10x-or-0x-scorecard

"10x productivity" is the kind of phrase that sounds like it has a number in it, which is how it sneaks past your bullshit detector. But the number is never attached to anything you can weigh. Ten times what, exactly? Show me the before and after of a single hour, a single invoice, a single shift. Nobody does, because the metric isn't the point — the shot of it is the point. It's a mood ring worn by CEOs who want you to feel the future arriving before you've seen the delivery truck.

So let's build a scorecard that doesn't ask the CEO's question. The CEO asks: "Did AI make the company more productive?" That question is unanswerable, because "productive" is a word that does whatever the speaker needs it to do. The worker's question is different: "What did I get back?" That question is answerable. It has a ledger. It has pay stubs and clock-out times and the size of the check on the first of the month.

The scorecard is MECE because the rows don't overlap and the columns don't miss anything:

The promise: quote the claim. Write it down. "10x productivity" means nothing until it's a sentence a human said out loud.
Hours: did they go down? Not "flexible" hours. Down. 40 to 30 is down. 40 to 40 with more outputs is not down; it's faster.
Headcount: did it actually shrink, or did the same people just absorb more work? Reorgs aren't cuts. Attrition isn't a layoff. If the team is the same size and the work doubled, that's not a productivity gain, that's a wage freeze wearing a costume.
PTO and workload: is there a policy that says "you may take time off," or a culture that makes taking it a career risk? One is a line item. The other is a trap door.
Wages: up, down, flat, adjusted for what the company says inflation is? Flat wages with "10x productivity" is a transfer in disguise. The surplus went somewhere. The scorecard asks where.
Decisions: was the AI rollout presented to the people whose jobs it changes, or announced at them from a slide deck? This is the row that separates "we built this together" from "we did this to you."

Score it. 1–2: productivity theater, a stage show with no curtain call. 3–4: a cost somewhere, to someone — maybe it's the workers, maybe it's the customers, maybe it's the thing you'll find out about in a class-action filing. 5–6: a different system, the rare thing that looks like it might actually be what the word "progress" means.

The test case is the antiwork CEO who demanded 10x output and granted 0x PTO. The ratio that matters isn't the one on the earnings slide. It's the one that compares what the CEO keeps to what the worker gets to keep. 10 to 0 isn't a ratio. It's a confession.

Let's finish the rows. The CEO's claim has three more columns you can check without a finance degree. Pay: did the check get bigger? Not 'we're investing in your future' — the number on the stub. If the company got 10x productivity and your pay went up 2%, the other 8x went somewhere, and it wasn't your pocket. Benefits: did the insurance get cheaper? Did the 401k match go up? Free pizza is not a benefit; it's a bribe to stay late. And workload: did the number of tasks per human go down? Because if the same two people are doing the work of twenty and calling it 'AI-enabled,' that's not productivity gain, that's a headcount freeze dressed as a miracle.

Try it on a real claim. I once saw a CEO brag that AI gave his team '10x productivity.' The team's hours stayed flat. Headcount stayed flat. The company's revenue rose 30%. The workers got a 3% raise. Score: hours down? No. Headcount down? No. Pay up? No. That's a zero out of three. The claim is not a metric; it's a mood ring.

The scorecard doesn't ask whether AI works. It asks whether you got paid for the work it did. That's a question you can answer with a pay stub and a clock. Everything else is a speech.