SJTU-led Paper 'Context Engineering 2.0' Traces the Practice Back 20+ Years
mdancho84 · x · 2026-11-14
A new paper from SJTU, SII and GAIR — Context Engineering 2.0: The Context of Context Engineering — argues that context engineering isn't a fresh invention of the agent era: related practices go back over 20 years.
Key points:
- Frames the central question of how machines can better understand human situations and purposes;
- Traces an evolution from early HCI frameworks around primitive computers to today's HAI (human-agent interaction) paradigm, and potentially human-level or superhuman intelligence;
- Offers a systematic definition, a historical and conceptual landscape, and key design considerations, aiming to lay a conceptual foundation for systematic context engineering in AI systems.
More from coding & agent
- Upcoming DeepAgents/LangGraph manual unpacks how agent state flows to model calls — ghumare64 · 2026-11-14
- Dev: my AGENTS.md reuses one stack across all projects, so I stop thinking about tech choices — marclou · 2026-11-14
- T3 Code adds MCP connector so any agent can control it and swap models freely — koltregaskes · 2026-11-14
- One diagram to teach LangChain tool calling: how AIMessage and ToolMessage pair up — ghumare64 · 2026-11-14
- Data curation RL is just SWE envs: code pretraining data increasingly consists of Claude-built, less-sloppy repos — willcb · 2026-11-14
- Solo dev builds real-time conference translation platform for 70 languages with an agentic harness — evisoft · 2026-11-14