Human-Driven Intelligence: Where AI & Automation Fall Short
summary
Dives into concrete limits: the Dora report finds 67% report improved code quality while also showing that each 25% jump in AI use yields a 1.5% throughput drop and 7.2% stability decrease, suggesting AI speeds riskier, larger changes. Shows AI helping with code review, pipelines and even brainstorming via ChatGPT, yet struggling with exploratory testing that resists automation and with ethics — facial recognition fails for Black women and loan/insurance models embed bias. Stresses retraining models, keeping tests maintained, and defining quality before automating.
AI-written summary · about
chapters (10)
- 0:00Human-driven intelligence intro
- 1:06Automation hype cycle
- 2:33AI’s impact on code quality
- 5:08Defining quality beyond function
- 8:02Strategy before tools
- 9:53Automation can’t fix testing
- 12:08AI transforming jobs
- 14:41AI ethics and bias
- 17:12Smart AI and automation use
- 21:10Future of human-driven intelligence
related talks
18mHumans + AI: The Future of Software Development Teams
41mThe Ironies of AI²
39mAI-assisted development: Common pitfalls and smarter practices
40mThe Developer RenAIssance: The Humans Powering the Age of AI
41mPanel discussion: AI in code development – hype vs reality
26mAl and Software: The Good, The Bad, The Scary
56mPANEL: Unleashing the Future: The Impact of Gen AI on Development Productivity
34mAI and Code Quality: Building a Synergy with Human Intelligence