Pocket Guide to Startup Essentials

Pocket Guide to Startup Essentials | Generated by AI

Home 2026.07

Question: What is Y Combinator’s “Pocket Guide for Startups”?

Answer

The YC Pocket Guide is a condensed version of Y Combinator’s core startup philosophy. It is basically the founder operating system YC teaches: ship fast, talk to users, find product-market fit, stay small, and iterate aggressively. YC published a similar list as “YC’s Essential Startup Advice.” (Y Combinator)

The whole idea can be summarized as:

A startup is not a smaller company. It is a machine for finding a repeatable way to create value.

1. Launch now

Do not spend 6 months building the perfect product.

Bad:

Idea
  ↓
Architecture
  ↓
Design docs
  ↓
Hiring
  ↓
Launch

YC style:

Idea
  ↓
Build ugly MVP
  ↓
Users
  ↓
Feedback
  ↓
Iteration
  ↓
Product-market fit

The market gives you the specification.

2. Build something people want

This is the YC equivalent of the “north star”.

Not:

“Can we build this?”

But:

“Will somebody be upset if this disappears?”

Examples:

Dropbox:
- Not “cloud storage technology”
- Problem: “I need my files everywhere”
Airbnb:
- Not “marketplace website”
- Problem: “Hotels are expensive, people have unused rooms”

3. Do things that don’t scale

Early stage founders should manually solve problems.

Examples:

Airbnb early:

Founder → visits apartments
Founder → takes photos
Founder → talks to hosts
Founder → gets listings

Not:

Build automated marketplace system first

Manual work teaches you the real problem.

4. Find 10-100 customers who love your product

YC does not care about vanity metrics.

Bad:

100,000 registered users

Good:

50 users
  |
  +-- use every day
  +-- complain when service breaks
  +-- pay money
  +-- recommend friends

The first 100 customers are your product design team.

5. Write code, talk to users

Many technical founders hide behind engineering.

Example:

Month 1:
  rewrite backend
  optimize database
  build Kubernetes cluster

Users:
  zero

YC founder:

Morning:
  talk to 5 users

Afternoon:
  implement what users requested

6. Find the 90/10 solution

Do not optimize everything.

Find the smallest thing that creates most value.

Example:

You want an AI coding assistant.

Wrong:

Build:
- own model
- own IDE
- own cloud
- own vector DB
- own agent framework

Right:

Day 1:
OpenAI API
+
VSCode extension
+
one killer workflow

Then expand.

7. Growth comes from product, not marketing

YC’s view:

Great product
      |
      ↓
Users tell others
      |
      ↓
Growth

Not:

Bad product
      |
      ↓
Facebook ads
      |
      ↓
Growth

Marketing can amplify something good. It rarely fixes something bad.

8. Stay small before product-market fit

Before PMF:

2 founders
100 users
fast iteration

After PMF:

engineering team
sales
marketing
operations

Hiring too early creates bureaucracy.

9. Ignore competitors

YC’s famous mindset:

Your biggest enemy is usually not another company.

It is:

building something nobody wants
moving too slowly
running out of learning speed

Competition matters after you prove demand.

10. The founder relationship matters

YC strongly emphasizes co-founder relationships.

Because startups are:

uncertainty
+
stress
+
failure
+
long periods without reward

A bad co-founder relationship destroys companies faster than technology problems.

For an AI startup in 2026, the YC version becomes

Week 1

Build:

AI agent
+
one painful workflow
+
10 target users

Example:

AI assistant for Chinese manufacturers

Instead of:
"AI platform"

Do:
"Generate export documents automatically"

Month 1

Goal:

10 users love it

Not:

10000 users registered

Month 3

Goal:

People pay

Example:

$50/month × 100 customers
=
$5k MRR

Now you have evidence.

YC’s entire philosophy:

Build → Launch → Talk → Learn → Iterate

repeat 100 times

The founder is basically running a scientific experiment.

YC Startup Library collects these principles and related founder resources. (Y Combinator)

For someone building AI agents/consulting products, the most relevant YC rule is probably:

Do not build an AI platform. Find one expensive human workflow and replace it.

That is how many YC AI companies are approaching the current wave.

References:

YC’s Essential Startup Advice (Y Combinator)
YC Startup Library (Y Combinator)
YC Startup School resources (startupschool.org)

Back Donate