This is Fine! A podcast about resilience engineering and software The 2025 DORA Report w/special guest Fred Hebert
12 snips
Mar 12, 2026 Fred Hebert, a Staff SRE and Lund student known for work on SLOs, error budgets, and the Law of Stretched Systems, discusses the 2025 DORA Report. He unpacks why the report reframes around AI-assisted development. They explore AI adoption models, survey limits, platform vs AI impacts, cognitive load and burnout, and how new capacity can be reabsorbed by organizational demands.
AI Snips
Chapters
Transcript
Episode notes
Yearly Model Changes Make Longitudinal Claims Tricky
- DORA's model changes year-to-year, adding and removing constructs like transformative leadership, which complicates longitudinal interpretation.
- Researchers iteratively fold or drop items when they no longer fit statistically, leaving gaps in public rationale.
Invest In Platform Engineering Before Betting Big On AI
- Prioritize platform engineering over chasing AI hype because platforms yield larger, clearer organizational benefits in the DORA model.
- Fred notes platforms act as multipliers and in the report often outperform AI's measured positive effects.
Distro Maintainers Finally Could Run Huge Test Suites
- Clint recounts Linux distro maintainers using heavy test coverage enabled by automation to catch regressions earlier.
- He contrasts past funding barriers for mass testing with present AI/platform-driven ability to run thousands more tests per package.