When Everyone's Content is AI-Polished, What Signal is Left?
As AI makes flawless prose and pitch-perfect résumés the default, one founder argues that polish is no longer a signal — and that the messy, contradictory, unfiltered details are what's left to trust.
I have started to distrust writing that is too polished.
That is a strange admission from someone who uses AI every day. At Clera, we use it to build products, research companies, organise information, and turn chaotic notes into something other people can actually understand. It saves us an enormous amount of time.
But I also spend much of my week reading candidate profiles, interview notes, outreach messages and posts from people trying to explain what they are good at.
More of it is flawless than ever. But less of it sticks with me.
You can give an AI tool three bullet points and get back a confident opening, a neat argument and a quotable final line. I do it too. Once everyone can produce that surface, the surface stops telling us very much.
LinkedIn recently announced it would reduce the distribution of what people have started calling “AI slop.” The company describes it as low-effort content that appears polished but lacks any real perspective or substance.
I understand the reaction. My feed is full of posts that are perfectly structured and completely forgettable.
The bigger change goes far beyond LinkedIn.
Polish Used to Tell Us Something
Clear writing has always been an imperfect signal. People have used editors, ghostwriters and résumé consultants for a long time.
Still, producing something thoughtful used to require a certain amount of effort. You had to decide what you believed, find the right examples, and work out how to explain them.
AI has dramatically reduced that effort. Someone can now have a half-formed opinion and present it like a conclusion they have defended for years.
Because of that, I find myself looking for different signals.
I want the oddly specific detail. The decision that was uncomfortable. The moment someone changed their mind. The observation another person might disagree with.
A recent Forbes article makes this point through stories told by Jeff Bezos and Jensen Huang. Bezos talks about working on his grandfather’s ranch. Huang talks about washing dishes at Denny’s.
The lessons they took from those experiences are fairly ordinary: work hard, stay humble, and take pride in difficult jobs. What people remember are the ranch and the dishes. Those details show us where the belief came from.
Hiring Makes This Impossible to Ignore
Building Clera has made me notice the same change in hiring.
Hiring looks wonderfully clean in a spreadsheet. A candidate wants a particular salary, title, company stage and working model. A company has a list of skills and years of experience it expects. Match the columns and introduce the people.
Real life is much messier.
Candidates regularly contradict their stated preferences — and I mean that in the best possible way.
Someone tells us exactly what they want. Then we show them a role that breaks several of their rules, and they want to speak with the founder immediately.
Perhaps the problem is unusually interesting. Perhaps they would own something important. Perhaps the founder explains the company in a way that makes them reconsider what they were optimising for.
None of that fits neatly into a résumé or a set of filters.
Hiring managers behave the same way. They write a precise list of requirements, reject candidates who satisfy every one of them, then become excited about someone with a less obvious background who understands the problem immediately.
This can be maddening when you are trying to build software around hiring. It is also what makes the problem so interesting.
I have come to believe a thoughtful introduction has value because it carries a small amount of personal risk.
When someone says, “You two should talk,” they are putting their judgment behind the conversation. They have seen something that may not be obvious on paper, and they are willing to explain why it matters.
AI can help us find people, remember details and notice patterns across far more conversations than any human could manage alone. That is a large part of what we are building at Clera.
The goal is to enter the conversation with better context — not to pretend it has become unnecessary.
What I Look for Now
When I interview someone, I care less about whether every answer sounds complete.
I want to know why they made a particular choice. What nearly changed their mind? What did they get wrong? What surprised them once they were actually doing the work?
Real answers often have loose edges. People pause. They correct themselves. They remember an important detail halfway through.
That gives me far more information than a perfectly delivered summary of their résumé.
The same applies to leaders writing online. I do not need another polished post telling me that culture matters or that resilience is important.
Tell me about the person you nearly didn’t hire. Tell me about a company policy you reversed after seeing how it worked in practice. Tell me about the moment one of your principles became inconvenient.
Then I can understand what you actually believe.
I used AI while working on this article. Of course I did — a ton.
It helped me test the structure and clean up sentences. Hiding that would feel performative. The distinction I care about is whether the tool helped me express an observation or supplied something that merely sounds like one.
If I cannot explain what I have personally seen, decided, or learned, I probably do not have much worth publishing.
AI helped make this article clearer. It did not build Clera, watch candidates repeatedly surprise us, or change its mind about what makes a strong match.
Those conclusions, including the ones I may eventually get wrong, are mine.
