Many users dismissed the Claude analysis finding no cheese-heart disease link after wealth controls as LLM psychosis or for relying on weak observational data instead of RCTs.
Based on 2 visible X reactions from 3 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@Altimor there are over a dozen RCTs on saturated fat and cardiovascular events, here’s a literature review: https://pubmed.ncbi.nlm.nih.gov/32827219/?utm_source=chatgpt.com it shows less saturated fat reduced cardiovascular events; why not use that instead of making weak associations on observational data
my version of llm psychosis https://x.com/Altimor/status/2075338030299426832/photo/1
Asked Claude to look into the correlation between cheese consumption and heart disease across countries, and it found ~0 correlation (after controls). Surprising given the current consensus on cheese / saturated fats as being very bad for cardiovascular health — you'd have expected a strong positive correlation.
Many users dismissed the Claude analysis finding no cheese-heart disease link after wealth controls as LLM psychosis or for relying on weak observational data instead of RCTs.
Based on 2 visible X reactions from 3 accounts; directional sample.
Ask a question below.
Published answers will appear here.