Keynote: Celebrating the 10 Years of Elixir
summary
Ten-year Elixir retrospective pivots to concrete plans: a funded PhD project builds saturated types (set-theoretic unions/intersections, singleton values) to infer types from pattern matching and guards; roadmap starts with structs then functions, but research may fail. Empirical claims get nuance — types don’t automatically cut testing or bugs, coverage can substitute simple type systems. Tooling work ships: Livebook desktop for interactive learning and Nx and Axon plus Explorer for GPU-backed tensors, neural nets and data frames, with early production uses.
AI-written summary · about
chapters (10)
- 0:0002:53 - Introduction
- 2:5412:53 - Why would we want Static types?
- 12:5417:41 - To type or not to type?
- 17:4219:14 - Rust's compiler + type system
- 19:1522:22 - Claim: Types lead to fewer tests
- 22:2329:27 - Claim: Code coverage catches all bugs that types do
- 29:2833:40 - Types as restrictions
- 33:4145:31 - Set-theoretic types
- 45:3249:03 - Developer and learning experience types
- 49:0452:36 - Machine learning
related talks
01h 03mKeynote panel discussion: The next 30 years
23mHow to teach Elixir to non-functional developers
39mFunctional Concepts in Elixir
16mKeynote
34mInterview with Louis Pilfold on Elixir Gleam
58mKeynote: Why Functional Programming Matters
01h 07mPipe forward : Using Elixir and F# together
01h 08mUnconference