Leadership in AI-assisted engineering
summary
Fresh telemetry shows 95% telemetry adoption and regular AI users reporting ~6 hours saved per week, while a local daemon reports 50% of merged code AI-generated (up from 27%). PR size almost doubles, producing sludge and volatile change failure rates; median PR throughput rises ~7.76% (mean 13.1%) with a top performer at 70%. Practical levers include durable system prompts, temperature tuning, self‑hosted models and agents (DevGen, Ferry, incident agents), plus telemetry + experience sampling + high‑participation surveys for impact measurement.
AI-written summary · about
chapters (11)
- 0:00AI leadership challenges
- 1:13AI adoption and productivity
- 6:11Measuring AI impact and ROI
- 8:39AI success strategies
- 10:24Psychological safety and AI transparency
- 12:31Trust in AI outputs and guardrails
- 16:22AI and employee skill development
- 18:46Unblocking AI with compliance and creativity
- 19:27Measuring AI effectiveness
- 23:59AI in software development lifecycle
- 27:11AI readiness and leadership mandates
related talks
17mLeadership in AI-assisted engineering
44mSpeed Is Easy, Trust Is Hard: AI's Real Impact on Developer Productivity
22mThe AI Productivity Paradox in Engineering
27mSoftware engineering with GenAI
46mRecalibrate - How AI shakes up software team
27mThe state of AI in software development: Insights across 400+ organizations
01h 01mAre Your Engineering Teams Prepared for the GenAI Revolution
52mAre Your Engineering Teams Really Prepared for the GenAI Revolution?!