The original poster making this claim used a t-test to compare means (https://x.com/RobLynch99/status/1734278713762549970?s=20). Turns out the data is not normally distributed, making a t-test worthless (https://www.statology.org/t-test-assumptions/).

There might be other tests to do, but for this specific setup, the claim has been debunked.