Keynote: Advent of Code, Behind the Scenes
summary
Creator recounts building Advent of Code from a 70-user prototype to 1.4 million users and 930k active solvers, surviving a midnight traffic spike by switching CGI → FastCGI, moving to AWS ALB scaling, then provisioning static capacity. Describes the three-tool workflow—input generator, part‑1 solver, part‑2 solver—to produce many distinct inputs per user, runtime calibration so Python/Ruby still win with the right algorithm, accessibility fixes, DDoS mitigation, and why leaderboards discourage LLM use.
AI-written summary · about
chapters (7)
- 0:00Introduction to Advent of Code
- 6:04Traffic estimation disaster
- 12:21Why people love Advent of Code
- 16:14How to build a good puzzle
- 24:37Puzzle creation process
- 33:49Growing pains and lessons learned
- 45:23Conclusion and impact
