Agentic Coding the Responsible Way | Stanford Converge | Vasyl Rakivnenko
Автор: Vasyl Rakivnenko
Загружено: 2026-01-07
Просмотров: 25
Are you using AI coding agents like Cursor, Replit, or Claude Code? Learn how to avoid the hidden risks of agentic coding tools—including code hallucinations, fabricated benchmarks, and black-box codebases.
In this lightning talk, Vasyl Rakivnenko, AI Technical Lead at Stanford Legal Design Lab, shares real-world lessons from months of using agentic coding in production. Discover the 5 R's framework for responsible AI-assisted development that helped his team leverage research assistants to build legal tech applications in weeks instead of months.
What is discussed:
→ The difference between traditional coding, vibe coding, and agentic coding
→ Why your codebase becomes a black box with AI agents
→ Real examples of LLM code hallucinations and fabricated code
→ How AI agents can accidentally delete your database
→ The 5 R's: Refer, Restrict, Review, Revert, Retest
#agenticcoding #aicodingtools #llm #softwaredevelopment #aicoding #cursor #claudecode #codingbestpractices #stanforduniversity #legaltech #codehalluciations #vibecoding #aiassisteddevelopment #machinelearning #softwaredevelopment #programming #techtalks
Доступные форматы для скачивания:
Скачать видео mp4
-
Информация по загрузке:
