Become a 10x developer by harnessing the power of Generative AI
summary
Generative AI coding assistants show concrete limits and concrete benefits: on 164 benchmark problems ChatGPT solves 65% correctly, GitHub Copilot 46% and Amazon CodeWhisperer 31%, failing mostly on long complex prompts, so use short focused prompts, descriptive method names and embedded unit tests; refactor benchmarks over 100,000 runs show PaLM/Google models 37%, GPT‑3.5 30%, Code Llama 18% so always review refactors and keep humans in the loop; duplication rises across 153 million lines, developers spend ~70% time reading code, and AI helps prototype, generate tests and find performance wins (quick sort ~95% faster on 1,000 elements).
AI-written summary · about
chapters (14)
- 0:00🚀 Build Stuff Conference – Get Your Tickets!
- 0:21Introduction: The Rise of AI Coding Assistants
- 3:10Weakness #1: Functionally Incorrect Code
- 6:05Understanding AI’s Performance on Complex Prompts
- 9:20Weakness #2: Issues with Refactoring Existing Code
- 12:45Weakness #3: The Challenge of Code Duplication
- 16:30How AI Coding Assistants Mimic Short-Term Contractors
- 19:40Human Ingenuity: How Developers Can Safely Use AI Tools
- 23:10The Importance of Writing Descriptive Code Prompts
- 26:00The Real Struggle: Reading Code, Not Writing It
- 29:15The Strengths of AI Coding Assistants in Prototyping
- 32:40Exploring New Technologies Faster with AI Tools
- 36:10Understanding and Improving Code Performance
- 40:05Generating Unit Tests with AI for Legacy Code
related talks
01h 02mCan you become a 10x developer using generative AI?
36mUnleashing Generative AI: Improving Developer Productivity in SRE
57mAccelerate Your Developer Productivity with AI Embrace the Future Now!
01h 02mPANEL - Unleashing the Future: The Impact of Gen AI on Development Productivity
41mWeb Development in the Times of AI
33mAI assisted software delivery
01h 16mAI in Software Development: Boosting Efficiency
01h 28mFuture of Software Programming with AI Bots at Hand