Cool, but who asked for this?

Cool, but who asked for this?

The 2026 hiring data, and the one skill AI still can't fake

The 2026 hiring data, and the one skill AI still can't fake

The gist…

The gist…

A couple years ago, LinkedIn's self-appointed prophets swore AI would be writing all our code by now, no devs required. Weird, then, that software engineers now make up a bigger slice of Big Tech hiring than they did back in 2019, not a smaller one.

Here's the part nobody's musing about: that slice got bigger because the whole pie shrank. Overall hiring at the world's biggest tech companies is down 25% since 2019. Design's down 48%, and new grad hiring is down 65%. AI didn't take your job. It kicked out the bottom rung of the ladder you were supposed to climb to get one.

I talked to three engineers who live this from the inside — Ingrid Zhao (senior engineer at Apple), Saeed Salehi (a 14-year technical lead at RTL), and Fatima Taj (senior engineer at Yelp) — about exactly where AI's code looks flawless and still gets rejected, and what that means for how safe your job really is.

What we get into:

  • Why the hiring "recovery" for engineers is really a story about who got pushed off the ladder entirely

  • The exact test Saeed uses to catch code that compiles perfectly and solves nothing ("what happens when this fails?")

  • Why Klarna quietly rehired the human agents it bragged about replacing with AI

  • The one shift that separates engineers stuck at senior from the ones promoted to staff

Transcript

HEY

I’m Tom, and I run Cold. I help startup founders and marketers make progress.

If you want to go deeper, see how I work, and explore ways we can collaborate, head here: