Join us this Friday, March 27th at 3:30 PM for the March LA-AI Mobile Meetup @ Innovation Portal

Something shifted this week. Not in the technology itself, but in the conversation around it. For months, the AI‑and‑jobs debate has been this abstract thing that lived in think pieces and panel discussions. Smart people arguing about whether the robots were coming. Then this week happened, and it stopped being a debate.

On Tuesday, Meta confirmed it was cutting several hundred employees across Facebook, Reality Labs, and recruiting, amid record AI spending and a broader efficiency push. That came on top of reports that the company may be planning cuts as deep as 15,000—around 20% of its workforce—in phases through 2026. Mark Zuckerberg has framed 2026 as the year AI will “dramatically change the way we work,” and the stock went up about 3% on the news. On Wednesday, HSBC made news for weighing up to 20,000 job cuts tied to a multi‑year AI overhaul. Crypto.com slashed about 12% of its staff, with its CEO stating flatly that companies that don’t make the AI pivot immediately “are going to fail.” And that’s just the headliners.

Block’s massive layoffs from late February were still sending shockwaves. Roughly 4,000 people—about 40% of the company—were let go in a single move. Jack Dorsey called it an AI‑driven restructuring, not a financial downturn. The stock surged more than 20% in the immediate reaction. Andrew Yang has called it the largest percentage layoff in S&P 500 history, though that characterization depends on how you define the comparison set.

That stock reaction is the part that should make you pay attention. Not the layoffs themselves, because layoffs happen. But the fact that cutting 40% of your people is now the kind of move that can get a strong approval signal from investors. That’s a new incentive structure. And incentive structures have consequences.

Two big reports landed this week that put numbers behind the feeling. Cognizant updated its AI impact assessment and found that 93% of Fortune 500 jobs face some level of AI disruption, about six years ahead of its original timeline. A Duke‑led survey of roughly 750 CFOs projected about 502,000 AI‑linked job cuts in 2026, a roughly ninefold increase from the 55,000 AI‑linked layoffs estimated in 2025, though still just about 0.4% of the U.S. workforce. But nine times last year is nine times last year.

Goldman Sachs has been one of the loudest institutional voices on this all quarter. Their research suggests that millions of jobs globally are exposed to AI automation, with scenarios that could mean on the order of tens of thousands of AI‑related job losses per month in the U.S., depending on how quickly adoption accelerates. Their economists have argued that the big story in 2026 in labor will be AI. If displacement comes faster than expected, it could push the Fed toward rate cuts—and they’re already signaling openness to that path.

The loudest doomsday voices, though, are often the same people selling the tools. Sam Altman has said the world “is not prepared.” Dario Amodei and Mustafa Suleyman have echoed similar timelines. But these predictions come from companies that directly benefit from the anxiety. Marc Benioff at Salesforce has been less diplomatic, accusing some firms of “AI washing”—using the technology as cover for cuts that were coming anyway. On the other side, firms like Citadel Securities and other labor‑market trackers point out that demand for software engineers is actually up in the low‑double digits year‑over‑year. The truth, as usual, is messier than either camp wants to admit.

The Dallas Fed published some of the most useful research on this. Their finding: AI tends to substitute for codifiable knowledge (the stuff you learn in school) but complements tacit knowledge (the stuff you learn by doing). That’s why entry‑level workers are getting hit hardest. Workers aged 22 to 25 in AI‑exposed fields saw a roughly 13% employment decline since 2022. Young software developers saw about a 20% drop in headcount over a similar period. But experienced workers in those same fields are seeing wages go up. AI isn’t primarily replacing expertise. It’s shrinking the runway you used to get there.

That’s a real problem, and it’s one that doesn’t show up fully in aggregate unemployment numbers. The headline rate is around 4.4%. Not scary. But it hides the fact that entry‑level hiring at the 15 biggest tech firms appears to have fallen sharply, with some analyses put it around 25% in a single year. The pipeline into the middle class is getting narrower, even if the middle class itself hasn’t collapsed yet.

IMF Director Kristalina Georgieva has said plainly that policymakers need to “wake up,” because AI is for real. The jobs being created and the jobs being destroyed aren’t the same jobs, don’t need the same skills, and aren’t in the same zip codes.

So what do you actually do with all of this?

The skill that matters most right now isn’t knowing how to use any particular AI tool. It’s knowing how to evaluate what AI gives you (see last week’s blog post on this subject here). Can you bring the kind of judgment that only comes from experience, the tacit knowledge that AI can’t replicate? That’s not a technical skill. It’s a human one. And it might be the most valuable thing you develop this year.

The workforce isn’t cleanly splitting into “kept” and “replaced.” It’s splitting into people who got curious early and acted and people who waited to see what would happen.

You’re reading this newsletter. You’re already in the first group.

Know someone who’d enjoy this?

Don’t miss the next issue.