinking. But the timing matters. Stanford documented a visible hiring gap within three years of ChatGPT’s release, and that gap has widened since.
That is early enough to watch in real time. If the pattern spreads, waiting for the retrospective will be a bad strategy.
I have been chewing on this question since 2019, when I wrote a piece about automation and reskilling and quoted Brynjolfsson saying people were falling behind because technology was advancing faster than our skills and organisations could keep up.
He is now one of the economists measuring part of the answer at Stanford.
So here is what this letter does. Every week, one technology, and the same question. What does it change for how humans work and live?
First, what the thing is. Then what has already changed, with receipts. Then where it may plausibly go for jobs and lives, with the reasoning shown so you can pull it apart.
The filter is the consequence, not the announcement. This newsletter is called After It Lands because that is the part I want to keep watching.
One thing I still cannot answer. As clerk jobs fell, 600,000 accountant jobs appeared. I wonder how many people moved from one group to the other.
That number would tell us more than either of the others, and I have never seen it.
I do not know what new roles this wave will create, how many there will be or how quickly they will appear. I also do not know whether the young workers missing today’s openings will have a path into them.
If you can see an answer from where you sit, in your team’s hiring plan or your own job description, I want to hear it.
Whatever lands on your job next, there will be no press conference.