In April, Meta began installing software on employee laptops that captures every keystroke, mouse movement, and click. The program is called the Model Capability Initiative. It takes periodic screenshots. It runs across hundreds of apps and websites. There is no opt-out.

The stated purpose is to train AI agents to perform the tasks that the employees themselves are performing. The company framed it as a way for workers to help improve models that struggle with basic computer behaviors like navigating dropdown menus and using keyboard shortcuts. Just do your daily work, the memo said. We'll take it from there.

The same week, Meta laid off 8,000 people.

I used to work there. Not when the keyloggers went in, but during an earlier era, one that was already deeply interested in watching you work. They never called it surveillance. They called it performance data. Output metrics. Cycle time. The tracking was granular, obsessively so, despite the fact that the data literacy behind it was poor. Nobody quite knew what the numbers meant, but everybody knew they were being measured.

What I remember most clearly is that it only went one direction. We didn't get reports on our managers' performance. We didn't see their activity data. We didn't know how our output metrics were being interpreted or by whom. It was an asymmetric relationship dressed up in the language of transparency. And if you're reading this from inside any large company right now, I'd bet that dynamic sounds familiar.

Meta is just the one that got caught being explicit about it. But every large company has its own version. They just bury it deeper. In the employee handbook. In the acceptable use policy. In the onboarding paperwork you signed on your first day without reading because you were too relieved to have the job. The terms of your employment have always been one-sided. What's changed is how much the company is willing to take, and how little they feel they need to offer in return.

Here's what the inside of corporate work looks like in 2026, by the numbers.

61% of American workers are languishing at work, struggling with engagement, motivation, or fulfillment. 83% of knowledge workers report some degree of burnout. Entry-level employees are the hardest hit, with associates and junior staff reporting the highest rates of reduced engagement. And six in ten workers are now doing something researchers are calling "job hugging," which means staying in a role not because the work is good but because the alternative feels worse.

That last one deserves a second read. More than half of the people sitting in meetings this week are not there because they want to be. They're there because leaving looks scarier than staying. The system doesn't retain people by locking the door. It retains them by making the hallway dark. These are people living under terms they never agreed to, in jobs that keep updating the deal without asking.

Meanwhile, the process of getting into the building has become its own kind of degradation. Think of it as the sign-up flow. Nearly half of UK job seekers have now been interviewed by AI, and they describe the experience as awkward and humiliating. 53% of job seekers experienced ghosting within the last year. Entry-level jobs routinely require three or more years of experience. When candidates don't hear back, they spray applications everywhere. When recruiters get flooded, they add more automation to filter. When the automation filters out good people, candidates learn to game the system. Both sides blame each other. The process gets worse.

I spent eight years in recruiting. I sat on the employer side of this. I watched the tools get more sophisticated and the experience get more inhuman in lockstep. The technology was never the problem. The problem was that nobody in the room believed the candidate's time was as valuable as the company's. That asymmetry again.

People like to say the pendulum swings. Employers have the power for a while, then the market shifts, candidates get leverage, and things balance out. I believed that for most of my career. I don't anymore. Here's why.

In the early and mid 2000s, the companies we now think of as monoliths were nimble. Meta, Amazon, Salesforce, Block, Snap, they were the disruptors. They moved fast. They were small enough to feel the consequences of a bad hire and smart enough to move quickly on a good one. Then they grew. And as they grew, they did something that had less to do with building great teams and more to do with competitive strategy: they hired en masse. Not because they needed all of those people, but because every engineer on their payroll was an engineer their competitor couldn't have. It was a brain drain play dressed up as hypergrowth.

That era is over. The leaders of those companies can see what's coming. Smaller, more agile companies, powered by tools that didn't exist five years ago, are going to outcompete them. That's why the cuts feel existential. Look at what happened in a single quarter: Meta guided $115 to $135 billion in AI capital expenditure while cutting 10% of its workforce. Amazon cut 16,000 jobs while spending $43 billion on infrastructure in Q1 alone. Block, the parent of Square and Cash App, cut half its entire workforce. Snap cut 16% of its people. Microsoft offered buyouts to 7% of staff. These aren't trims. They're transformations. They're trading headcount for compute because they've done the math and the math says the future needs fewer people.

When a company records your keystrokes to train an AI agent and then lays you off in the same week, that's not a cycle. That's a ratchet. It only turns one direction. The terms get updated. You get notified, if you're lucky. And the "agree" button is the only one that works. The question isn't whether the pendulum will swing back. The question is whether the pendulum still exists.

If you zoom out, the picture starts to look less like collapse and more like dispersal. Some people read the terms. And canceled.

In Victoria, British Columbia, a stock photography platform called Stocksy United is owned by its photographers. Not in a vague, metaphorical sense. Nearly 1,000 contributing artists are shareholders with voting rights. Stocksy pays 50% royalties on standard licenses and 75% on extended, numbers that make the industry standard at Getty Images and Shutterstock look extractive by comparison. Revenue has topped $10 million. The founders came from iStockphoto, watched it get sold to Getty, watched the royalties get scraped back and the quality decline, and decided to build a model where the people doing the work couldn't be sold out from under themselves.

In the UK, a network called CoTech links over two dozen worker-owned technology cooperatives, firms doing web development, data tools, design, and consulting, all owned by the people writing the code. Outlandish, one of its members, has about 20 worker-owners building applications for charities and universities. CoLab, based in Ithaca, New York, is a full-service tech cooperative with a collaborator network spanning the globe. These aren't communes. They're functional businesses where developers and designers decided they'd rather own their shop than rent themselves out to someone else's.

And in Denver, an 800-person taxi cooperative called Green Taxi Co-op controls a third of the metro ride market. They launched in 2015, back when everyone assumed Uber was inevitable.

These aren't startups in the Silicon Valley sense. They're not trying to scale to a billion-dollar valuation. They're trying to build something that works for the people inside it. And they're working.

On a different axis entirely, the solopreneur economy is accelerating in ways that would have seemed absurd a decade ago. Solo-founded startups went from 23.7% of new ventures in 2019 to 36.3% by mid-2025. A full tech stack for a one-person business now costs $3,000 to $12,000 a year, a 95 to 98% reduction compared to hiring equivalent staff. Operating margins for these businesses run 60 to 80%.

Zoom just launched a "Solopreneur 50" program on May 4th, awarding grants to one-person businesses powered by AI. They got nearly 3,000 applicants from 48 states. The biggest category wasn't technology. It was services and consulting. People monetizing what they already know.

Something is happening here, and it isn't a trend piece about the gig economy. The cooperatives are people who read the terms, rejected them, and wrote their own. The solopreneurs are people who decided they'd rather be the platform than the user. Both groups are proving that the basic economics of work have changed. The infrastructure to do meaningful work outside of a corporation now costs less than your annual health insurance premium.

But here's the honest gap. And I think it matters more than the success stories.

Everything I just described still requires you to be entrepreneurial in the traditional sense. You need to find your own clients. Build your own audience. Set up your own payment infrastructure. Handle your own taxes, contracts, liability. The cooperatives solve the governance problem but not the discovery problem. The freelance platforms solve discovery but commoditize you into a bidding war. The solopreneur stack solves execution but assumes you want to become a one-person media company first.

What doesn't exist yet is the thing that would actually make the future of work accessible to normal people. A way to say "here's what I'm good at" and have that connect directly to someone who needs it. Without a 20% platform cut. Without submitting to an algorithm that ranks you against ten thousand other profiles. Without having to become a content creator just to get hired.

The tools for building have gotten radically cheaper. The tools for finding, for matching, for creating genuine economic relationships between a person with a skill and a person with a need, those are still stuck in the Upwork era. The exit from the corporate model exists. The bridge to what comes after it hasn't been built yet.

Every few months, your phone asks you to accept updated terms of service. You don't read them. Nobody reads them. You scroll to the bottom, tap "agree," and move on with your day. The company gets a little more access. You lose a little more control. And because it happens gradually, it never feels like a decision. It just feels like the way things work.

Employment has become the same agreement.

The terms keep changing. The surveillance expands. The stability shrinks. The buyout packages arrive. The AI gets trained on your output. And every morning you show up and tap "agree" again, not because the terms are good, but because the alternative is logging off entirely, and nobody's built a platform worth switching to yet.

But some people are building. The cooperatives wrote their own terms. The solopreneurs decided they didn't need a platform at all. The infrastructure isn't finished, and the bridge between "I have a skill" and "someone will pay me directly for it" is still under construction. But the foundation is real, and more people walk across it every day.

I'm not going to tell you to cancel. I'm going to tell you to read the terms. All of them. Notice what you've agreed to. Notice what's changed since you signed up. Notice who the agreement is designed to protect.

And then decide if you want to keep clicking "agree."

Thanks for reading,

Tyler

If this hit, forward it to someone who needs to read it. If you're not subscribed, you can fix that HERE.

Sources and further reading: