Sagan's Baloney Detection Kit for the Age of AI Slop

Your feed is full of confident nonsense.

Some of it is written by people.

Some of it is generated by tools that can sound certain while being wrong in a freshly laminated way.

Either way, you need a filter.

Carl Sagan gave us one in 1995.

His Baloney Detection Kit is still painfully useful because the basic problem has not changed:

People believe neat stories too fast.

AI just made the neat stories cheaper.

So here are Sagan’s nine tests, translated for the age of AI slop, productivity hype, and screenshots pretending to be evidence.

---

1. Get Independent Confirmation

One source is a rumor wearing shoes.

The move: Find another source that did not copy the first one.

If a “study says” claim is everywhere, find the actual study.

If a chart is spreading, ask where the data came from.

If an AI summary gives you a citation, open it.

Do not outsource trust to a screenshot.

Screenshots are not evidence. They are evidence-shaped rectangles.

2. Look For Serious Disagreement

Good ideas can survive criticism.

Weak ideas need fan clubs.

The move: Before accepting a framework, tool, or trend, find the best critic of it.

Not the loudest hater.

The best critic.

The person who understands the claim and still has objections.

If you cannot find serious disagreement, be careful. You may be standing inside a funnel.

This applies hard to productivity systems. I poked at that in why most goal-setting frameworks are backwards.

3. Authority Is Not Proof

“A founder said…”

“A top performer does…”

“A famous investor believes…”

Fine.

Interesting.

Still not proof.

The move: Separate the person from the evidence.

Smart people can be wrong.

Successful people can be wrong outside the narrow lane that made them successful.

Famous people can be paid.

Do not hand your brain to someone because their bio has shinier nouns.

4. Generate More Than One Explanation

Your first explanation is usually convenient.

That does not make it true.

The move: List three other explanations before you settle.

Your content is not growing?

Maybe the algorithm changed.

Maybe the offer is unclear.

Maybe the hook is weak.

Maybe the audience does not care yet.

That fourth one hurts, which is why it deserves attention.

If you are blaming the platform for everything, read the real reason your content is not growing before you build a conspiracy board out of analytics.

5. Be Suspicious Of Your Favorite Theory

The idea you love most needs the roughest inspection.

Cruel? Maybe.

Useful? Absolutely.

The move: Ask, “Would I believe this if someone else said it?”

This question is rude in the correct direction.

Side project not working?

Maybe the market needs more time.

Maybe your positioning is bad.

Maybe the offer is boring.

Maybe you are protecting the version of the story where you were almost right.

That one stings.

Check it anyway.

6. Put A Number On It

“AI saves me hours.”

How many?

“This method doubled my output.”

Compared to what?

“This tool changed my workflow.”

Which part?

The move: Add numbers before you let a claim impress you.

You do not need a laboratory.

You need a baseline.

Before and after.

Time spent.

Money saved.

Errors reduced.

Output shipped.

Claims get less magical when they have to stand next to numbers.

Good.

That is the point.

7. Test Every Link In The Chain

Bad advice often hides inside a long chain.

“Build in public, then you get an audience, then you get customers, then you get a business.”

Lovely little staircase.

Also fragile.

The move: Check each link.

• Does building in public actually attract the right people?
• Does attention turn into trust?
• Does trust turn into buying?
• Does buying turn into a business that can last?

One broken link breaks the chain.

Do not let a smooth sentence carry four untested assumptions.

8. Prefer The Simpler Explanation First

Simple does not mean shallow.

It means fewer moving parts.

The move: Before choosing the dramatic explanation, test the boring one.

Launch flopped?

Maybe the algorithm shifted.

Maybe the headline was unclear.

Start with the headline.

New habit failed?

Maybe your identity is misaligned with your deeper values.

Maybe you scheduled it at a terrible time.

Start with the terrible time.

The boring answer is not always correct.

It is just a better first suspect.

9. Ask If The Claim Can Be Wrong

This is the big one.

If a claim has no losing condition, it is not advice.

It is a fog machine.

The move: Ask, “What would prove this false?”

If the answer is nothing, walk away.

Examples:

• “The right opportunity appears when you are ready.”
• “If it did not work, you did not believe hard enough.”
• “The method works, but only when you fully surrender.”

No losing condition.

No test.

No thanks.

How To Use The Kit

Do not run all nine tests on every sentence you read.

You have a life.

Use the kit when a claim makes you nod too fast.

That is the danger signal.

When something feels perfectly aligned with your frustration, ambition, or fear, slow down.

Ask three questions:

• Who else confirms this?
• What is the simpler explanation?
• What would prove it wrong?

Those three catch a lot of nonsense.

Not all of it.

Enough to make you harder to fool.

The Point

Sagan did not write the kit for AI content.

He did not need to.

The human weakness is older than the tool.

We like certainty.

We like clean stories.

We like advice that flatters the thing we already wanted to do.

AI just produces more of that, faster, cheaper, and with smoother grammar.

So keep the kit close.

Use it before you share the claim.

Use it before you buy the course.

Use it before you rebuild your life around a thread.

If you want to build the larger habit, books that teach clear thinking under pressure are a good next step.

But start here:

Do not let fluent nonsense wear your face as a mask.