AI Access is not AI Ability A Cautionary Tale on the Importance of Dev
summary
Claim: having ChatGPT access is not enough — teams need AI access ≠ AI ability and measured training. Recent Meta study shows developers expect ~24% speedups but actual task time increases by 19%, while one experienced user sees 25% faster; typical reported gain is about four hours a week (≈10%). Practical program uses six 90‑minute workshops covering tokenization, embeddings, sampling, agents, RAG and measurement, partners with DX to track real engineering metrics and promises to report quantified results at CTO Craft Con London in November.
AI-written summary · about
chapters (9)
- 0:00Introduction by LJ
- 1:05AI access vs AI ability
- 5:00AI as a unique tool category
- 9:14False sense of mastery with AI
- 13:02AI tools and productivity illusion
- 16:34Developing true AI mastery
- 21:49Measuring real productivity gains
- 23:53Study on AI ability impact
- 25:59Closing and next steps
related talks
59mAI is Eating the Developer Career! Wait, what? Maybe it's the other way around?
51mAssessing AI's progress
01h 16mAI in Software Development: Boosting Efficiency
23mThe AI productivity paradox: Why feeling faster isn't the same as going faster
55mAI Assisted Engineering
46mMeasuring the Impact of AI on Developer Productivity
32mHow to Make AI Work for You: A Deep Dive Guide into ChatGPT for IT Professionals
41mPanel discussion: AI in code development – hype vs reality