Most people treat startup ideas like they’re hunting for a rare Pokémon. They sit in a room, maybe with a whiteboard and some overpriced coffee, and try to "think" of a billion-dollar company. Paul Graham, the guy who co-founded Y Combinator and basically saw the birth of Reddit, Airbnb, and Dropbox, says that is exactly how you fail.
If you’re trying to think of startup ideas, you’re already behind. The best ideas don't come from brainstorming sessions. They come from living a specific kind of life and noticing when something feels broken. It’s about observation, not invention.
The Problem With "Startup Ideas"
Honestly, the phrase "startup idea" is kinda misleading. It makes it sound like a finished product you just need to unwrap. In reality, most successful companies started as something that didn't even look like a business.
Google was just a better way to search for stuff in a sea of terrible 1990s web directories. Facebook was a way for Harvard kids to see who was in their classes. Neither of them started with a massive business plan. They started because someone noticed a gap.
Graham’s core thesis in Paul Graham: How to Get Startup Ideas is that you should stop trying to think of ideas and start looking for problems. Specifically, problems you actually have.
Why you should solve your own problems
If you solve a problem you have yourself, you at least know the problem exists. That sounds stupidly obvious, but you'd be shocked how many people build things for "users" who don't actually exist.
👉 See also: Dubai Islamic Bank Stock: Why Local Investors Are Betting Big Right Now
When you build for yourself, you are your first user. You know exactly what the "crappy version one" needs to do to be useful. If you’re building a social network for pet owners (a classic "bad" idea Graham often cites), you’re guessing what they want. You’re building a model of a user that might be totally wrong.
Live in the Future and Build What’s Missing
One of the most famous bits of advice from Graham is to "live in the future."
This isn't some sci-fi advice. It basically means you should be at the leading edge of some kind of change. If you were a programmer in 2008, you were living in the future of mobile apps before everyone else. If you're playing with LLMs and AI agents right now, you're living in the future.
When you live at the edge, the things that are missing become obvious to you. You don't have to "think" of them. They just irritate you.
- The Leading Edge: Find a field that is moving fast.
- The Irritation: Pay attention to the things that make you say, "Why is this so hard?"
- The Build: Create the thing that fixes that irritation.
Beware of "Schlep Blindness"
This is a big one. A "schlep" is a tedious, unpleasant task. Most people have an unconscious filter that prevents them from seeing ideas that involve too much boring work.
Take Stripe. Before Stripe, every developer knew that accepting payments online was a nightmare. You had to talk to banks, deal with weird regulations, and write hundreds of lines of code just to take a credit card. It was a massive schlep.
Thousands of programmers saw this problem. But almost all of them ignored it because their brain told them, "Ugh, dealing with banks sounds miserable. I’ll just build a recipe-sharing app instead."
The Collison brothers didn't have that filter. They dove into the schlep. Because they were willing to do the boring, hard work of fixing payments, they built a company worth billions. If an idea looks hard and unsexy, that's actually a good sign. It means you’ll have less competition.
The "Well" vs. The "Swamp"
When you're starting out, you have two choices for the "shape" of your idea.
- The Swamp: Something that a lot of people want a little bit. (Think: another generic task manager).
- The Well: Something that a small number of people want a lot.
You want the well.
It is much easier to expand from a small group of people who love you than to try and make a large group of people stop being indifferent. Facebook started with just Harvard. If it had launched for everyone at once, it probably would have died. It needed that initial "well" of intensity to get moving.
How to tell if your idea is a well
Ask yourself: Who wants this right now? Who wants it so much they’ll use a buggy, ugly, incomplete version of it? If you can’t name a specific group of people who are that desperate, your idea is probably too shallow.
Don't Dismiss the "Toy" Ideas
Many of the biggest companies in the world looked like toys when they started.
📖 Related: Why the United States Dollar to Chinese Yuan Rate is Hurting Your Wallet Right Now
- Apple I: Just a circuit board for hobbyists.
- Twitter: A way to tell people what you had for lunch.
- Airbnb: Renting out an air mattress on a floor.
People dismiss these because they don't look like "serious business." But if a lot of smart people are obsessed with something that looks like a toy, pay attention. It usually means they've found a new way to do something that hasn't been professionalized yet.
The "Organic" Way to Get Ideas
If you want to get good at this, stop trying to force it. Instead, focus on these three things:
- Learn a lot about things that matter. Become an expert in something, preferably something technical or rapidly changing.
- Work on projects that interest you. Build stuff just because it's cool.
- Work with people you like. Most great ideas come from conversations between smart people who are bored with the status quo.
If you do these three things, your brain will start noticing "gaps" in the world. Those gaps are your startup ideas.
Actionable Next Steps
If you're currently staring at a blank page trying to find "the one," try this instead:
✨ Don't miss: Why How Much a Dollar in Pesos Costs You More Than Just the Exchange Rate
- Audit your "schlep" list: For the next week, write down every time you encounter a process that is annoying, slow, or involves too many emails. Don't worry if it sounds "boring."
- Look at your browser history: What are you searching for that doesn't have a good answer? What "workarounds" have you built for yourself using spreadsheets or hacks?
- Find the "edge": Spend an hour a day reading about a field you know nothing about but that is growing fast (like synthetic biology, carbon capture, or new AI architectures).
- The "Someone Should" Test: Every time you say, "Someone should really build X," write it down. Then ask yourself why you shouldn't be that person.
The goal isn't to find an idea that looks like a giant company on day one. The goal is to find a problem that is real, solvable, and worth the "schlep" to fix.