The Unoptimised Human — When Competence Becomes the Ceiling

Something shifted in July. Not a single event, but a pattern that became impossible to ignore: we’ve spent the last two years optimising everything with AI, and we’re starting to see what we’ve lost in the process.

The word I keep coming back to is averaging. AI doesn’t create. It synthesises the most statistically likely output from the sum of human expression. Feed it enough brand guidelines, enough competitor copy, enough “best practice” templates, and you get something polished, professional, and utterly indistinguishable from everyone else who did the same thing. I called it “Competent Beige” in one of this month’s pieces, and the response told me I’d struck a nerve. Because everyone recognises it. Everyone’s seen it. Most people just hadn’t named it yet.

The deeper problem isn’t that AI produces mediocre content. It’s that competence is seductive. When the output reads well and looks right, there’s no friction to signal that something’s wrong. The dashboard is green. The engagement metrics are fine. The email open rates are stable. And underneath all of that, the brand is slowly dissolving into the same voice as every other brand using the same tools.

This month I explored that tension from multiple angles: vibe marketing, the confirmation economy, emotionally intelligent AI, and what happens when your bank replaces you with an algorithm. The thread connecting all of them is the same question: in a world where AI can make everything competent, what’s your actual competitive advantage?

Here’s what I wrote about it.

The Pieces

“Competent Beige”: When AI Optimises the Soul Out of Your Brand

This was the piece that started the month’s thread. I noticed three LinkedIn posts from three different brands in completely different sectors that could have been written by the same person. Same rhythm. Same safe, balanced tone. Same vaguely inspirational sign-off. That’s not a coincidence; it’s a structural consequence of using AI as a voice rather than a tool. The post names the problem and proposes a simple test: if you removed your logo from your AI-generated content, could anyone tell it was yours?

The Vibe Marketing Paradox: Technical Competence Without Brand Intelligence

The vibe coding conversation has been one of the more interesting cultural moments in tech this year. Someone with no engineering background prompts an AI to build functional software, and it works. But the same question applies to marketing, and almost nobody is asking it. Thousands of brands are doing the marketing equivalent of vibe coding: prompt AI, get polished output, ship it. The output is technically competent. It’s also strategically empty. This piece explores why “good enough” content is actually a competitive disadvantage when everyone has access to the same tools.

The Confirmation Economy: How Modern Branding Manufactures Reality

In 2023, a Belgian sommelier bought a bottle of wine for €2.50 and convinced a panel of professional wine experts it was exceptional. He didn’t change the wine. He changed the context. The label, the story, the social proof, the environment. Every layer of reality was constructed, and the experts confirmed what they’d been primed to see. This piece examines how modern branding operates on the same principle, and why AI is accelerating the gap between perceived value and actual value.

Attention Architecture: Emotionally Intelligent AI

This one went further than the others. Instead of just critiquing AI’s averaging tendency, I proposed an alternative: what if AI systems were architecturally designed to process emotional context, not just semantic content? Drawing on the original “Attention Is All You Need” transformer paper, I sketched what an emotionally intelligent attention mechanism might look like. The argument is that the next frontier isn’t bigger models or faster inference; it’s systems that understand why something matters, not just what it says.

What I’m Watching

The EU AI Act deadline hit on August 2nd. The regulatory framework that’s been theoretical for two years is now enforceable. The implications for marketing teams using AI-generated content are significant: transparency requirements, risk classifications, and accountability structures that didn’t exist a month ago. If your team is using AI to generate customer-facing content and you haven’t audited your compliance position, now is the time. I wrote about this ahead of the deadline, but the real story will unfold over the next quarter as enforcement begins.

The “vibe coding to vibe everything” pipeline is accelerating. What started with software development is spreading to design, marketing, legal, and finance. The pattern is always the same: AI makes the technically competent part easy, and organisations mistake that competence for strategic value. The companies that figure out the difference first will have a meaningful advantage.

The Question

Here’s what I’ve been sitting with all month: If AI can make everything competent, what’s left that only you can do?

Not what you should do. Not what’s “authentic” in the brand guidelines sense. What is the thing that, if you removed it, the output would be technically fine but fundamentally wrong?

I think the answer has something to do with taste, conviction, and the willingness to be distinctive at the cost of being universally palatable. But I’m still working on it. I’d genuinely like to hear yours.

David Chadderton is the creator of the STAR Framework and the author of three books, most recently Dear Algorithm, It’s Not Me, It’s You. He writes The Unoptimised Human on Substack, where he explores the intersection of human behaviour, AI, and the psychology of decision-making. By day, he’s a Chief Marketing Officer. By night, he’s the person who can’t stop asking why people do what they do.