Next year, a new MN law will require tech firms to report results from experiments involving > 1000 people.
This could be huge for public knowledge. It could also pollute science, since fast-moving firms don't always have time to validate tests & didn't expect them to become public.
This validation is essential for reliable science. In a new post, I report on work to validate problems with 22% experiments in the Upworthy Archive & update the scientific record.