Plan to Throw the First One Away
summary
Argues for planning to toss the prototype and shows what that means: a working Android prototype built with Java, Python Flask backend and Odoo, an emulated economy/traffic/network simulator, NFC Java cards for signing, and a git-based ledger using Git history plus OpenPGP to prevent double-spend. Timeline is 13 months with a prototype due in ~4 months, team of ~9, four hires so far, and a surprising compression from 30,000 lines to 2,969 lines after rewriting in Kotlin—so the prototype works and then gets reimplemented in Go/Kotlin.
AI-written summary · about
chapters (30)
- 0:00Introduction
- 1:33Imposter Syndrome
- 2:44Secondguessing Myself
- 4:04ManMonth
- 5:09ManPower
- 7:34Scheduling
- 9:17Second System Effect
- 10:37Other Observations
- 12:13Background
- 12:38How did this get started
- 15:47Can you scale this
- 17:19Team
- 18:23Moving Parts
- 20:33Blockchains
- 21:17Throw the First One Away
- 21:47Pecky
- 23:51The Plan
- 27:48Principles
- 29:43Features
- 30:50Crypto
- 34:37Backend
- 35:53Simulation Engine
- 37:48Precommit
- 41:55Hiring
- 43:31Kotlin
- 45:12Backend stuff
- 46:14The simulator
- 47:16What we learned
- 48:17I wish I had figured out sooner
- 49:26Are there any questions
related talks
01h 10mWhy Projects Succeed: Lessons learned from the Android OS
35mWhy Projects Succeed: Lessons Learned from the Android OS
56mThe Best Simple System for Now
19mDon't Build to Last: Strategies for Designing and Productionising Experiments
36m"Engineering to embrace change" Or "How clumsy engineering can lead to powerful research – Part II"
46mWhy Projects Succeed: Lessons Learned from the Android OS
43mThe Engineering Playbook for Growing Android Teams
39mI don't have any technical debt