Stanford prof's LLM peer-review workflow: argue back and forth, bucket issues by severity
anshulkundaje · x · 2026-10-29
Stanford professor Anshul Kundaje says his most satisfying and efficient results come from collaborating with LLMs over several rounds of argument rather than accepting first outputs. Since GPT6 Astra (Ultra) review quality stepped up, and prompting the model to clearly separate fatal flaws, big-picture issues, technical issues, nitpicks and wishlist items improves it further.
Related event: Stanford Researcher Sees Step-Change in GPT6 Astra's Peer Review Quality(2 posts)→
More from coding & agent
- Seroter's daily reading list: Gemini 4 Carbon, state of context engineering, agent-shaped judgment — rseroter · 2026-10-29
- Obsidian tip: the most useful search skill is simply naming your notes well — dSebastien · 2026-10-29
- PostGIS MCP Server lets you query spatial databases via natural language — modelcontextprotocol · 2026-10-29
- Harbor v0.25 ships anti-reward-hacking toolkit: tampering detection, cheating prevention for agent evals — ajratner · 2026-10-29
- Anthropic says no regression behind Claude Code usage complaints, blames heavy sessions — bcherny · 2026-10-29
- LlamaIndex launches Document Parsing Protocol: 140 rules to standardize OCR output for agents — llama_index · 2026-10-29