How (not) to build a startup?
summary
Argues that startup risk comes from internal choices not frameworks: picks Magento then rewrite to avoid platform limits, chooses Vert.x + Kotlin → Spring Boot after performance and maintenance pain, builds and drops bespoke pieces (own auth server with vulnerabilities, three CI/CD rewrites), wastes weeks load-testing a product catalog to millions, stalls over linter and API crusades and unfinished 70–90% features. Emphasizes daily CI/CD, simple monitoring, hiring teammates who show up for incidents, and using constraints to force focus on product over technical ego.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:52Why startups differ
- 3:37Building the dream stack
- 8:09Avoiding enterprise mistakes
- 13:48Reality of ugly code
- 15:37Problems we didn't expect
- 21:41Copying big tech patterns
- 24:11Overengineering traps
- 28:08Technical ego pitfalls
- 31:22Doing too much at once
- 33:07B2B vs B2C misunderstanding
- 34:57Architecture for nothing
- 36:47What actually helped
- 43:31Importance of constraints
- 45:21Final lessons learned
- 47:07Q&A
related talks
48mIndustry Myth Busting
55mShiny objects are cool, but so is building products people use
35mUnderstanding Startup Pains and How To Overcome Them
25mEgoless Engineering
52mUnicorn's baby steps: what should you know as a 1st-time startup CTO
42mWinning as a Start-Up by Failing Fast
39mStop Focusing on Tech Debt
41mWhy Promising Nothing Delivers More and Planing Always Fails