Going Beyond Copilot with AI Agents
summary
Claims that code assistants miss 95% of developer time spent outside the IDE and that ~30% of accepted completions need fixes, then shows a concrete system: zencoder plugins for VS Code and JetBrains with rep gring and an agentic repair pipeline that compiles, runs tests, lints and iteratively fixes generated code. Demo asserts commit‑ready unit tests and a code‑improve agent, runs in the cloud with third‑party and in‑house models, is closed beta now and goes public end of October with free access until then.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:00Coding assistants today
- 3:52LLM assistant limits
- 9:13Developer time with AI agents
- 10:18What are AI agents
- 14:28Adding AI agents to workflow
- 16:57ZenCoder with AI agents
- 18:20Live demo: ZenCoder agents
- 22:56ZenCoder beta and future
- 23:37Q&A
- 23:58Agents for API docs
- 25:06Copilot Chat vs AI agents
- 26:51Test-driven development
- 28:16Handling dependencies
- 29:37Agents for architecture
- 31:25Context permanence
- 32:52IP and self-hosting
- 34:42Closing
related talks
06h 22mUsing AI to Enhance Developer Productivity
18mCoding Reinvented: How AI Agents Are Shaping the Future of Software Development
25mFrom AI Chaos to Production Ready: Using AI Agents to Improve Your Code Quality
30mHow AI Agents Will Change the Future of Programming
25mImproving developer happiness with AI
41mGoing beyond copilot with AI Agents
16mCoding Reinvented How AI Agents Are Shaping the Future of Software Development
24mGitHub Copilot Agents + Sonar | The Blueprint for Agentic Development