Expert View: Knowledge Graphs Bring Massive Trouble to AI Agents but Remain Indispensable Short-Term
tokenbender · x · 2026-07-28
Data and AI expert @tokenbender expressed strong frustration with knowledge graphs (KGs) while discussing the Kimi K3 paper. He pointed out that KGs bring significant engineering headaches in entity extraction, event creation, and updates.
However, he admitted that at least in the short term, KGs remain an unavoidable component in both agent harnesses and model training. He previously highlighted a key takeaway from the post-training section of the K3 paper: agents possess a self-evolving knowledge graph capable of web-scale exploration.
Related event: K3 Paper Breakdown: Self-Learned Reasoning and KG Debates(2 posts)→
More from coding & agent
- DeskPilot: open-source native Python desktop client for local LLMs with MCP and sandboxed tools — poofph · 2026-09-23
- Opus 5.5 turns a single image into a Three.js game menu in one simple prompt — majidmanzarpour · 2026-09-23
- Yale PhD student open-sources his paper figure scripts, packaged as a Skill for Claude Code and Cursor — burny_tech · 2026-09-23
- Dev uses Opus 5.5 with Lean to formally verify Claude Agent SDK, yielding 16 bug-fix PRs — bcherny · 2026-09-23
- Chrome flags compromised passwords; user wants a Gemini agent to reset 50+ of them — doooyle · 2026-09-23
- REQL gives coding agents persistent working memory, pipeline views and multi-agent coordination — cl0wnfire · 2026-09-23