We're All Running Each Other Through ZeroGPT

A new hobby has quietly swept through offices, group chats, and comment sections, and nobody will admit to it.

People are copying each other's writing into AI detectors.

The LinkedIn post that's a little too polished. The executive email that suddenly has paragraph structure. The newsletter from the guy who used to write in all lowercase and now opens with "In today's rapidly evolving landscape." Paste it into ZeroGPT, watch the little meter, and render your verdict.

We have all become amateur AI detectives. None of us asked for the badge.

The DIY moment, again

It feels a bit like when HGTV hit the airwaves and everyone became a contractor. Suddenly your neighbor had Opinions about subway tile and a strong stance on open-concept living. They'd never hung a cabinet, but they could talk.

Same thing now, except the renovation is happening to our writing.

Everyone is suddenly verbose. Everyone is eloquent. Everyone's cold email reads like it went to grad school. And the em dashes — oh, the em dashes — are everywhere, multiplying like rabbits, punctuating sentences that never needed punctuating.

(Yes, I know. Look at me. We'll get to that.)

How the arms race actually works

The cycle is beautiful in its stupidity.

Someone uses AI to write something. Someone else suspects AI was used. They run it through a detector to confirm their suspicions. Word gets around. So the original writer starts using AI to sound less like AI — please make this warmer, more human, less like a press release.

And everyone involved pretends none of this is happening.

We've accidentally built a world where AI writes the content and AI evaluates whether AI wrote the content, while humans referee from the sidelines, squinting at em dashes like sommeliers.

The point here isn't whether the detectors work. That's the boring question, and the answer is "sort of, not really, who cares."

The interesting question is the one nobody's asking out loud: why do we suddenly care so much?

"This sounds like ChatGPT" is a book review now

Here's what's actually going on. "This sounds like ChatGPT" has become a cultural critique, the new "this is giving corporate," a way to say I don't trust this without explaining why.

And the real giveaway is almost never the writing itself. It's the change. When someone who texts in fragments suddenly emails you three balanced paragraphs with a tidy call to action, the alarm doesn't go off because the writing is bad. It goes off because it isn't theirs.

We're not detecting AI. We're detecting a personality transplant.

Now walk this into a trades business

Contractors are about to inherit this entire debate, and most of them don't see it coming.

For twenty years, homeowners have complained about the same things: slow follow-up, calls that never came, the estimate that was "in the mail." Communication was always the gap.

Now the contractor who hated email can sound responsive, organized, and thoughtful — overnight, with a tool that costs less than a tank of diesel.

And some homeowner, somewhere, is going to read that tidy follow-up and squint: Did AI write that?

Maybe. Probably, in part.

But here's what the squinting misses. Nobody hires a plumber because he writes a beautiful email. They hire him because something is leaking and the water needs to stop.

AI can't stop the water. It can't crawl under the house, read the panel, or size the unit. It can't do the work.

What it can do is help a good contractor sound like the professional he already is — instead of losing the job to someone less skilled who simply answered faster.

That's not a trick. That's the gap finally closing.

The thing we're actually measuring

For years, we judged communication by one standard: was it helpful?

Now we're sliding toward a different standard: do we think a robot was involved?

That's a worse standard. A useful estimate email is useful whether the contractor typed it at 6 a.m. or roughed it out with help over coffee. The homeowner with a flooded basement does not want artisanal, hand-forged prose. They want to know when you're coming.

The real risk was never that contractors would use AI. The real risk is sounding like every other contractor who uses AI — the same chirpy template, the same three em dashes, the same "We're thrilled to serve our valued customers!" that fools no one and warms nothing.

So here's the move, and it's the whole game:

The future won't belong to the people who never touch AI. It won't belong to the people who let AI do all the talking, either.

It'll belong to the ones who use it to sound more like themselves. Not less.

Claude helped me write this. On ZeroGPT, it passes as ~94% human.