I wrote my first post on LinkedIn using AI, and it failed miserably.
I gave it the best examples. I followed rules I took from a 'LinkedIn guru'. I was happy if I got 1 impression per minute, and I refreshed endlessly to watch the number go up.
So I stopped trying to copy what worked for others. I started writing the posts myself.
I realized they don't need to be perfect. They don't need to follow a structure. They just need a good hook and a personal touch.
People don't follow you for "perfect content", they follow you for your journey. The only person that can share your journey is you. That's what makes you special.
Now anything I write hits thousands of impressions or more. I'm not optimizing for impressions, but it's nice to have for three months of consistent posting.
That experience taught me what AI is actually for. And it's not what most people think.
You can feel the slop before you can name it.
When your coworker sends you a deliverable, and you know they haven't internalized what they wrote.
When your classmate sends you "their part" of the work, and it looks like something you can generate in five minutes.
When you open a deck for a journaling startup, and it's called "Memora", again.
AI uses beats. It has the same rhythm, the same sentence length, the same confidence. It will always produce something that is 80% good, the mean of everything "good" ever written. Just enough not to say no to, but nothing special.
The world doesn't need a thousand 80% pieces of work, it needs ten that are 100%.
The reader can tell, not because of what was written, but because of how structured it feels. Even chaos can have structure.
Here's the part most people miss.
The producer of the slop saves an hour. Every reviewer downstream loses ten minutes.
Multiply across the economy. AI generated decks reviewed by AI summarizers. AI generated cold emails filtered by AI agents. Hiring decisions made on AI-generated applications, screened by AI-generated rubrics. We are speedrunning a future where AI generates the content for AI to consume, and humans pretend to be in the loop.
Producers save time and feel productive. The result is a system bloated with slop. We've built a machine that punishes the careful and rewards the careless.
AI isn't the problem. Treating it as a writer is.
AI should be a challenger. Your job as a human is to take the feedback and choose what to keep.
This shift is what makes the difference between AI that makes you lazy and AI that makes you sharper.
Here's how I actually use AI without producing slop.
Rule 1: Output judged by what works vs. output judged by who I am.
I let AI write my code. It's a means to an end, the product of a vision.
I never let AI write my emails, my posts or my decks. These are reflections of who I am. They need to tell the story the way I understand it, not how we want people to take it.
Humans are innovators that think outside the box. We find creative solutions to targeted problems.
AI can tell you founders need money to scale. A human sees that an investor's hesitation isn't about the numbers, it's about whether you can defend the numbers when challenged. That's the kind of thing only a person with skin in the game ever notices, and it's the kind of thing that decides whether the round closes.
Don't let AI choose what you want to build. Let AI do the building once you have personal conviction for it.
Rule 2: AI can't make a mistake on purpose.
Fundamentally, AI builds output by what is probabilistically the most likely next word. It will be right 80% of the time.
Real voice is irregular, unexpected, and personal. The intentional breaks in pattern. The metaphors and allusions. Personal references with fragmented context. These are probabilistically unlikely.
You can't have conviction behind a script. You have conviction because it is personal.
Rule 3: AI shields you from being wrong, which means your gut never trains.
You get better at decisions by being wrong and noticing.
Human beings have a sixth sense. The gut feeling that tells you if something is right or wrong is trained on centuries of evolution.
When you make a decision and you are wrong, you internalize the mistake. People who outsource their thinking to AI are realizing that they have no conviction. They need external validation.
After years of getting answers from a system that agrees with them, they've lost the instinct that says something is off.
I found myself asking if coming to Dubai to build was the right choice given the current situation. AI laid out all the pros and cons. I saw the whole picture, but I couldn't establish conviction on yes or no. However, I didn't feel anything when I read the risks and I knew from that alone that my gut is saying: go to Dubai. So far, so good.
Lived experience makes decisions that don't make sense on paper but are 100% correct. AI can't do that. Not because it's stupid, but because it has no skin in the game.
It's impossible to have an intellectual conversation with someone who doesn't have an opinion and a personal attachment to what they say.
In a world flooded with AI output, the things AI can't do become valuable.
Taste. Opinion. Lived experience. The willingness to throw away the first version. The patience to write your own first draft.
In five years, the craftspeople will win, because they are capable of critical thought.
Taste is the moat. Craft is the premium. The asset class is human.
I'm building Grill.
Grill never generates. It pressure tests what you wrote by asking questions that challenge your assumptions. It makes sure you understand what you're shipping, so you've built conviction in your gut, not from what's good on paper.
There's a longer arc. I'm building an operating layer for people who refuse to ship slop. Grill is the entry point.
If this is the company you want to see exist, join the movement.
usegrill.com