Pg Startup Hubs

---
name: pg-startup-hubs
description: Paul Graham's analysis of startup hubs, Silicon Valley, and where to locate a startup or build a startup ecosystem. Use this whenever the user asks whether to move to Silicon Valley or another city for their startup, whether location matters, why some cities produce startups and others don't, how a city, university, or government could become a startup hub or attract founders, how immigration or talent policy affects startups, or what makes Silicon Valley work. Also use it for remote-vs-hub debates and "can we build our own Silicon Valley" questions.
---

# PG Startup Hubs

> **Essay text is not bundled.** Paul Graham's essays are copyrighted, so this repo ships only links. Run `scripts/fetch_essays.py` once to download them into `references/essays/` for local use. The script sits two folders above this skill (`../../scripts/`) in both the repo and a plugin install, and needs no arguments from there. If a file there is missing, read the essay from its URL in `references/essays.md` instead.

The core claim: death is the default for startups, and a few places supply the antidote. The antidote is people, not buildings or incubators. "If you could get the right ten thousand people to move from Silicon Valley to Buffalo, Buffalo would become Silicon Valley."

Read `references/essays/` for exact passages, especially `hubs.txt`, `siliconvalley.txt`, `startuphubs.txt`.

## Why hubs work

- Two components: being somewhere startups are the normal, cool thing to do (which pushes people from interest to actually starting), and chance meetings with people who can help. "In the Valley, lightning has a sign bit."
- Density produces chance meetings; different startups need different kinds of help (Sean Parker for Facebook, a database guy for another); and enough people with the same ambitions set the social norms so your peers push you forward instead of back to the mean.
- People in hubs help each other with no expectation of return, partly because startups are rarely killed by competitors so the game isn't zero-sum.
- Investor boldness may be the main advantage: West Coast investors repeatedly snatched YC startups from Boston investors who saw them first. "Boldness is the essence of venture investing," and the bold ones are bold because they understand startups, like skiers on diamond slopes.
- Hubs are markets, and markets centralize despite technology because face-to-face contact makes deals happen.
- Startups beget startups on roughly a five-year cycle (Shockley 1956, Fairchild 1957, Intel, Kleiner Perkins, Google 42 years later). There's no middle ground: Chicago is negligible next to Seattle.

## Should you move?

- All other things being equal, yes, to the biggest hub. The argument that London startups needn't move to the Valley equally proves small-town startups needn't move to London. "The difference between cities is a matter of degree." This is about cities, not countries.
- Reasons not to: a family, immigration barriers (one of the Valley's biggest weaknesses), an industry-specific center (entertainment in LA/NY), or a definite funding offer where you are. "You generally shouldn't pass up a definite funding offer to move."
- The question isn't whether you need it but whether it brings any advantage, because competitors who take the advantage will beat you. "We don't need to be in Silicon Valley" is a sign you aren't thinking about the question right.
- Seed-stage startups are mobile ("just a couple founders with laptops"), so seed funding is national or international, not regional, and local YC clones don't work. Wufoo succeeded in Tampa; the Wufoos are exceptionally disciplined.

## Why the US, and what a rival would need

Ten reasons startups condense in America: immigration (a mecca must let people in), wealth, not a police state, better universities (private competition; equalized systems produce no great ones), you can fire people, work isn't identified with employment, not too fussy (German incorporation capital rules; "when you hit something that would have killed Apple, prune it off"), a large domestic market, venture funding especially angels, and "dynamic typing for careers" (late bloomers over failed prodigies). Europe doesn't lack balls; it lacks examples.

To beat the Valley: a town that puts cars last (the Valley is "one giant parking lot"), low capital gains tax, smarter immigration including a founder visa, letting in the other 95% of great programmers. The US has under 5% of world population; exceptional performance implies immigration.

## How to build a hub

- Two limiting reagents: rich people and nerds. You attract both with a first-rate research university in a town both want to live in. Miami has rich people and no nerds; Pittsburgh and Ithaca have nerds and no rich people.
- Personality matters: old buildings, walkable neighborhoods, independent cafes and bookshops, tolerance of oddness (most good startup ideas seem a little crazy), youth, a center to walk to. "You don't build a silicon valley; you let one grow." Technology parks and incubators miss that the key stage is three guys in an apartment.
- A university could be built fast by hiring 200 top young researchers with large bonuses all at once, because professors choose colleagues above all.
- Pittsburgh recipe: encourage the youth food boom, preserve old buildings strictly, become the most bike- and pedestrian-friendly city, make CMU a talent magnet, stay tolerant, and get the university out of the way (no IP claims, leaves of absence, free time; "weighted by market cap, the best thing a university can do is literally nothing"). Investors come last and slowly; cheaper startups, crowdfunding, and YC make that less fatal.
- Buying one: roughly a million dollars per startup to relocate, a billion for a thousand, the price of a football stadium. The hard part is picking; piggyback on eminent Silicon Valley angels' portfolios. Start with 30 (startups need other startups), no legalistic leaving conditions, no renovated-warehouse incubator, convertible debt that converts only in a big round. Political will is "microscopically small."
- Programs with "innovation" or "entrepreneurship" in the name almost always disappoint. Aim at something specific (better batteries), not at innovation.

## How to run a session

1. Clarify which question it is: a founder deciding where to be, or a city/university/government trying to attract startups.
2. For founders: apply the degree argument and the four legitimate exceptions. Be honest that remote is a disadvantage even if survivable.
3. For ecosystem builders: check the reagents (university? people under 30 want to live there? tolerant? walkable? investors?), name what's missing, and prescribe the organic path plus the "treat founders as users and ask what they want" method. Warn against incubators, business plan competitions, and innovation centers.

## Source essays in `references/essays/`

- `hubs.txt` Why Startup Hubs Work (Oct 2011).
- `startuphubs.txt` Why to Move to a Startup Hub (Oct 2007).
- `siliconvalley.txt` How to Be Silicon Valley (May 2006).
- `america.txt` Why Startups Condense in America (May 2006).
- `maybe.txt` Can You Buy a Silicon Valley? Maybe. (Feb 2009).
- `pgh.txt` How to Make Pittsburgh a Startup Hub (Apr 2016).
- `95.txt` Let the Other 95% of Great Programmers In (Dec 2014).