How to Get Startup Ideas (Without Sitting Around “Thinking”)

In partnership with

This week we’re coming back to one of Paul Graham’s most famous essays:
“How to Get Startup Ideas.”

If you care about finding something real to build, this essay should be required reading.

And before we go further, let’s clear something up:

NTE Pro has 6,779 startup ideas (and growing).

Paul Graham would tell you not to “sit around trying to think of startup ideas.”

So which is it?

Here’s the truth:

Paul isn’t against ideas.
He’s against fabricating them in a vacuum.

The Dangerous Way People Try to Get Ideas

Most people try to get startup ideas like this:

They open a blank doc.
They ask, “What’s a billion-dollar idea?”
They brainstorm.

Paul calls this a mistake.

The best startup ideas don’t feel like ideas.
They feel like:

“Why doesn’t this exist yet?”
“This is broken.”
“This should be easier.”

They come from proximity to problems.
From living slightly in the future.
From noticing something that annoys you before everyone else catches up.

Not from cleverness.

So Where Does NTE Fit?

If NTE were saying:

“Pick one of these ideas and go build it.”

That would contradict Paul.

But that’s not what NTE is.

NTE is not a vending machine.
It’s a pattern library.

When you study thousands of early-stage concepts, especially ones that haven’t been built yet, you start to see:

• What types of problems recur

• Where markets are forming

• What founders are not building yet

• How timing shows up before it’s obvious

It trains your eye.

And that’s exactly what Paul is advocating for.

Live in the Future. Notice What’s Missing.

One of Paul’s core lines:

“Live in the future and build what seems interesting.”

That’s the game.

But most people aren’t actually close to the frontier.
They’re reading TechCrunch headlines from six months ago.

This is is one of the reasons we built WhoFiled and why it matters.

WhoFiled tracks Form D filings, GitHub momentum, Product Hunt launches, Reddit builder posts, Hacker News reveals, podcast conversations, investor activityand filters it through your context.

Not as noise.
As signal.

So instead of asking, “What should I build?”

You start noticing:

• Why 4 companies just raised in the same niche

• Why open-source activity is spiking in a category

• Why investors are clustering around a theme

• Why something exists now that didn’t 18 months ago

That’s frontier awareness.

That’s Paul-aligned.

Want to get the most out of ChatGPT?

ChatGPT is a superpower if you know how to use it correctly.

Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.

Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.

Download the free guide

The Real Trick

Paul says the best startup ideas are often things that look like bad ideas to most people.

Stripe looked painful.
Airbnb looked weird.
Microsoft looked tiny.

They were solving problems that felt annoying, unsexy, tedious.

He calls this noticing what others ignore.

NTE exposes you to thousands of “raw” ideas, not polished unicorn narratives, so you can see how un-obvious good ideas look early.

WhoFiled shows you where energy is forming before it’s consensus.

Put those together and you’re not inventing ideas.

You’re recognizing them.

What This Means for You

If you’re sitting around trying to think of a startup idea, stop.

Instead:

• Study markets forming.

• Notice what frustrates you.

• Watch where technical momentum is building.

• Pay attention to founders operating slightly ahead of mainstream awareness.

And when something feels obvious to you but invisible to others, that’s the signal.

That’s what Paul is pointing at.

Why We’re Staying With These Essays

We keep coming back to to these essays because they don’t hype you.

They should sharpen your thinking.

They remind you that startups aren’t about cleverness.
They’re about proximity.

And tools only matter if they increase proximity to truth.

NTE Pro exists to sharpen your pattern recognition.
WhoFiled exists to surface early signal through your lens.

Neither gives you the answer.

They help you see it when it’s forming.

That’s the real alignment.

And that’s how you get startup ideas
without ever sitting down to “come up with one.”