The AI productivity boom: real gains or a mirage?

The Fortune piece promises a lot in a single breath: productivity liftoff has begun, tied to a doubling in 2025, certified by one of Stanford’s original AI gurus.

A takeoff, a number, a guru. It reads clean. Too clean.

Before the takedown, grant the hope. If AI really is finally bending the productivity curve, that’s not a footnote — that’s the story executives, workers, and policymakers have been waiting on for years. The idea that invisible software could quietly raise living standards has been the techno-optimist’s favorite bet.

But hope doesn’t excuse vague math.

Start with the phrase at the heart of the headline: “productivity liftoff.” What exactly are we talking about? Output per worker? Output per hour? Firm-level value added? Or is it the sheen on venture-capital slides, where a single engineer “augmented by AI” is counted as ten?

The article offers a claim and an authority. It does not offer a repeatable metric. No definition, no time series, no way for an outsider — or a budget director — to test the idea a year from now. Strip away the headline and we’re left with an assertion, not a measurable phenomenon. Why would anyone rearrange national policy on that basis?

Here’s what they won’t tell you: jargon can masquerade as evidence.

Economists have argued for decades about how to capture productivity gains from new technologies. Software, the internet, and now AI often change quality and create new categories rather than just increasing the volume of old ones. When streaming replaced DVDs, the “output” didn’t show up as more discs pressed; it showed up as better access and different behavior. The official statistics lagged, sometimes for years.

That context matters when someone drops a phrase like “doubling in 2025” without anchoring it to a known, auditable series. Doubling what, exactly? Model parameters? Private investment in AI startups? The number of pilot deployments inside Fortune 500 firms? These are very different stories. Words without a ledger are useful for headlines and investor decks. Not for deciding tax policy or worker training budgets.

Then there’s the source architecture. The Fortune piece leans on one Stanford-linked authority. Experts matter. Universities do produce deep work. But when a narrative of sudden economic lift rides primarily on a single name and a single number, it starts to look less like reporting and more like a pitch.

Follow the money.

Who stands to benefit if markets and governments internalize the idea that productivity has already turned a corner? Startups framing themselves as the picks-and-shovels of a new boom. Venture funds that need a macro story to justify their paper valuations. Corporate leaders eager to tell shareholders that any short-term pain is the necessary prelude to an efficiency windfall. And yes, policymakers who want a justification to fast-track “innovation-friendly” deregulation.

Convenient, isn’t it.

The Fortune story sits in a long American tradition. In the 1980s, executives rushed to embrace “office automation” the moment consultants started promising leaps in white-collar productivity. Companies spent heavily on early PCs and proprietary software. Some workflows improved. Many didn’t. The economic data took its time, and when it arrived, the so-called miracle looked more like a modest makeover. The rhetoric had run several years ahead of the receipts.

The danger isn’t just intellectual sloppiness. It’s misallocated risk.

If productivity truly flips, capital allocation should shift toward retraining, durable infrastructure, and tools that broaden access to these gains. If it’s mostly PR, then capital chases illusions and leaves workers exposed. That’s why the absence of sectoral detail in the Fortune piece matters so much.

Some industries — cloud services, software platforms, back-office automation — are natural early beneficiaries. They have structured data, repeatable processes, and managers already primed to buy “efficiency.” Others — caregiving, creative crafts, construction — face messy environments, regulatory friction, and cultural resistance. A uniform “liftoff” headline papers over the reality that any gains will land unevenly, both across sectors and within them.

Who captures the upside? Who absorbs the shock?

Defenders of the Fortune framing will counter that early indicators are exactly that — early. They’ll say you spot turning points before the official statisticians do. That if models are scaling, gains will inevitably compound into the broader economy. There’s a grain of truth there: adoption curves rarely show up cleanly in quarterly reports.

But scaling model size and scaling social impact are different beasts. A more capable system might supercharge code generation at a major software firm, while leaving a nurse’s daily routine unchanged. It might boost margins without boosting wages. Without transparent, widely accepted metrics, declaring liftoff now is less a conclusion and more a bet.

Here’s the deeper concern. When powerful narratives outrun measurement, policy and corporate strategy often chase the story. Governments rush to “support innovation” with targeted tax credits and permissive oversight. Corporations accelerate automation programs to keep pace with perceived peers, not proven gains. Workers find themselves in the crosshairs of restructuring justified by a line in Fortune rather than a line in the national accounts.

Here’s what they won’t tell you: narratives shape incentives long before evidence catches up.

So what would start to make a claim like “productivity liftoff has begun” credible?

Not a bigger headline. Not a more famous guru. A few concrete things: transparent metrics tied to real economic output; corroborating analyses from multiple, independent institutions; and a sector-by-sector accounting of where, specifically, the gains are emerging — and where they are conspicuously absent. Without that, “doubling in 2025” is a slogan waiting to be stapled onto a slide deck.

I don’t dismiss the possibility that AI will meaningfully lift productivity. I’ve spent enough time with workers who feel these tools changing how they do their jobs to know something real is happening at the edges.

But when a single authority and a single number are asked to carry the weight of a historic turning point, that’s not analysis; that’s atmosphere. And atmosphere is exactly what sells when everyone’s been primed to expect liftoff.