"Workers Are Getting More Done, and It's Not Because of AI"

There's a new article from The New York Times this morning that should make every knowledge worker pause their endless Slack scroll and sit up straight. Talmon Joseph Smith reports that American workers are more productive than they've ever been — and the gains have almost nothing to do with the AI tools dominating every tech conference keynote. The economy keeps growing faster than expected, even as immigration has slowed and waves of baby boomers continue retiring. The engine? Prime-age workers, 25 to 54, quietly getting more efficient at their jobs.

The framing matters because it arrives at an interesting inflection point. For the past two years, the dominant tech narrative has been that AI is about to transform productivity — that Copilot, Claude, and their ilk represent the next great leap forward. And individually, the early returns are real. GitHub users report coding 55% faster with Copilot. BCG consultants finish tasks 25% quicker with AI assistance. By the end of 2025, 41% of American workers had used generative AI at work, up from 31% a year earlier, according to the St. Louis Fed's tracker. But the macro numbers tell a different story: total factor productivity across major economies remains stubbornly flat, barely nudging above the 1.5% trend line.

The productivity engine isn't a language model. It's millions of people getting better at what they do every day.

So if the productivity gains aren't coming from AI, where are they coming from? Some of it is the mundane but real accumulation of better tools that predate the LLM boom — cloud infrastructure, collaboration software, automation pipelines that have been maturing for a decade. Some of it is organizational learning after the chaos of 2020-2022: companies figured out remote and hybrid work, streamlined their meetings, and stopped pretending everyone needed to be in an open-plan office to be effective. But a lot of it is simply people getting better at their jobs the old-fashioned way — experience, judgment, and the quiet efficiency that comes from doing something for long enough to stop second-guessing every decision.

This is also a soft but definitive rebuttal to the "quiet quitting" panic that consumed business media a few years ago. Remember when every LinkedIn post and CNBC segment was about how remote workers were coasting, disengaged, barely doing the minimum? The data never really supported that story, but it made for good outrage content. The current productivity numbers suggest the opposite: workers weren't slacking. They were just done with performative busyness, and it turns out that cutting the theater leaves more room for actual work.

The San Francisco Fed has been tracking a related phenomenon they call the Productivity J-Curve. The idea, borrowed from historical patterns around electrification and the internet boom, is that major technology adoptions initially depress measured productivity before they lift it. Firms keep legacy operations running while learning to integrate new technology. Factories in the 1920s used electric motors as direct replacements for steam engines without redesigning their workflows — it took decades before they reorganized around the new capability. The Fed's research suggests we're in the early, flat part of the AI curve, and the surge is still ahead. But here's the twist the NYT piece surfaces: we might not need to wait for AI to deliver because humans are already delivering.

There's a measurement problem lurking under all this that doesn't get enough attention. Most AI spending — training models, building datasets, developing software — gets expensed as a cost rather than capitalized as an intangible asset in national accounts. Meanwhile, 92% of S&P 500 enterprise value now sits in intangible assets. If you're spending billions building AI capabilities and accounting treats that like buying office supplies, you're systematically understating both investment and returns. The real productivity picture might be brighter than the official numbers show, and some of those "organic" human gains might have AI assistance baked in after all — just not in ways that show up in the statistics yet.

Another angle worth considering is demographic. The prime-age workforce that's driving these numbers is smaller than it used to be, thanks to an aging population and tighter immigration. Fewer bodies producing more output is a simple formula for higher per-worker productivity. That's good news for GDP per capita, but it also means the margin for error is thinner. A smaller, more productive workforce can deliver great numbers in a growing economy, but it's also more brittle — one sector-specific disruption or a public health shock hits harder when there are fewer people to absorb the blow.

There's also an uncomfortable wrinkle for the AI industry buried in these headlines. If productivity is climbing without AI, then the "adopt or perish" sales pitch that every enterprise SaaS company has been running loses some of its urgency. The argument that you need AI to stay competitive looks different when the evidence suggests you're already getting more competitive without it. That doesn't mean AI won't matter — the J-Curve suggests it will matter enormously — but it does mean the timeline for when it starts mattering at scale might be longer than the quarterly earnings narrative would have you believe.

The quiet story here is that American workers were never the problem the pundits diagnosed. The great remote work experiment, for all its messy implementation, didn't destroy productivity — it may have unlocked some of it. The obsession with monitoring, badge swipes, and "return to office" mandates was always more about control than output, and the data keeps suggesting that letting people work the way they work best is not just humane but economically rational. Whether AI eventually supercharges those gains or not, the headline worth celebrating today is simpler: people are good at what they do, and they're getting better.

As Talmon Joseph Smith put it in his NYT piece, after all the chatter about quiet quitting and remote work slackers, it's not even a matter of debate anymore — just check the scoreboard.

Sources: The New York Times — "U.S. Workers Are More Productive Than Ever. And That's Without A.I." (July 14, 2026); Informed Clearly — "AI Productivity Paradox: Why 2026's Tech Boom Isn't in the Numbers"; Fortune — "Why AI is raising worker productivity but not making the economy more productive" (May 2026); Bureau of Labor Statistics — "Productivity up 0.3 percent in first quarter 2026"