Keynote: Rails Doesn't Scale
summary
Key claim: Rails does scale, but linear scalability is boring — adding servers or Heroku Dynos only buys predictable throughput. Concrete tools and examples: HAProxy, mod_perl roots, Docker, Erlang experiments and Heroku deployment patterns; a legacy app gets rebuilt in four months and ships weeks before press; a 1,300‑word cold email leads to joining 37signals in a month. The larger claim is social: community scales and teaching multiplies impact, producing hyperlinear returns. A later startup, Operable, is built, burns out after a 110‑hour week and eventually closes its doors.
AI-written summary · about
chapters (24)
- 0:00Introduction
- 1:02Expectations
- 1:20Scalability
- 2:23Linear Scalability
- 2:55Linear Scalability Boring
- 3:47Technical Scalability
- 4:48Story Time
- 5:38Pearl
- 6:30ModPearl
- 10:02The Java Years
- 13:29Ruby
- 13:57Building at 37 Signals
- 15:18Postmortems Nose
- 18:17We had no idea what were doing
- 19:01It took a village
- 20:24The flip side
- 20:54We are nice
- 21:33What makes Rails special
- 22:14Enabling people
- 23:32Lessons Ive learned
- 28:43Learning
- 29:01Fighting Hero Culture
- 29:54Lessons Learned
- 32:00Resources
related talks
27mRails Orbiting the cloud: Mastering the scalability with Kubernetes
52mKeynote: Startups on Rails in 2024
54mKeynote: The Success of Ruby on Rails
52mKeynote: Startups on Rails in 2024
42mHow Rails fosters a diverse and competitive tech ecosystem in the era of big tech
48mWeb scale for the rest of us
32mBuilding HUGE Web Apps: Rails at 1,000,000 Lines of Code
41mForget Scaling: Focus on Performance