A love letter to legacy
summary
Legacy systems keep money flowing and often arrive as a ten‑year‑old zip file with source and no trustworthy tests, so the claim is that understanding and incrementally renovating beats big‑bang rewrites (most rewrites fail long before completion). The talk shows using GenAI tooling to read docs, find security candidates and generate glue code, describes a network‑isolated Copilot loop for safe research, and warns that AI can create instant legacy if governance and tests are missing.
AI-written summary · about
chapters (13)
- 0:00Introduction and speaker background
- 2:37What is legacy software?
- 6:58Legacy systems in real life
- 10:50Complexity turns into legacy
- 13:20Maintenance death spiral
- 16:09AI and legacy code interaction
- 22:35Legacy systems as cathedral renovation
- 26:26Security and AI limitations
- 30:25Working with AI on legacy systems
- 33:15Avoiding big bang rewrites
- 36:10Practical AI benefits today
- 39:25Plan B for AI dependency
- 40:30Q&A and closing remarks
