Summary of "Founders at Work"

Core Idea

Start before you're ready: Build prototypes and ship fast; perfect plans kill more startups than imperfect launches.
Solve problems you experience: Best products emerge from founders' own frustrations (email, blogging, bookmarking, DVRs).
Execution beats ideas: First-mover advantage matters far less than building something users actually want and iterating relentlessly.

Self-fund as long as possible to maintain control; consulting work can bridge early runway.
Ship cheap, iterate fast: Minimize burn rate (use free tools, virtual hosting, Linux servers); move fast when momentum builds.
Don't research yourself into paralysis: Naive optimism about "impossible" problems can be an advantage; avoid studying why competitors failed.
Kill projects that aren't working: Side projects can become the company (Gmail, Flickr), but abandon what doesn't grow.

Build working code constantly; never wait for perfection before testing assumptions with real users.
Focus ruthlessly: Prioritize the 80% solution over perfect edge cases; feature creep kills startups.
Simplicity wins: Don't overengineer; design for actual user behavior, not assumptions.
Tagging/simple features can revolutionize discovery: Emergent behavior (real-time events, collaboration) comes from user creativity, not planned features.

Choose complementary skills + deep trust: Technical + business balance matters more than hiring generalists.
Establish ground rules before conflicts arise: Agree on decision-making, exit conditions, and dissolution terms upfront.
Hire slowly for cultural fit: Bad early hires set terrible precedents; transparency and honesty beat artifice.
Separate technical founders from management roles: Engineers rarely make good CEOs without explicit development; bring in experienced operators early if growth is explosive.

Raise when momentum is high, not desperate: VCs call weekly once you're growing; struggling pitches get ignored.
Avoid VC if you can bootstrap: Maintains control, forces prioritization, prevents over-scaling (Basecamp, Firefox, craigslist all thrived without external pressure).
Don't take more money than necessary: Larger rounds create pressure to hit unrealistic milestones and force unwanted pivots.
Negotiate hard on control: Board seats, veto rights, and dissolution terms matter more than valuation.
Know your walk-away number before negotiations: Prevents panic-driven bad deals; understand acquirer motivation before selling.

Organic word-of-mouth > paid advertising: Build product so good users evangelize it; don't plan marketing campaigns.
Users will tell you what they want—dig deeper: Don't just add requested features; understand root problems.
Self-policing communities scale better than corporate control: Let users flag/moderate content; saves moderation costs and builds trust.
Listen obsessively: Company direction shifts from community suggestions, not founder vision alone.

Founder personality shapes company culture permanently: Invest in values and people practices early; transparency beats spin.
Small teams with P&L responsibility outperform functional silos: 2-3 person teams accountable for both customer satisfaction and profitability.
Don't grow faster than culture can scale: ArsDigita's 50% hiring spree destroyed values; cap growth at 25%/year for culture transmission.
Don't hide behind licensing restrictions or bureaucracy: Face-to-face problem-solving with users/employees builds loyalty.

Expect 4+ hours sleep and emotional instability for 18+ months: Normalize the pain; cofounder conflict is normal—choose partners you trust enough to argue with.
Pivot from vision to vision, not solution to solution: Your original plan will be wrong; stay committed to the problem.
Turn disasters into competitive advantages: Personal problem-solving and transparency with users build trust that processes can't replicate.
Bridge financing buys time for better options: Don't make permanent decisions during panic.

Define your walk-away criteria NOW before fundraising or acquisition discussions begin.
Ship a prototype this week to test one core assumption with real users.
Document cofounder agreements (decision-making, exits, equity) before conflicts arise.
Build 3-6 months of runway through self-funding or revenue, then raise only what you need.
Set up feedback loops with 10+ real users; listen obsessively before adding features.