AI Layoffs Will Stop in 6 Months. For an Uncomfortable Reason

Looks like big tech companies got hit with a pretty steep subscription bill this week.

I don’t even want to imagine the panic that must be going on in Silicon Valley right now.

Just go on Twitter and see for yourself: people are way more on edge about AI than usual.

It’s not exactly a secret either: AI has been building itself a really bad reputation lately.

Even the Pope got on board, calling for AI to be “disarmed.”

People got booed at college graduations, like what happened to Gloria Colfilt and Eric Schmidt, Google’s former CEO.

Then there are the massive AI related layoffs:

Coinbase was one of the first to kick 14% of its employees to the curb.

Right behind it came Cloudflare, laying off 1,100.

Meta cut nearly 8,000.

Cisco, around 4,000.

LinkedIn, 875 people.

Intuit, 3,000 employees.

Wix let go of 1,000.

Groupon (didn’t even know they still existed) laid off hundreds.

Along with a bunch of other companies.

And all of this in under 20 days.

I told you this in the last video: all it took was Coinbase showing off its restructuring plan for everyone else to run and copy it.

And look how far this has already gone: tech layoffs this year already top 100,000, and we’re barely into May.

TrueUp is projecting 2026 could close out at 370,000, the worst year since 2023.

I have a theory, tell me what you think: with all the frantic hiring during COVID (when everything went digital overnight and everyone thought that boom would last forever), companies filled up with every kind of engineer imaginable.

But let’s be honest: not every hire turned out to be the next legendary 10x engineer who takes a company to another level.

Because with that much hiring, mediocre engineers always slip through the cracks: they pass the technical interviews, sell themselves as super seniors, and end up doing their job halfway.

Then AI showed up and blew the lid off, exposing everyone.

And look, this isn’t just my theory. The industry itself is already confirming it.

Check out what Uber’s CEO, Dara, said in an interview for The Diary of a CEO:

So let’s be objective about this: if out of that 90% using AI, only 30% are exceptionally productive, why the hell would you keep the other 60% on payroll?

You fire them, and all that money you used to spend on salaries, benefits, bonuses, equity, whatever, you hand it over as an open bar of tokens to your top performing engineers.

Oh, and by the way: there are already startups, like Tenex, paying their best engineers over a million dollars a year, and not by the hour, but for what they produce with AI leverage.

They’re going to be the superstars of the tech industry.

So I guess we’re stuck living, breathing, and swallowing tokens, hoping we make it onto that list of chosen ones for the technological shift.

But apparently these structural changes, and the wild spending on AI models, aren’t exactly working out great for these companies.

Check out what Uber’s COO, Andrew Macdonald, said on the Rapid Response podcast:

Sure, that 25% doesn’t mean much if that code doesn’t add any real value to the product in the end.

That’s the whole point of coding, right? To actually make people’s lives better somehow.

Also, in that same interview, Andrew brought up when their CTO, Pravin, went viral for saying they’d already burned through their entire 2026 AI budget, and he said that back around April.

This also reminds me of a tweet I saw yesterday while writing this script, where a consultant claims a company spent 500 million dollars in a single month on Claude because nobody thought to set a usage limit.

Genuinely insane.

It’s like tokens are the popcorn you eat at the movies.

A handful falls on the floor and you don’t even look.

The bucket’s still full anyway.

Except here, every kernel that falls costs millions.

And don’t even get me started on Sankar’s tweet, showing that 82% of every dollar spent on AI never even makes it to production: it goes into fixing bugs, rework, and reviews.

All of this is a signal to executives at big tech companies that their plan to replace employees with AI has two problems.

The first: apparently token costs are already surpassing what it used to cost to keep the employees they let go.

The second: when your tokens run out, AI just stops.

It won’t put in two extra hours because something needs to ship to production.

If you don’t pay, there are no tokens.

You fired humans to save money, and bought a subscription that bills you straight into bankruptcy.

You handed years of work over to a machine that has no loyalty, no discretion, and nothing to lose if it makes a mistake.

The employees you let go knew what to do when things broke.

And here’s my message for you, still stuck wondering if you’re about to lose your job.

All of this happening right now is showing us one thing.

AI came to blow the lid off, to separate the ones who were hiding from the ones who actually know how to do their job.

Whoever knows how to use AI like an extension of their own body is the one who’s going to make serious money.

So the question shouldn’t be: is AI going to take my job?

Your question should be: am I going to be able to become part of that 30%?

Because the other 70% already knows what’s coming.

Thanks for watching.