#61 | I don’t optimize, therefore I am!

TL;DR: Machines can fake moments better than we live them. But do we really care whether an Instagram or TikTok post is real? And what actually proves you’re human when every moment can be generated?

👋 Happy New Year (still),

Oops! I did again. But you probably do this too. We all paste our words into an AI now and ask it to make us sound more like ourselves.

I recently wrote about identity and performance. It was certainly true, but felt flat on the page.

The AI version came back smoother, more conversational, the kind of line that (hopefully) gets shared because it sounds like someone (me) just understood something real (whatever that means, though).

Interestingly, when I read it, I couldn’t tell if it was better because it was actually closer to how I think, or just closer to how I imagine I sound when I’m thinking well.

That gap—between what we actually are and what sounds like us—that’s what today’s edition is about.

We’re all watching the difference between ourselves and a version of ourselves that’s somehow more “us” than we are.

And the question beneath that gap is: if a machine can nail the surface of authenticity, what’s left to actually prove we’re thinking, changing, human people underneath?

When any moment can sound/look authentic, does authenticity still matter?

For years, the internet ran on one unstated rule: if something sounds like it came from a real person, it probably did.

That’s gone now.

AI models can write diary entries about loneliness that feel lived-in. Apologies may sound genuinely ashamed. And confessions may land like someone just told you something true.

They’re not trying to trick anyone. They’re just good enough that the difference doesn’t matter.

What I find interesting is that we don’t actually care as much as we think we do.

Research on AI-authored emotional messaging indeed shows people prefer synthetic content when it feels true to their experience.

We stream AI music without hesitation, share generated art, and we like captions that sound authentic. Even when—especially when—we suspect they might not be human.

The moment something works emotionally, the question of whether it’s real becomes almost optional. And most of us choose not to ask.

Which means a single post—no matter how raw, vulnerable, or specific—can’t function as evidence that a human wrote it anymore. Can it?

But we’re also not sure we need it.

What happens when ‘rawness’ becomes just another aesthetic

Instagram spent a decade teaching us: polish is proof.

The grid, the filters, the composed moments. Authenticity meant looking like you had it together. That was the test, and we all learned how to pass it.

Then the tools got too good. Perfect lighting, perfect composition, and perfect versions of you became trivial to generate.

So perfect that they started to feel fake.

No wonder the culture flipped almost overnight.

​Instagram’s head recently argued​ that the polished grid is dead, that raw content is now the only proof that something is real.

Unflattering photos, messy rooms, BeReal’s supposedly random pings. The assignment switched from “look polished” to “look like you’re not trying.”

In my opinion, however, we didn’t stop performing. We just changed what performing looks like.

I’ve retaken “candid” photos five times to get the right kind of bad (well, I forgot to turn off the photo lights in the first four pics).

Friends time their BeReal posts to coincide with something interesting happening. Photo dumps are carefully arranged to look chaotic.

We learned the choreography of looking like we’re not choreographing anything.

And, weirdly, we seem to be okay with this.

​Studies on social media authenticity​ show that what matters isn’t whether something is real — it’s whether it feels relatable. If it reflects something true about the viewer’s experience, it certainly connects.

Neither polish nor rawness proves anything. But both work, as long as they feel true.

So what actually proves you’re human?

The only answer that holds up is this:

Proof you’re human isn’t in one moment. It’s in the pattern you trace over the years.

Not a brand pattern. The actual, crooked shape of a life lived in public.

Old posts that now embarrass you. Opinions you’ve reversed, with both versions still sitting there.

Phases where you cared about something for three months, then dropped it completely.

Half-formed ideas you posted anyway. Moments where you contradicted yourself week to week, all visible, unretouched.

A model can generate an infinite number of thoughtful captions about growth. But it cannot have actually grown (let’s see if that remains true).

It also cannot look back at what it said last year and feel the weight of being wrong. Neither can it have the before and after that time gives you.

You do. That’s the gap that matters.

And that’s the gap the machines haven’t closed yet. And maybe that’s the only gap that matters.

What your archive actually proves

If I look back at my own trail, I see versions of myself I no longer claim.

Phases where I sounded certain about things I’m now uncertain about.

Stretches where I was struggling and showed it. Ideas I built and then abandoned. All still there.

And yes: I’m choosing which contradictions to highlight. Even my proof is edited. But that’s the entire point.

The fact that there’s a trail at all—messy, uneven, not optimized for engagement—is what reads as human.

We’re stuck with our own contradictions. And maybe that’s the real thing about us.

The cost of proof

But don’t lean back. This doesn’t give you a new way to prove you’re real.

In fact, it asks for something harder: visibility of contradiction.

In other words, it is the willingness to let the old versions of you stay public. Permission to change without cleaning up the evidence.

In a world built on curating your best self, that’s almost like surrender, isn’t it?

But if the question is what truly reads as a thinking person rather than a system, it’s not that one perfect post an AI refined for you.

It’s the crooked line. The contradictions you keep visible. The choice to be different in public without optimizing it.

Time will tell.

Cheers,

Mark
The AI Learning Guy
👋⚡😎

Interesting Sources

• ​Mosseri essay — Instagram head Adam Mosseri on AI, the “death” of the curated grid, and raw content as proof of reality
• AI-authorship study — Research on how AI‑generated emotional messaging affects authenticity, moral disgust, and consumer responses
• Social authenticity — Longitudinal study on feeling authentic on social media and its impact on mental health
• BeReal design — Analysis of how BeReal’s app mechanics structure “authentic” self‑presentation
• Deepfake trust — Study on how AI‑generated deepfake videos shape public trust and democratic perceptions

Note: No single website has all the answers. This list serves as a starting point for those who want to explore or satisfy their curiosity about AI.
Links: Links with * are affiliate links. See disclosure below.