Pick a number between 1 and 10, and go with your gut. If you landed on 7, you just joined a crowd: about one person in three does the same. People are predictable, and that stays mildly interesting until you bring a machine into it.
A LinkedIn post making the rounds runs the same test on the models we trust with real decisions. Ask Claude, ChatGPT, Gemini, or Grok to pick a number, and the original poster says the answer lands on 7 almost every time. Humans have bias, but the writer's point is sharper: AI carries far more of it, and hides that behind a confident tone.
That party trick stops being funny the second you remember how often you ask these tools 'is this good?' When millions query the same model with the same question, they all get nudged toward the same answer. Here is what you get from this piece: a simple rule that flips AI from a judge into a map of the crowd.
Your top salespeople should be spending their time creating revenue. Instead, they’re updating CRM records, scheduling meetings, chasing signatures, and managing the administrative work surrounding every deal. Those tasks matter, but they’re not the highest-value use of your most expensive talent. The free guide 5 Reasons You Aren’t Hitting Quota: The Sales Operations Gap Examined helps sales leaders identify the operational gaps that can pull reps away from selling. BELAY supports sales teams by matching them with U.S.-based Sales Assistants who manage the work around the sale, from CRM management and lead routing to follow-up coordination.
*Ad
The number 7 says everything about how these models work
Models are built to return the most likely answer. That is exactly what you want for facts, grammar, and a fast summary of a dense report. It is poison for taste, because taste is the art of not choosing the obvious thing.
Seven feels random to you. It is not random to the model. It sits at the statistical center of a billion human guesses, and the machine lands there because landing there is its whole job.
Why the crowd's answer becomes worthless
The creator drops one line that I keep repeating to myself: 'if everyone can do it, no one wants it.'
Think about the scale. Millions of people type 'which logo is better?' into the same handful of models.
They all get steered toward the same safe pick, and that pick stops being tasteful the moment it becomes everyone's pick.
This is the part that sold me. The tool is not rating your quality. It is showing you the exact spot where the whole internet already stands.
What this quietly does to everyone's work at once
Here is the consequence few people flag. When a tool optimizes for the most likely answer and millions lean on it, output converges. Your LinkedIn hook starts to sound like every other LinkedIn hook.
I think this is the real cost of outsourcing judgment. The convergence is the cost. You do not get a bad answer: you get the same answer as your competitor, your neighbor, and the person you want to stand out from.
In partnership with
Discover laptops with built-in intelligence.
Shift your workforce into hyper-productivity mode with Dell Pro laptops powered by Intel® Core™ Ultra with Intel vPro®. Shop on Dell Premier.
*Ad
The contrarian rule: find the 7, then leave
The advice from the author is blunt: never ask AI for its opinion. Use it to find the 7, then go somewhere else.
I love this framing because it does not tell you to quit the tool. It changes the job you hand it. The model becomes a map of average, so you know precisely where not to land.
Six moves that turn AI into a taste engine
Kill the flattery first. Do not ask 'is this good?' Ask for the three weakest parts of your draft and why a skeptical reader would bounce.
Then make the default visible. Ask what most people would write here, and you have named the zone to avoid. Have the model list every cliche your niche is posting right now, then ban all of them.
'Make it better' only gets you a shinier 7. Ask for a stranger angle, a sharper counterargument, or one very specific example instead. Feed it work you love and explain what makes it yours, so it has a reference point that is not the average of the web.
One more that people miss: taste expires. What felt fresh six months ago might be everyone's 7 today, so audit your go-to styles before they calcify.
Reading six moves is easy. Making them a reflex takes reps. 3 Minute AI runs three-minute lessons that each end with a task you actually do in a built-in chat lab, so you can practice asking for the weak spots instead of the compliment. Daily lessons and two full courses are free.
Turn AI into Your Income Engine
Ready to transform artificial intelligence from a buzzword into your personal revenue generator
HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.
Inside you'll discover:
A curated collection of 200+ profitable opportunities spanning content creation, e-commerce, gaming, and emerging digital markets—each vetted for real-world potential
Step-by-step implementation guides designed for beginners, making AI accessible regardless of your technical background
Cutting-edge strategies aligned with current market trends, ensuring your ventures stay ahead of the curve
Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.
*Ad
A naming example you can steal
Say you are naming an AI writing tool. The old way: type 'give me a name,' then pick a favorite from the list. You drown in mashups of 'Write,' 'AI,' and 'Genius.'
The better way flips it. Ask for the 20 most predictable names first, map the crowded zone, then start your real brainstorm outside it. Bring your own instinct, and maybe a human friend or two.
The same trick works for logos, headlines, and hooks.
Ten minutes to leave the 7 behind
Today, open your AI, ask it for the obvious take on whatever you're working on, write that answer down, and then deliberately build in the opposite direction.
If you want the six moves and the naming example laid out step by step, here's the exact workflow for turning AI into a taste engine.
Worth 10 minutes if you keep shipping work that feels competent but somehow blends into everyone else's.
Credits to the original creator.
Five reads for anyone tired of the 7
Seven really is the "random" number: YouGov asked Britons to pick a number from 1 to 10, and a third said 7 while almost nobody touched the edges. This is the human bias the models inherited.
AI models have favorite numbers: TechCrunch covers an experiment where GPT, Claude, and Gemini each kept returning the same "random" pick between 0 and 100. Proof the 7 problem goes well past 1 to 10.
AI boosts creativity, shrinks variety: A Science Advances study found AI-assisted stories were rated more creative but ended up more similar to each other. The convergence cost from this issue, measured.
Verbalized Sampling, the anti-7 prompt: Stanford-led researchers show that asking a model for five answers with their probabilities pulls it off the default and boosts creative diversity. A one-line trick you can test tonight.
Paul Graham on taste for makers: A classic essay arguing that good taste is real, learnable, and mostly about seeing past what everyone else accepts. The human half of the taste engine.
Poll: You ask AI "is this good?" and it says yes. What do you actually do?
Hit reply and tell us why.