MED01 · P2 locked 2026-10-14 · P3 next
Historical Evidence Landscape of Medical Cannabis
What did medical cannabis and cannabinoid research consist of, decade by decade, and what did it and did it not establish?
Not medical advice. These are research records about evidence quality, written for education. Talk to a clinician about any treatment decision.
Who did what
| Role | System | Model, as recorded |
|---|---|---|
| Researcher · P1 | ChatGPT | not recorded |
| Researcher · P1 | Gemini | not recorded |
| Researcher · P2 | ChatGPT | GPT-5.6 Sol |
| Researcher · P2 | Gemini | Gemini Spark |
| Editor and verifier | Claude | not recorded |
| Cross-auditor | ChatGPT | not recorded |
| Cross-auditor | Gemini | not recorded |
Brief and prompts
- ClaudeBrief3.8k words
- ClaudeCorrection prompts926 words
- ClaudeCross audit prompts1.4k words
- ClaudeKickoff prompts989 words
- ClaudeBrief · P24k words
- ClaudeCorrection prompts · P21.4k words
- ClaudeCross audit prompts · P22.8k words
- ClaudeKickoff prompts · P21k words
Independent research (raw, as received)
- ChatGPTChatGPT correction 1 · P1348 words
- ChatGPTChatGPT pass · P19.9k words
- GeminiGemini correction 1 · P12.2k words
- GeminiGemini pass · P111.9k words
- ChatGPTChatGPT correction 1 · P2354 words
- ChatGPTChatGPT stage 1 · P210.1k words
- ChatGPTChatGPT stage 1 disclosure · P2776 words
- ChatGPTChatGPT stage 2 · P210.2k words
- GeminiGemini correction 1 · P22.9k words
- GeminiGemini stage 1 · P29.9k words
- GeminiGemini stage 1 disclosure · P21.5k words
- GeminiGemini stage 1 rerun · P211.8k words
- GeminiGemini stage 2 · P26.7k words
Verification, critique and corrections
- ClaudeCorrections2.4k words
- ClaudeEvidence extraction491 words
- ClaudeVerification & critique2.3k words
- ClaudeVerification log3.2k words
- ClaudeVerification & critique · P25k words
Ledger and synthesis
- ClaudeClaim ledger16.5k words
- ClaudeClaim ledger, plain language2.1k words
- ClaudeSynthesis3.7k words
- ClaudeSynthesis, plain language1.4k words
- ClaudeClaim ledger, plain language · P24.1k words
- ClaudeSynthesis · P24.9k words
- ClaudeSynthesis, plain language · P22.1k words
Cross-model audit
- ChatGPTChatGPT secondary audit1k words
- ClaudeClaude secondary audit537 words
- GeminiGemini secondary audit6.3k words
- GeminiGemini secondary audit987 words
- ChatGPTChatGPT secondary audit · P211.7k words
- ClaudeClaude secondary audit dispositions · P21.4k words
- ClaudeClaude secondary audit verification · P21.4k words
- GeminiGemini secondary audit · P29.5k words
P1 audit: ChatGPT conditional pass that caught two verifier errors; Gemini pass, with a finding logged on the auditor for non-verbatim quotations. P2 audit: ChatGPT accepted; Gemini rejected (quotations not found in the sources), replaced by a source-level spot check by the research director, 8 of 8 matched.