Is the AI Layoff Wave Real?

There's a hot debate going around software circles lately: is the AI-driven layoff wave actually real? Opinions split into two camps. One side says it's already happening — the data shows U.S. software engineering job postings are shrinking, which spells long-term trouble for anyone squeezing into computer science programs. The other side says the data being cited is flawed, that it ignores broader macroeconomic factors, and that the whole premise is false — demand, they argue, is actually rising. Honestly, this kind of debate has the flavor of an annual rerun: once or twice a year, everyone rehashes whether AI, automation, or software in general has finally come for our jobs.

The Layoff Debate Heats Up — Both Sides Have Receipts

Both camps actually have some evidence behind them. Let's start with the more visible layoff wave. From my trip to Silicon Valley two years ago through this year's trip to San Francisco, I kept hearing local friends bring up layoffs — the sense of it was palpable. Tech companies also can't keep a secret, so layoff numbers are essentially public information. According to TechCrunch's tally, roughly 70,000 people had been laid off by May of this year alone. And TechCrunch skews heavily toward software and startups — old-line "legacy" companies barely register on their radar — yet just about every software and internet company you can think of, big or small, new or old, shows up on that list. In April alone, 22,000 people were laid off across the U.S. software and internet industry.

Among the mountain of tech-layoff headlines, two companies stand out to me as particularly interesting. The first is ProductHunt, the site where the software industry launches new products — it cut 60% of its staff in the second half of last year. That's a layoff ratio nearly on par with Elon Musk's cuts at Twitter; it's genuinely unclear whether a mid-sized company can survive a cut that deep, but ProductHunt did it. And keep in mind, ProductHunt should be a beneficiary of the AI boom — AI companies need to advertise their products somewhere, so ProductHunt should be riding that wave. The company itself said its finances were actually solid and site traffic was running high; the cuts, it said, were purely a "strategic decision." What was that strategic reasoning? I think most people can guess: AI replacing human labor.

The other company is Duolingo. Plenty of people in Taiwan use it too — the app is user-friendly, most of its features are free, it teaches languages from all over the world, and its social media game is excellent, with an adorable owl mascot whose memes are everywhere. Duolingo employs a lot of writers and translators, and last year it had 80 million users and revenue of more than NT$16 billion (roughly $500 million), beating expectations, with its stock hitting new highs again and again. Life seemed good — surely its employees were safe?

But last year Duolingo ran two rounds of layoffs, cutting a large swath of the staff who wrote and produced course content — and it looks like more cuts are coming. What's worse, some of the people who were kept on were kept specifically to train AI on that content, which is essentially digging their own graves on the clock. Duolingo is about as clean an example of AI replacing humans as you'll find.

Here's another example closer to home. With the latest round of updates from several major AI players, a lot of people have discovered you can now build a website with essentially zero coding background. Sure, that was technically possible before too, but it's increasingly starting to feel like you barely need to think at all to pull it off now.

Building small websites has always been one of those fallback jobs a software engineer can lean on. Here's an example that may not be perfectly representative: an engineer I know has kept himself afloat for years — nearly long enough to retire — just building websites for small bed-and-breakfasts in a rural township. These aren't exactly high-powered clients, but the demand is real. Once that kind of demand can be met almost instantly with AI — say, the B&B owner's kid learns to do it with AI at school — the people who served that small, niche demand will lose their jobs too.

Facebook, with nearly 70,000 employees, and Duolingo, with only about 700, are both cutting staff — and now even a B&B owner can build their own website. Are humans going to be left with nothing but jobs like wearing Duolingo's owl mascot costume? (And even that job probably won't be safe for long — a humanoid robot won't complain about the heat inside the suit.)

At the same time, though, the other camp has its own evidence: AI is genuinely hot right now, and that's driving demand for labor up, not down. Top consultancies like BCG and McKinsey have seen orders surge, because so many large companies need help implementing AI — for the moment, these firms have more work than they can handle.

Layoffs and Hiring, Side by Side — Employers Just Want Different Skills

Here's my own take, from the ground. I work in the internet industry, and right now the field actually needs more engineers, because in the short run, the changes AI is bringing have increased demand — there's simply a lot of new work that needs doing to adapt to it. So it's entirely understandable that big tech companies are often laying off and hiring at the same time — cutting people even as they're recruiting — it's just that the people going out the door and the people coming in aren't the same people. That said, I wouldn't rule out the absurd scenario where someone gets fired because their company decided it needed "AI talent," only to land a new job at another company by claiming they know AI... Silly as that may sound, it can happen. But for now, as far as I can tell, overall job demand is growing — even people who get laid off can largely be absorbed by new demand elsewhere, provided they're willing to pick up some AI skills along the way.

Even in the language-learning industry — arguably the sector hit hardest by AI — companies in the U.S. are still landing investment, and OpenAI itself has directly invested in language-learning software. So the opportunities and the jobs are still out there; they just may not look like what we originally imagined.