Before the Startup: Why Your Instincts Are Wrong and How to Fix Them

I am a big fan or Paul Graham. Paul is a programmer, writer, and investor. In 1995, he and Robert Morris started Viaweb, the first software as a service company. He later founded the startup accelerator Y Combinator and he shaped modern Silicon Valley. He funded 3,000+ companies (including Airbnb, Stripe), and writes influential essays on technology and startups. 

If you are thinking about starting a company, Paul has a warning for you: Your instincts are probably going to get you killed.

In his lecture "Before the Startup," Paul argues that entrepreneurship is not a natural act. It is deeply "counterintuitive," meaning the things that feel like the right thing to do are often the exact things that will cause your company to fail.

Here is a deep dive into why startups are so dangerous to your intuition, and how you can actually prepare yourself to succeed.

1. Skiing Analogy: Why You Need to Rewire Your Brain

Paul compares starting a company to learning how to ski. If you are standing at the top of a steep, icy hill, your survival instinct screams at you to lean back to slow down. But if you lean back on skis, you lose control and fly down the hill disastrously. To survive, you must suppress your instinct and lean forward.

Startups are the same. Y Combinator partners spend hours telling founders not to do specific things. The founders usually ignore the advice because it feels wrong, only to come back a year later saying, "I wish we had listened".

To survive, you have to accept that you cannot trust your gut on business decisions yet. You need to learn a new set of unnatural habits.

2. Trust Your "People" Instincts

There is one area where your instincts are actually perfect: People. While you don't know business, you have spent your whole life interacting with humans. A common mistake technical founders make is meeting a potential co-founder who seems smart but gives them a "distasteful" vibe. They ignore the vibe because they think business is supposed to be distasteful.

Paul warns: Do not do this. If you have misgivings about someone, trust them. You should pick co-founders the way you pick friends. In fact, Paul defends the idea of a "monoculture" in early startups... hiring your friends is safer because trust and speed are more important than diversity in the very early days

3. Trap of "Playing House"

One of the most dangerous instincts young founders have is the desire to "look" like a startup. Paul calls this "Playing House".

We have been trained by school to jump through hoops: get the grades, join the clubs, pad the resume. Founders carry this into business by focusing on the external markers of success:

• Incorporating the company.
• Renting a cool office in a trendy neighborhood.
• Hiring their college friends.
• Trying to raise venture capital immediately.

The Counter-Example: Mark Zuckerberg. Paul points out that Zuckerberg was a "complete noob" at the mechanics of startups. Facebook was originally incorporated as a Florida LLC... which is comically the wrong way to structure a high-growth tech company. He likely knew nothing about how to raise an Angel round.

But he won anyway. Why? Because he wasn't "playing house"; he was obsessed with his users. While other founders are busy trying to act like CEOs, successful founders are busy making something people actually want

4. You Can’t "Hack" a Shark

Throughout your life, you have likely learned how to "game the system." You learned how to cram for tests to get an A without mastering the material. In the corporate world, you might learn how to look busy by sending emails late at night or even changing the clock on your computer to fake a timestamp.

In a startup, gaming the system stops working:

• There is no professor to grade you.
• There is no boss to trick.
• There are only users.

Paul compares users to sharks. Sharks are too stupid to be fooled. You can't wave a red flag and trick a shark; they only care about one thing: "Meat or no meat". Users are the same. If your product doesn't solve their problem, no amount of marketing tricks will save you.

Note: Paul advises that whenever you hear the term "Growth Hacks," you should mentally translate it to "Bullshit"

5. The "Parenting" Reality: Why You Should Wait

Paul offers a controversial piece of advice to college students: Do not start a startup yet.

He compares a startup to having children. It is a button that, once pressed, changes your life irrevocably. It takes over everything.

• The Larry Page Example: It might look cool to be Larry Page (Google co-founder). He has billions of dollars and big jets. But he also has zero freedom. He has been running as fast as he can since he was 25. If he takes a week off, a backlog of problems accumulates that only he can solve. He can never backpack through Thailand or explore a random hobby just for fun ever again.
• The Mark Zuckerberg Example: Paul explicitly highlights the loss of serendipity for Zuckerberg.
  • No Backpacking: Zuckerberg will never get to "bum around" a foreign country. He can never backpack through Thailand
  • State Visits: If he travels, it is effectively a "state visit" or he has to hide incognito at the George V hotel in Paris
  • The Trap: Paul notes that "Facebook is running him as much as he is running Facebook"

Your early 20s are the only time you will be able to live with zero overhead and zero responsibility. You can take risks, travel cheaply, and explore random projects. If you start a company now, you sacrifice that serendipity forever.

6. How to Get Ideas: "Live in the Future"

If you sit down and say, "I am going to think of a startup idea," you will almost certainly come up with a bad one. You will think of something that sounds plausible but solves no real problem.

The best way to get ideas is to "turn your brain into the type that has startup ideas unconsciously"

"Girlfriend in Taiwan" Example: Paul tells the story of a fellow student at Harvard in the mid-90s. He wanted to talk to his girlfriend in Taiwan, but long-distance calls were expensive. Because he was an expert in networks, he wrote software to turn sound into data packets and ship them over the internet. He didn't do it to build a startup; he did it to save money. He was "living in the future," so the idea of Voice over IP (VoIP) was obvious to him before it was obvious to the world.

Airbnb Example: You don't even need to be a tech wizard. The founders of Airbnb (Brian Chesky and Joe Gebbia) were art school graduates. They weren't experts in code; they were experts in design and organizing people. They solved a problem relevant to their lives, and the business followed.

Conclusion

Follow your genuine intellectual curiosity. If you become an expert in a specific domain and hang around other smart people, the startup ideas will eventually find you

Questions?

How do you measure success?

• The Wrong Metric: You think the measure of success for a startup is fundraising
• The Right Metric: Real founders know the only thing that matters is making something users want. Paul compares users to sharks—you cannot trick them. They only care about "meat or no meat"

Why is 'just learn' better than taking entrepreneurship classes?

1. The Linguistics Analogy: Theory vs. Practice

If you want to learn how to speak French, a linguistics class will teach you about the structure of languages, but it won't actually teach you how to speak French

• The Problem: Entrepreneurship classes teach you the theory of startups (mechanics, fundraising, structure), which is not what you need to know.
• The Reality: To succeed, you don't need expertise in startups; you need expertise in your own users. This cannot be taught in a classroom; it can only be learned by actually starting the company and interacting with users

2. Business Schools Teach "Management," Not Creation

Business schools were designed to train the "officer corps of large companies". They focus on management, but management is a problem you only have after you have already succeeded.

3. "Just Learning" Creates Domain Expertise

"learn powerful things"

• The Source of Ideas: The best startup ideas come from "domain expertise". Larry Page didn't succeed because he studied business; he succeeded because he was a genuine expert in search
• Living in the Future: When you focus on learning hard subjects (like physics or networks), you get to the "leading edge" of technology. When you live at the leading edge, startup ideas become obvious to you before they are obvious to anyone else.

4. Unconscious vs. Conscious Ideation

If you take a class on entrepreneurship, you are likely trying to consciously think of startup ideas, which usually results in bad, plausible-sounding ideas.

• The "Just Learn" Advantage: By following your genuine intellectual curiosity and working on side projects, you turn your brain into the type that generates ideas unconsciously. You solve problems because they are interesting to you, not because you are trying to start a company.
• Example: The student who invented a version of VoIP (Voice over IP) didn't do it to build a startup; he did it because he wanted to talk to his girlfriend in Taiwan for free and he understood networks

The component of entrepreneurship that actually matters is domain expertise. You get that by "just learning" powerful things, not by studying the mechanics of how to run a business.

How can I tell if a problem is genuinely interesting?

1. The "Intolerance" Test
   The most practical way to tell if you have a "taste for genuinely interesting problems" is to check your tolerance for boring things

• The Test: Can you tolerate "known boring things" like literary theory or working in middle management at a large company?
• The Result: If you can tolerate those things, you likely lack the instinct for interesting problems. If you find working on boring things "intolerable" and can only focus when you are genuinely engaged, your internal compass is likely working correctly'

2. The "Fractal Stain" Analogy

• The Edge: Every point on the moving edge of that stain represents an interesting problem.
• The Interior: Things inside the stain (like steam engines) are no longer interesting because they are already established. To find interesting problems, you must locate yourself at the "leading edge".

3. It Seems Obvious to You (But Not Others)

If you successfully "live in the future" (by becoming an expert at the leading edge of a field), the most interesting problems will not feel like "startup ideas." They will feel like things that simply "ought to exist"

• The Evidence: The solution will seem "obvious to you" while appearing "uncannily prescient" to everyone else.
• Example: The student who invented VoIP (internet calling) didn't think he was building a startup; it just seemed obvious to him that audio should be sent as data packets so he could talk to his girlfriend for free.

4. Skepticism from Outsiders (and Parents)

A sign that you are working on a genuinely interesting new problem is that other people, specifically your parents, think you are wasting your time.

• The "Twitch" Factor: Paul Graham admits that when he first heard of Twitch.tv (broadcasting people playing video games), his reaction was "what?". It seemed trivial, but it was a genuinely interesting problem to the founders and their users. History is full of successful founders whose parents thought they were working on unimportant things.