Best Simple System for Now
summary
Defines the Best Simple System for Now and shows concrete rules: start by deploying one VM via rsync and avoid Kubernetes until you actually need scale; build a four‑line Python stub to convert HTTP→AMQP instead of overengineering a gateway; invokes TimSort and Python’s import this, Gall’s law, Kakaku and Advent of Code forcing algorithm changes; recounts HP‑UX memory mistakes; cites WhatsApp (500M users, 13 engineers), SQLite (≈600× test volume) and economic tools like Cost of Delay and the trust me once experiment.
AI-written summary · about
chapters (7)
- 0:00Intro
- 3:44Agenda
- 5:05Defining the best simple system for now
- 26:17Do not try this at home
- 34:41Why no one else is doing this
- 40:09And then what?
- 43:00Outro
related talks
56mThe Best Simple System for Now
22mSimplicity: The Key to Software Success
01h 07mHow Simple Is "As Simple As Possible"?
01h 17mWhen Less Is More: Designing & Decoding Complex Distributed Systems
26mThe Power of Simplicity: Crafting Durable Software Systems
50mBuilding an Application Platform: Lessons from
01h 02mTools for simplified large scale systems
49mHypergrowth scaling made simple