AI won’t fix developer productivity (unless you fix context first)
summary
Claims that concrete improvements happen once teams engineer context not just code generation: a context engineering layer ties GitHub Slack Jira Confluence and PR history into searchable, ranked inputs so models like Claude and Cursor stop hallucinating and thrash less. Evidence includes a customer anecdote of five days to write 12 lines, Stack Overflow stats (53% slowed waiting, 63% spend up to two hours searching), and engineering patterns that actually get built: memories, agentic search with an LLM observer, and pre-ingest embeddings plus ranking and LLM-as-judge.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 2:09Traditional context gathering inefficiency
- 5:01AI tools amplify context problems
- 7:28What a context engine does
- 9:37Data and relationships in context engineering
- 12:51Memories and personalized context
- 14:58Working context examples
- 17:26Assembling context for AI agents
- 21:01Why context engines are vital
- 22:07Final recommendations
related talks
19mContext-Aware AI Dev : Next Wave of Engineering Productivity
27mAI for software development: A reality check
35mFrom Prompt Engineering to Context Engineering: Lessons from an AI Code Review Product
27mFrom AI Assistants to AI Coworkers: How Engineering Teams Ship Faster with Enterprise Context
22mThe AI Productivity Paradox in Engineering
32mAI and Engineering Closing the Gap Between Expectations and Reality 1
26mBeyond the code: Leading engineering teams in the GenAI era
13mMeet the Speakers