The loudest data point in the AI-and-jobs debate comes out of Stanford, and it isn’t encouraging if you’re a 24-year-old developer. Australia’s own numbers, published the same year, tell a different story, and Dovetail Software co-founder and CEO Benjamin Humphrey says his own hiring history explains part of why.
The US Story Doesn’t Match What Australia’s Data Shows
Researchers at Stanford’s Digital Economy Lab tracked millions of US payroll records and found something specific: employment for 22-to-25-year-olds in AI-exposed occupations like software development now sits 19% below where it would have landed had it kept pace with older, less-exposed peers, a gap that has widened steadily since the study first ran in 2025. Coding and customer service show the sharpest divergence. More experienced workers in the same fields show no comparable drop.
Australia’s Tech Council released its own Tech Jobs Report the same year, and it points somewhere else. The country’s tech workforce was recorded at close to 977,000 people by November 2025, and the Tech Council’s analysis found no discernible decline in early-career technical employment since generative AI became widely available in late 2022. Software engineering was the largest technical occupation in the report and grew 7% year on year.
His own hiring history offers one plausible piece of the puzzle, though not a full theory for the whole economy.
Dovetail Never Bought Into the Boom It Grew Alongside
Dovetail spent its early years building through the same cheap-money period that inflated hiring across the industry. “During that period of zero interest rates and high SaaS valuations, everyone was just spending huge amounts of money on hiring,” he says. “I don’t feel like that period was particularly notable for businesses caring about customers so much as they were caring about the growth of the employee count, and also deploying capital they’d raised in a cheap market.”
Dovetail’s own growth through that stretch was driven by something else entirely. Existing customers hired their own people and added licenses as their teams grew, rather than Dovetail expanding headcount to chase the market it was competing in. A company that never staffed up to match a hiring bubble has no bubble to correct for now that AI has arrived, and no junior cohort to quietly stop replacing.
That distinction matters for how the Australian data reads. A separate 2026 Stanford update found the divergence for young US workers had grown wider, not narrower, over the eighteen months since the first version of the paper, evidence the pattern is structural rather than a one-off correction. Companies that hired aggressively into the boom and are now trimming toward the model AI makes possible will show up in exactly the kind of numbers Stanford is tracking. Companies that never inflated their junior pipeline in the first place have nothing equivalent to cut.
Sydney’s engineering market rewarded that discipline rather than punishing it. A talent pool deep enough to hire well without a hiring war meant Dovetail’s headcount decisions were driven by what the business actually needed, not by what competitors were bidding for. The same logic that kept Dovetail out of the zero-interest-rate hiring spree is the logic Australia’s broader tech workforce numbers now seem to reflect: growth built on demand rather than on cheap capital and headline headcount doesn’t produce a generation of juniors who become disposable the moment the market tightens.
None of this settles the argument for every company or every country. Australia’s early-career numbers could still move as AI adoption deepens, and the Tech Council itself has said it will keep watching the data closely as adoption grows. But the story that AI is quietly emptying out the bottom rung of the tech workforce is, for now, a US story more than a global one, built on a hiring model Australia’s own tech sector never fully copied. Dovetail’s own decade of hiring without a headcount chase is one account of why that gap might exist, and it’s a harder story to tell than the Stanford numbers, because discipline rarely makes for the same kind of headline a collapse does.
#Australia #Tells #Story