There’s a certain tone that dominates AI advice right now, and once you hear it, you can’t unhear it.
Confident. Calm. Slightly smug.
Delivered like a TED Talk distilled into three bullet points and a stock photo.
It sounds helpful.
It sounds reasonable.
It sounds like it knows what it’s talking about.
Most of the time, it doesn’t.
A lot of AI advice is just vibes wearing a lab coat.
You’ve seen it everywhere:
- “Be more intentional.”
- “Add constraints.”
- “Think about your audience.”
- “Use AI as a collaborator, not a replacement.”
None of these statements are wrong. That’s the problem.
They’re not useful either.
They function like motivational posters for thinking. They create the illusion of insight without forcing a single decision. You nod, feel vaguely affirmed, and move on exactly as before.
Authority without specificity is just noise that knows how to stand up straight.
What makes this worse is that AI systems are very good at reproducing that tone. Give them a prompt like “give me advice about writing with AI” and they’ll happily generate ten paragraphs of polished generalities that sound profound and say almost nothing.
And people mistake that polish for depth.
This is how “clarity cosplay” spreads.
Advice that looks precise but never commits. Guidance that gestures toward action without ever naming one. Statements so broadly agreeable that disagreement would require inventing a straw man.
It’s not that the advice is incorrect.
It’s that it’s uninhabited.
No one had to believe anything to write it.
Real advice has fingerprints. It reflects tradeoffs. It reveals what the speaker values more than something else. It risks being wrong in a specific way.
Vibe advice avoids all of that. It floats safely above context, insulated from consequence.
You can tell you’re dealing with vibes when the advice still “works” no matter what you’re doing.
If the same tip applies equally to:
- writing a novel
- designing a landing page
- choosing a thumbnail
- planning a life change
It’s not guidance. It’s wallpaper.
AI didn’t invent this problem, but it accelerates it. Because once advice is detached from lived decision-making, it becomes infinitely reproducible. Models are trained on oceans of content that already favors generality over commitment, so they return exactly what they’re given.
And then people ask, “Why does all AI advice sound the same?”
Because most human advice already did.
The uncomfortable part is that vibes-with-authority advice thrives because it protects everyone involved. The person giving it doesn’t have to take responsibility for outcomes. The person receiving it doesn’t have to change anything.
No friction. No failure. No accountability.
Just a warm sense that something wise occurred.
Contrast that with advice that actually helps.
It sounds narrower. Stranger. Occasionally annoying.
It says things like:
- “This will make your work worse in the short term.”
- “Pick one audience and disappoint the rest.”
- “Stop refining this. Use it badly for a month.”
- “If you won’t choose, nothing downstream matters.”
That kind of advice doesn’t scale well. It doesn’t trend. It doesn’t look good in a carousel post.
But it does something dangerous.
It forces a decision.
This is why people often prefer vague AI advice to concrete human critique. Vibes don’t demand courage. Specificity does.
And to be clear, this isn’t a call to reject AI advice wholesale. It’s a call to stop confusing tone with substance.
When reading or generating advice, ask one simple question:
What would I actually have to do differently if I followed this?
If the answer is “nothing concrete,” you’ve found vibes.
If the advice doesn’t make you slightly uncomfortable, slightly defensive, or slightly resistant, it probably isn’t touching anything real.
Authority isn’t how something sounds.
It’s what it costs to say.
AI can help articulate decisions. It can sharpen thinking that already exists. It can surface options you hadn’t considered.
What it cannot do is replace the work of choosing.
So the next time a piece of AI advice feels profound, pause.
Ask where it stands.
Ask what it excludes.
Ask who it would annoy.
If the answer is “no one,” you’re not looking at insight.
You’re looking at vibes.
— Sven
