@emollick
Here is the key table showing the 30 day effects (which they misreport in the paper) & which apparently uses a completely different test, but the ChatGPT group remains ahead. The errors shouldn’t fill you with confidence about the study, though. https://t.co/hqocnAsF38