Sources
Selected reports, research, surveys, and official product/documentation links used to ground this wiki.
Scope check: sources were selected to support the central question: how AI will change software engineering work, productivity, quality systems, team structure, roles, and skills over roughly the next decade.
Research and evidence
- Stack Overflow 2025 AI survey
- Stack Overflow 2024 Developer Survey
- DORA 2025 report
- Google DORA 2025 announcement
- GitHub Copilot productivity experiment
- GitHub Copilot code quality research
- METR 2025 experienced developer RCT
- McKinsey developer productivity with generative AI
- Uplevel Data Labs Copilot study
- BLS AI impacts in employment projections
- BLS software developers outlook
- Gartner GenAI upskilling press release
- GitClear AI code quality report
- GitHub Copilot official
- Cursor official
- Claude Code official
- OpenAI Codex official
- Amazon Q Developer official
- Google Gemini Code Assist official
- JetBrains AI official
- Sourcegraph Cody official
- Windsurf official
- GitLab Duo official
- Snyk DeepCode AI official
- Sonar AI CodeFix official
- CodeRabbit official
- Qodo official
- Diffblue official
- mabl official
- Applitools official
- Launchable predictive test selection
- Datadog Watchdog
- Dynatrace Davis AI
- Braintrust official
Testing, verification, and reliability
Added 2026-05-20 to support the Testing in the Age of AI page.
- GitHub Spec Kit — spec-driven development toolkit
- AWS Kiro — agentic spec-driven IDE
- Martin Fowler — Exploring Generative AI
- Systems Correctness Practices at Amazon Web Services (CACM)
- EARS — Easy Approach to Requirements Syntax
- Parse, don't validate (Alexis King)
- Semgrep — static analysis
- CodeQL — semantic code analysis
- Kani — Rust bounded model checker
- TLA+ — specification language
- Hypothesis — property-based testing
- Stryker — mutation testing
- OSS-Fuzz — continuous fuzzing
- Antithesis — deterministic simulation testing
- What's the big deal about deterministic simulation testing?
- Pact — consumer-driven contract testing
- Promptfoo — LLM evals and red-teaming
- Inspect — UK AI Security Institute eval framework
- OpenTelemetry — GenAI semantic conventions
- Argo Rollouts — progressive delivery
- OpenFeature — vendor-neutral feature flags
Interpretation notes
Balance: the evidence supports real productivity and quality potential, but the effect is uneven. Controlled demos, enterprise surveys, and mature-codebase field experiments answer different questions. Treat this wiki as a synthesis, not a single-number forecast.