Stay real

Read time: 5 mins

Hi Reader,

So many of us are beginning to sound like ChatGPT.

Not in the obvious ways. Not—the—em—dash—discourse. Not one particular word.

I mean the shape of the writing: the tidy hook, the clean list, the neat ending that makes you nod… and then forget it existed.

This matters because attention is human. And humans have a radar for effort. When we sense there’s no person behind the words, we don’t engage. We stop trusting it.

“We expect more from technology and less from each other.” — Sherry Turkle (Sociologist of technology)

A friend asked me to help refine a LinkedIn post about recent work. Normal request. I gave him the kind of feedback I’d want someone to give me: what’s the point, why does it matter, what do you actually believe?

He came back with a new draft that was smoother and at the same time emptier.

It had the polish of a TED talk transcript about a life he hadn’t really lived. The “message” was there, but the decision wasn’t. The bit where you can feel a person putting their name behind a sentence.

Then it clicked. I’m not interested in collaborating with a loop. If you’re not willing to do the minimal human work of articulating why something matters, it doesn’t deserve minimal human attention.

A few days later, the same feeling showed up again, in a different form.

I was in a meeting, and someone said, casually, “Oh, we’ll just ask ChatGPT.” Ah, come on, man! A lot of my work I do alone. Working with real people on projects is precious to me. If we’re going to outsource the thinking, what are we doing here?

We all do this. Me too. The difference is that some of us still want the tool to be a support, while others are starting to treat it like the author.

You can feel it in the structure. It hints at a ‘hard season’ with zero specifics, then pivots into a universal insight that could’ve been written by anyone. The writing is built to be frictionless: a hook engineered to hold you, a list designed to simplify, a closing sentence designed to sound wise. It’s competent. It’s smooth. It’s also as forgettable as elevator music.

Once you see that pattern, you start noticing it everywhere: personal brands, “thought leadership,” comment sections filled with suspiciously perfect encouragement from accounts that clearly didn’t sit with the post for even ten seconds.

This is just one of the ways AI is making our expression more generic. It’s trained to produce the most acceptable, average version of an answer. When millions of people rely on the same model to draft public writing, the average starts to dominate the culture. Then the culture feeds the next drafts. The loop tightens.

If you want a quick test, read the captions under the Instagram posts of your friend's business.

We already understand this with feeds. Social platforms show you what keeps you watching. Over time, the loudest and most extreme version of a topic becomes the most visible version, and people begin mistaking visibility for truth.

Chatbots are a different kind of filter. They can only answer from what they’ve been fed and what they’re allowed to say. That means the “reasonable” answer is always constrained: by data, by incentives, by safety boundaries, by whatever the system has learned to avoid, soften, or steer around.

Spend enough time inside that style of language, and it leaks back into you. Into your phrasing. (The number of people I hear say, "It's not X, it's Y." Did we talk like that 3 years ago?) Into your confidence. Into the kind of conclusions you reach quickly. Into what you start repeating before you’ve actually thought it through.

So here’s the anchor I’m trying to keep, just as a way of staying myself:

Ask “why” more than you ask for output.

Why am I seeing this? Why does this feel true? Why am I about to post this? Why did I phrase it like that?

I caught myself doing it yesterday. I was halfway through rewriting some pages on my website for the 12th time in less than 12 months, and I could feel the “perfect wrap-up sentence” arriving. The kind that sounds wise. I deleted it. It wasn't wrong, but it wasn’t mine yet. I hadn’t earned it.

That’s the work.

Escape the algorithm

VIDEO: Ben Affleck (with Matt Damon) on why AI won’t replace filmmakers. His take: models like ChatGPT/Claude/Gemini “go to the mean” (so the writing is “really shitty” and “not reliable”); they’re useful as support (examples, logistics, cost-saving), but “always” rely on “human artistic aspects”, and a lot of the “AI will change everything” talk is about justifying valuations and data-center spend.

​

CARTOON: “Yes, the planet got destroyed. But for a beautiful moment in time, we created a lot of value for shareholders.” Tom Toro, New Yorker

If you got this far,

you're part of 71% of subscribers, so thank you, Reader.

If a section has particularly interested you, click the link of the corresponding section. It helps me prepare future newsletters.

If this was forwarded to you, consider subscribing.

See you soon.

​Peace,

Has

​​​