When to be Weird
summary
Argues specific criteria for when to adopt new tech: low cost + high benefit wins, high cost + low benefit loses. Cites concrete wins — typed programming (LSP-driven productivity, TypeScript and Rust adoption), immutability by default, sum types and compiler-driven refactors that catch missing cases, macros/auto-deriving eliminating class-of-bugs. Flags niche but useful ideas like STM and green threads that require language/runtime support and so remain limited. Uses Docker as an example of the right-sized change that succeeded where radical rewrites failed.
AI-written summary · about
chapters (14)
- 0:00Weird in software
- 3:15Being weird in tech
- 6:23Ideas in software dev
- 9:11Functional programming weird?
- 12:19Typed programming win
- 16:56Immutability by default
- 18:43Sum types adoption
- 21:32Sum types not fully won
- 23:00Pattern matching & auto-deriving
- 25:03Cool niche features
- 27:33Rust macros & meta-programming
- 28:57Type classes & traits
- 29:42Docker won over containers
- 30:44Q&A
related talks
52mWorse Is Better, for Better or for Worse
39mThings that Matter
32mMaking Sotware Better, Despite Having Users
01h 02mWorse Is Better, for Better or for Worse
58mSoftware Engineering Languages
45mDynamic vs Static Having a Discussion without Sounding Like a Lunatic
22mA Plea for Boring Tech
09mInterview with