Al and Software: The Good, The Bad, The Scary
summary
Claims that AI already speeds coding with autocompletion, unit-test and docs generation, and that the next big gain is post-commit pipeline diagnostics that explain CI/CD failures and propose fixes. Demonstrates use of third‑party models like ChatGPT, Claude 2 and Meta Code Llama while arguing for an abstraction layer to swap models and evaluate quality. Emphasizes security guardrails — model choice, prompt hygiene, data handling and zero data retention contracts — and measures impact via DORA metrics plus qualitative developer satisfaction.
AI-written summary · about
chapters (13)
- 0:00Welcome
- 0:42AI in software delivery
- 3:11AI beyond code generation
- 5:57AI’s impact on stakeholders
- 8:04Security and privacy with AI
- 10:10Tech stacks for AI
- 12:58Measuring AI’s business impact
- 16:50Skills for AI adoption
- 19:17AI versus past tech trends
- 20:21Fears around AI and automation
- 22:08Exciting AI use cases
- 23:55Toil versus technical debt
- 24:37Closing thoughts
related talks
13mMeet the Speakers
51mCode Smarter, Not Harder: AI-Powered Dev Hacks for All
01h 16mAI in Software Development: Boosting Efficiency
01h 28mFuture of Software Programming with AI Bots at Hand
45mThe AI Revolution in Software Development: From Code Writing to Quality Assurance
49mMagic of Automation and Everyday Chores
46mPanel Discussion: AI's Impact on Developers
47mAI and the Future of Software Development