The Mistake Almost Everyone Makes
In November 2012, Paul Graham published "How to Get Startup Ideas," and its first lesson is a warning against the very thing the title promises. "The way to get startup ideas is not to try to think of startup ideas," he wrote. "It's to look for problems, preferably problems you have yourself."
This is the mistake almost every aspiring founder makes. They sit down, often with a whiteboard and a co-founder, and try to brainstorm a billion-dollar idea. The output of that session is predictable: a list of ideas that sound plausible when described but that no one actually wants. Graham calls these "made-up" or "sitcom" startup ideas, the kind a TV writer would invent for a character who's supposed to be a startup founder. They have the surface texture of a real startup but no underlying demand.
The reason brainstorming fails is structural. When you try to invent an idea, you're working from imagination, and imagination is bad at predicting what people actually need. When you look for problems, especially ones you experience yourself, you're working from reality, and reality is where demand lives. The shift from "what would be a good idea?" to "what's a real problem I keep running into?" is the entire reorientation Graham is asking for.
This connects to a hard truth about markets. Most ideas that sound exciting in a pitch are solutions hunting for a problem. The durable ones start from a problem and find the solution. It's the same lesson at the core of product-market fit: demand has to exist before you build, and you find demand by looking, not inventing.
The Three Traits of Great Ideas
Graham gives a precise description of what the best ideas look like. "The very best startup ideas tend to have three things in common: they're something the founders themselves want, that they themselves can build, and that few others realize are worth doing."
Each trait does specific work.
Something the founders want. If you want it, at least one person does, and you have a built-in user to learn from: yourself. Founders who build for their own need have an enormous advantage, because they can sense whether the product is good without running a survey. Apple traces back to Steve Wozniak building the kind of computer he himself wanted and couldn't otherwise afford. Facebook grew out of a world Mark Zuckerberg already lived in, an online campus life that felt natural to him and his classmates at Harvard.
Something they can build. An idea you can't actually create isn't an idea, it's a wish. The best founders are at the leading edge of a technology, which means the gaps they notice are also gaps they're equipped to fill. The ability to build is what turns a noticed problem into a real company.
Something few others realize is worth doing. This is the trait people underrate. If everyone already sees the opportunity, you're in a crowded race. The most valuable ideas are the ones that look unpromising or trivial to most people, because that lack of competition gives you room to grow before anyone else takes it seriously.
The intersection of all three is rare, and that rarity is why great ideas are valuable. Plenty of ideas satisfy one trait. Something you want but can't build is a daydream. Something you can build that everyone sees is a commodity. The magic is in the overlap, and the overlap usually comes from deep personal involvement in a fast-moving field.
| Trait | What it gives you | Failure mode without it |
|---|---|---|
| Founders want it | A real user to learn from (you) | Building for a customer you don't understand |
| Founders can build it | The ability to actually ship | An idea that stays a wish |
| Few realize it's worth doing | Room to grow uncontested | A crowded race you're late to |
Notice, Don't Think Up
One of Graham's most useful reframes is about the verb you use. "The verb you want to be using with respect to startup ideas is not 'think up' but 'notice.'"
This sounds like a small distinction, but it changes the whole activity. "Think up" implies invention from nothing, sitting in a room generating possibilities. "Notice" implies awareness, catching something that's already there but that most people walk past. The best ideas, Graham argues, aren't manufactured. They're spotted by someone standing in the right place with their eyes open.
The catch is that you can only notice what you're positioned to see. A problem invisible to outsiders is obvious to someone embedded in the field. This is why domain immersion beats brainstorming. The founder of Stripe noticed how painful online payments were because they were trying to build things that needed payments. The founders of dozens of vertical software companies noticed gaps because they worked in those industries first.
Noticing also requires a particular openness. Many founders see a problem and unconsciously dismiss it, thinking "surely someone has solved this" or "this is too obvious to be a real opportunity." That reflexive dismissal is the enemy. The skill is to notice the gap and then take it seriously instead of explaining it away. Often the reason no one has solved an obvious problem is simply that solving it requires unglamorous work everyone avoids, which we'll return to.
Live in the Future
One of Graham's most quoted lines from the essay is also among his most actionable: "Live in the future, then build what's missing."
The idea is that startup opportunities are most visible to people who already inhabit the leading edge of a changing field. If you're working with the newest technology, the newest tools, the newest ways of doing things, you experience the future before most people do. And from that vantage point, the gaps are obvious. The things that "should exist but don't yet" feel like daily frustrations rather than abstract opportunities.
When the web was new, the people living in that future noticed what was missing and built it. When smartphones were new, the same thing happened. When AI tools became newly capable, founders living at that edge immediately felt the gaps between what was suddenly possible and what actually existed. In each case, the founders weren't predicting the future from the outside. They were living in it and reporting what they found absent.
This is why Graham advises young would-be founders to focus less on coming up with ideas and more on getting to the leading edge of something. "If you're at the leading edge of a field that's changing fast, when you have a hunch that something is worth doing, you're more likely to be right." The strategy isn't to search for ideas. It's to put yourself somewhere that ideas become visible, and then pay attention.
Getting to that edge is a learning problem. It means going deep enough into a fast-moving domain that you feel its frontier. That kind of deliberate, compounding learning is exactly what separates founders who spot opportunities from those who chase them, a theme we explore in how elite founders build knowledge systems.
Why the Best Ideas Look Bad
A surprising property of great startup ideas is that they tend to look bad at first. Graham makes this explicit, and it's one of the essay's most liberating insights.
Consider the examples. Google was the umpteenth search engine entering a market everyone considered solved and unprofitable. Facebook was a social network just for college students, a tiny, unserious-looking market. Apple sold a computer kit to hobbyists when the idea of a personal computer struck most people as pointless. Airbnb asked strangers to sleep in other strangers' homes, which sounded both unsafe and absurd.
In each case, the apparent badness of the idea was protective. Because the idea looked unpromising, smart and well-resourced competitors stayed away, giving the founders room to build before the opportunity became obvious. By the time the idea looked good to everyone, the original founders had a commanding lead.
Graham's practical advice follows from this: turn off the filter that makes promising-but-weird ideas seem unworthy. Most people have a mental filter that screens out ideas that sound strange, small, or low-status, and that filter is exactly what causes them to miss the best opportunities. The ideas that survive everyone's filter are, almost by definition, the crowded ones. The valuable ideas are sitting just past the filter, looking a little too odd to take seriously.
This requires a specific kind of courage, the willingness to pursue something that your friends will politely doubt. The discomfort of working on an idea that sounds bad is, paradoxically, a signal you might be onto something real.
Schlep Blindness
Closely related is a phenomenon Graham named in a companion essay: schlep blindness. A "schlep" is a tedious, unpleasant task. Schlep blindness is the way founders unconsciously avoid ideas that involve a lot of schleps, without even realizing they're avoiding them.
Stripe is Graham's prime example. The problem Stripe solved, making online payments easy for developers, was obvious and painful for years. Yet few talented founders attacked it, because doing so meant wading through the genuinely unpleasant world of banking relationships, compliance, and financial regulation. The schleps scared people off before they consciously evaluated the opportunity. The Collison brothers were willing to do the unglamorous work, and they built one of the most valuable companies of their generation in a space hiding in plain sight.
Schlep blindness is dangerous precisely because it operates below awareness. You don't think "this idea involves too much annoying work, so I'll skip it." Your mind quietly routes around it, and you never seriously consider it at all. The defense is to deliberately examine the ideas you've dismissed and ask whether you rejected them on merit or just because they looked like a lot of work.
The upside is large. Because schleps repel most founders, the ideas full of them are under-competed. If you're willing to do the hard, boring parts that scare others off, you get access to opportunities that look bad for a reason no one wants to say out loud: they're a pain. That pain is your moat. It's closely related to the way unaddressed difficulty becomes a knowledge debt that quietly kills startups when founders avoid the hard learning a problem demands.
Organic Ideas vs Made-Up Ideas
Graham draws a sharp line between two kinds of ideas, and the distinction is one of the most practically useful parts of the essay.
Organic ideas grow out of your own life. You experience a problem, it nags at you, and eventually you realize you could solve it. These ideas come pre-validated by at least one real user, you, and they tend to be grounded in genuine demand because they started from genuine frustration. Most of the famous startups began this way.
Made-up ideas are generated deliberately, usually in a brainstorming session, to fill the slot labeled "startup idea." They sound plausible. You can pitch them. But they're disconnected from real experience, and they usually solve problems that sound real but aren't urgent for anyone. Graham's term for the worst version is the "sitcom" startup idea: it has the form of a startup without the substance.
The trap is that made-up ideas are easier to produce on demand. If you need an idea this week, you can manufacture a plausible one by lunch. Organic ideas can't be summoned on a schedule. They come from accumulated experience and require patience. This is why Graham advises would-be founders not to force it. If you don't have an organic idea yet, the answer isn't to manufacture one. It's to go get the experience and learning that organic ideas grow from.
| Dimension | Organic ideas | Made-up ideas |
|---|---|---|
| Source | Real problems in your own life | Brainstorming sessions |
| Validation | At least one real user (you) | Sounds plausible to listeners |
| Demand | Grounded in genuine frustration | Often solves a non-urgent problem |
| Availability | Can't be forced; comes with experience | Producible on demand |
| Typical outcome | The famous startups | The forgettable ones |
How to Apply the Method
Graham's essay is a method, not just an observation. Here's how to run it.
Get to the leading edge of something. Pick a fast-changing field and go deep enough to feel its frontier. This is the single highest-leverage move, because it's what makes good ideas visible. You can't notice gaps in a field you only understand superficially.
Keep a problem journal. Every time something frustrates you, especially in your work or in a fast-moving domain, write it down. Don't filter for "is this a startup?" Just capture the friction. Over months, patterns emerge, and the recurring frustrations are your candidate ideas.
Audit your dismissed ideas for schlep blindness. Review the ideas you've rejected and ask honestly: did I reject this on merit, or because it looked like tedious, unglamorous work? The pile of "too much hassle" ideas often contains the best opportunities.
Turn off the badness filter. When an idea sounds too small, too weird, or too low-status, treat that as a possible signal rather than a verdict. Ask what would have to be true for this bad-sounding idea to be great. The best ideas survive this question.
Build for yourself first. If you want it, build it for you. Being your own first user gives you a feedback loop nobody else can match, and it's how you tell an organic idea from a made-up one.
Don't force it. If you don't have a real idea yet, resist the urge to manufacture one. Go accumulate experience and learning instead. The idea will come when you're positioned to notice it.
Once you have a candidate, the next questions are about the terrain and the timing, which is where the idea maze and the discipline of getting your first users through doing things that don't scale take over.
Building the Noticing Habit
The throughline of Graham's method is that good ideas come to prepared minds. You can't schedule a flash of insight, but you can systematically build the conditions that make insight likely: deep knowledge of a fast-moving field, a habit of capturing problems, and a mind trained to notice rather than dismiss.
This is fundamentally a learning practice. Living at the leading edge means continuously absorbing the newest developments in your domain. Noticing what's missing means paying attention to your own frustrations and the frustrations of others, then capturing them before they fade. The founders who reliably generate good ideas aren't luckier. They've built a system for staying at the edge and catching what they see there.
A capture habit makes this concrete. As you read about your field, highlighting the developments, problems, and observations that stand out with Glasp's web highlighter turns passive reading into a searchable record of what's changing at the frontier. A lot of leading-edge knowledge lives in talks, interviews, and conference videos, and YouTube Summary by Glasp lets you extract the key points from an hour of founder or technical content in minutes, then highlight what matters.
Over time, this collection becomes a map of your field's frontier. When you sense a gap, Glasp's AI chat lets you query everything you've saved, testing whether your hunch connects to patterns you've noticed before. Seeing what others at the edge are highlighting through the community feed adds a second set of eyes on the same frontier. The method isn't to think harder. It's to position yourself to notice, and then make sure you capture what you do.
Frequently Asked Questions
Why does Paul Graham say not to try to think of startup ideas?
Because deliberately trying to invent ideas produces "made-up" ones that sound plausible but lack real demand. When you brainstorm, you work from imagination, which is bad at predicting what people actually want. When you look for problems you genuinely experience, you work from reality, where real demand lives. The instruction isn't to stop thinking. It's to start from problems instead of from the desire to have an idea.
What does "live in the future, then build what's missing" actually mean?
It means putting yourself at the leading edge of a fast-changing field so you experience emerging possibilities before most people do. From that vantage point, the gaps between what's now possible and what actually exists feel like everyday frustrations. Those gaps are startup ideas. The advice is to get to the frontier of a domain, live there, and then build the obvious-to-you things that are still missing.
Why do the best startup ideas look bad at first?
Because an idea that looks obviously good attracts immediate competition from smart, well-funded people. An idea that looks small, weird, or unpromising scares those competitors away, giving the founders room to build a lead before anyone takes the opportunity seriously. Google, Facebook, and Apple all looked unpromising early on. By the time they looked good to everyone, the founders were already far ahead.
What is schlep blindness and how do I overcome it?
Schlep blindness is the unconscious tendency to avoid startup ideas that involve a lot of tedious, unpleasant work, without even realizing you're avoiding them. Because the avoidance happens below awareness, the defense is deliberate: review the ideas you've dismissed and ask whether you rejected them on merit or just because they looked like a hassle. The unglamorous, schlep-heavy ideas are often the least competed and most valuable.
I don't have a startup idea yet. What should I do?
Don't manufacture one. Graham's advice is to go get the experience that organic ideas grow from. Pick a fast-changing field and go deep until you reach its leading edge. Keep a journal of problems and frustrations you encounter. Build the habit of noticing gaps rather than dismissing them. The idea will come when you're positioned to see it, and a forced, made-up idea is worse than no idea at all.
Conclusion: Become the Kind of Person Ideas Come To
Paul Graham's method flips the usual advice on its head. You don't get startup ideas by trying to get startup ideas. You get them by becoming the kind of person they come to: someone at the leading edge of a changing field, attuned to problems, willing to take bad-sounding and schlep-heavy ideas seriously, and building things they themselves want.
That's a position you cultivate over time, not a trick you apply in an afternoon. It rests on continuous learning, deliberate noticing, and the discipline of capturing what you observe before it slips away.
Start building those conditions now. Highlight the developments and problems at your field's frontier with Glasp's web highlighter, turn founder and technical talks into searchable notes with YouTube Summary, and query your growing record of observations with Glasp's AI chat. Live in the future, keep your eyes open, and capture what you notice. The ideas will find you.