AI productivity hype must meet reality, not fantasies

The Big Think piece makes a blunt, provocative claim: AI could trigger the biggest productivity boom ever. That’s a plausible headline, and a headline a lot of technologists are actively building toward. But plausibility isn’t inevitability; the article sketches an optimistic arc without digging into who gets the wins, how we’ll track them, or why earlier tech booms often left most people waiting at the station while capital rode first class.

Let’s start with where Big Think is right. Better tools usually do raise output. Teams that wire generative models into their workflows — from software development to marketing to legal drafting — really can shave hours off routine tasks and accelerate experimentation. I’ll be honest, watching a skilled group amplify itself with AI feels like reading William Gibson for the first time: you get a jolt of “oh, this is what the next decade might look like.”

But productivity is an aggregate story; paychecks are not.

Historically, the first-order gains from automation tend to flow to the entities that own the means of automation. If a model reduces labor hours, employers capture the savings unless wages, bargaining power, or regulation force some redistribution. That’s not a footnote to the productivity story; it is the story. Cloud computing made infrastructure cheaper and empowered startups, yet most of the profits pooled at a few hyperscale platforms. AI risks replaying that pattern: higher output on paper, fatter margins for owners, and median living standards barely nudging.

The Big Think argument mostly treats productivity as a tide that automatically lifts boats. But tides follow gravity, and our gravity is set by institutions.

Here’s the thing: productivity as economists record it — output per labor hour — is a blunt instrument for human welfare. Gains from digital tech often show up in corporate earnings and consumer surplus more than in neat productivity lines. Free or nearly free AI tools can improve day-to-day life while leaving official measures murky; automation that boosts output but funnels income upward can make those same measures look great even as workers feel squeezed.

That measurement mismatch matters because policy runs on dashboards. If official productivity spikes but wages stall, central banks, treasuries, and regulators are flying through crosswinds. A surge in AI-enabled output might not convince a central bank to ease up if labor markets still look tight or inflation is sticky. Lawmakers might assume that headline growth reflects broad prosperity and underreact on worker support or tax design. Big Think’s boom narrative doesn’t really grapple with that disconnect between what the statistics say and what households feel — and that gap is exactly where political blowback usually brews.

There’s also a missing chapter on how this plays out across industries. Some sectors — software, design, marketing, back-office operations — are already seeing AI as a force multiplier. Others, from logistics to manufacturing to healthcare, face heavier regulatory, physical, or organizational constraints. That staggered adoption means we shouldn’t expect a clean, synchronized “biggest boom ever” so much as a patchwork of mini-booms and laggards. And patchwork gains are politically fragile; workers in slower-moving sectors will still vote.

Policy is the lever the piece barely touches.

If you genuinely buy the premise that AI can unlock a huge productivity surge, you’re implicitly buying a second premise: outcomes will hinge on institutions. Education systems decide who can use these tools well. Tax codes decide how gains are shared. Competition policy decides whether a handful of platforms gate access. Labor law and norms decide whether workers have any say when their workflows are rewritten by prompts and APIs.

Sure, but the article treats the boom as mostly a tech story, not a political economy story. That’s like talking about the railroads without mentioning land grants, monopolies, or the people who got priced out of the new routes.

Concrete questions follow from its thesis:

• Do we update antitrust rules so a few AI stack owners don’t quietly meter and tax every downstream productivity gain?
• Do we strengthen worker voice so efficiency doesn’t just mean “do more with fewer people for the same pay”?
• Do we build real retraining systems so displaced workers can move into higher-value roles rather than being pushed into lower-paid gig work?

Those aren’t natural consequences of better models; they’re explicit political choices.

A common counterargument is that this line of critique is too gloomy. Maybe AI really does democratize expertise. Maybe the small design shop or family business can use the same models that a giant multinational uses and punch way above its weight. That’s already happening on a small scale: one-person operations producing work that used to require a small team.

I buy some of that optimism. But democratized access to tools is not the same as democratized ownership or bargaining power. When everyone can tap the same models, competition can intensify and margins can shrink at the edge, even as value accrues in the center to whoever runs the infrastructure, owns the data, or controls distribution. The history of digital platforms is a long lesson in “access up, take rates up.” And when that tension gets sharp enough, you don’t just get think pieces, you get populist backlash.

So here are the implications I wish sat alongside Big Think’s headline:

• If AI raises measured productivity, ask who captures the gains; headline growth won’t tell you.
• Measurement gaps will skew policy responses; better metrics and smarter institutions matter as much as faster models.
• The political decisions we make now — on competition, taxation, and training — will decide whether any boom shows up in household balance sheets or just in earnings calls.

Big Think does a useful thing by asking us to picture the biggest productivity boom ever. My bet: if we don’t tune the rails it runs on, the charts will look spectacular while the experience on the ground feels strangely familiar.