"seriously"? Scott uses humor to present some rather unsettling facts, and while I personally find his mannerism crass... the assertion remains plausible.

>doesn't replicate

Which studies are you referring to exactly?

There was some fair criticism of confounding variables in Universe 25 (the first page google result most people cite), but the follow up Rat Utopia eventually compensated for the carrying capacity arguments.

There were other more recent experiments where lab assistants would play with the rodents during cage cleaning, and that stimulus was proven to be enough to keep test subjects cognitively functional.

https://www.youtube.com/watch?v=aNSHZG9blQQ

Best of luck, =3