Block just cut 4,000 people. Forty percent of its entire workforce. Jack Dorsey posted about it on social media like he was dropping a product update. The reason, per Dorsey himself: AI tools can now do what those people did. Not someday. Now.

So is AI actually killing jobs in 2026, or is everyone just panicking? Honest answer: both, but not equally, and the people getting hit hardest are not who most of the headlines are about.

The macro numbers are actually less scary than the vibes suggest. Snowflake surveyed companies and found that 77% report AI is creating jobs, not eliminating them. Of organizations that saw both gains and losses, 69% came out net positive. That stat is doing a lot of work for optimists like Marc Benioff, who keeps dismissing layoff fears even as his own company cut around 1,000 people this quarter.

Audrey will tell me those Snowflake numbers are funded by a company that profits off AI adoption. Fair point. But the signal is still real: 45,363 tech layoffs since January 1st, and only 9,238 of them, about 20%, are directly tied to AI. The rest are regular old business decisions. AI is not the apocalypse. It is a pressure multiplier.

The Entry-Level Trap Nobody Is Talking About Enough

Here is where it gets genuinely worrying, and where I think the cheerful "transformation, not elimination" crowd is missing the point. Gartner found that 21% of companies have already stopped entry-level hiring because AI handles that work now. By 2027, that number could hit 50%. A third of companies expect to eliminate those roles entirely by the end of this year.

Think about what entry-level jobs actually are. They are how you learn. They are the thing between your resume and your first real paycheck. Customer service, data analytics, IT operations: those are down 37-40% in headcount at AI-forward companies. The junior roles that used to teach you the job are disappearing faster than any senior position.

I am not a 22-year-old trying to break into tech right now, and I am grateful for that. If I were, I would be terrified. Not because AI took my dream job. Because the on-ramp to that dream job just got bulldozed.

The people who keep saying "AI creates more jobs than it destroys" are technically not wrong. But those new jobs in cybersecurity and software development require skills that you usually built by doing the entry-level work first. You cannot skip the tutorial and jump to the final boss. The pipeline is breaking.

What Should Actually Happen

Companies need to stop treating AI adoption as a headcount reduction strategy with good PR. If you are cutting junior roles because AI does those tasks, you have an obligation to build a different on-ramp for early-career workers. Apprenticeships, rotational programs, something. The economic ROI on AI is real, Snowflake clocks it at $1.49 per dollar invested, but if the gains stay at the top and the losses stack at the bottom, we do not have a transformation story. We have an inequality story with better branding.

Policymakers also need to update unemployment systems that were not built for this. More than half of people who qualify for benefits do not claim them because they do not know they are eligible. That number gets uglier if entry-level layoffs keep accelerating through 2026.

AI is not coming for your job the way the scary headlines want you to believe. But if you are three years out of college and your career was supposed to start this year, it absolutely came for yours.