AI didn’t cut 4,000 jobs at Block. A decade of cheap capital did.

Black Thursday at Block. A 40% reduction in workforce — over 4,000 people gone. One of the largest single-day cuts in S&P 500 history, as a percentage of headcount. Jack Dorsey says it’s because of “intelligence tools.” Wall Street loves the narrative and the stock popped 24%.

I used to work at Block, leading marketing, partnerships, and sales for bitcoin hardware products — Bitkey and Proto. I have enormous respect for the people there and for what the company has built. But I think the narrative around these cuts is wrong.

This isn’t an AI story. It’s a workforce correction wearing an AI costume.

Here’s what I saw inside Block: the Bitcoin hardware team was small, scrappy, mission-aligned. We had a single decision-maker, limited layers, and the full backing of leadership. We shipped new products in hard-to-explain industries because we could move fast. The team attracted some of the best people at the company — not because of the org chart, but because it felt like a startup. Everyone knew everyone.

Yet when I interacted with other parts of the company — the centralized functions, the shared services — it was a different world. Layers of approvals. People constantly out on PTO. A globally distributed team with employees in locations where the business didn’t even have operations — which meant a single approval could slip a week just on time zones alone. The contrast was stark.

Jack has let an organization swell before. Look at Twitter — a product that grew to world-changing scale while the company itself grew even faster, and financial performance went sideways for years. Whatever you think of Elon Musk’s methods, cutting 70% of that workforce didn’t destroy the product. Jack watched that happen from the board. Now he’s doing his own version at Block.

The math tells the real story. Block’s golden age — when the company shipped its most important products — was well before the pandemic. By the end of 2019, on its way into the bloated headcount era, the company had 3,835 employees and counting. By 2022, that number had more than tripled to over 12,000. The company acquired Afterpay during the peak for more than Block’s entire current market cap. When interest rates were near zero and crypto was surging, Square’s stock was flying — and the company hired like the party would never end.

The party ended.

Now they’re cutting back to under 6,000. Block isn’t alone — Salesforce eliminated 5,000 roles after their CEO said AI could handle 50% of the work. Amazon has cut 30,000 jobs across two rounds. Every one of them cited AI. But Challenger, Gray & Christmas data shows AI was directly responsible for about 55,000 of the 1.17 million total U.S. job cuts in 2025. That’s under 5%. The SF Standard is already calling the pattern “AI washing.”

So what’s actually happening? Companies massively over-hired during the zero-interest-rate era and now they’re correcting. AI is a convenient narrative that makes a painful restructuring sound forward-looking instead of overdue. Empire building became the default incentive structure — your seniority correlated with the number of reports you had, which created a gravitational pull toward growing teams whether the work demanded it or not. Jeff Bezos said the best teams can be fed by two pizzas. We lost the plot on that a long time ago.

And here’s the part I find most interesting: the functions getting cut hardest — marketing, creative, comms, BD — are the ones furthest from AI, not closest to it. Meanwhile engineering, the function most technically set up to be replaced by AI tools, is being supercharged by them and demand is almost limitless. That tells you this isn’t about machines replacing people. It’s about restructuring workforces that were bloated before ChatGPT existed.

A lot of large company work is performative. Meetings about meetings. Slack reactions so someone can prove they’re busy. Coordination costs that scale exponentially with headcount. The smaller the team, the less performative work is needed. The larger the team, the more you need to perform work to seem like you’re working. AI should, over time, expose which work is real and which is theater. That’s probably healthy. But let’s not pretend these cuts were caused by AI — they were caused by cheap money and misaligned incentives.

For the thousands of talented people affected — at Block and everywhere else — this isn’t abstract. These are real careers and real livelihoods. Many of the people I worked with at Block were exceptional. They deserve an honest accounting of why this is happening — not a narrative that makes it sound like a machine learned their job overnight. The remedy isn’t to fear AI — it’s to learn how to use it before someone who does takes your seat.

AI will transform work. It already has. But today, the biggest workforce shifts aren’t being driven by what AI can do. They’re being driven by what cheap capital made us forget: that leaner, focused, mission-aligned teams will outperform large, layered organizations that grew too fast on borrowed money and borrowed time.

I know because I lived it. A small team inside Block, building new products in hard industries, shipping faster than teams three times our size. Not because we had better tools — because we had fewer layers, clearer ownership, and a reason to be there. That’s always been the formula. AI didn’t change it.