I've read the original paper by Jaeggi et. al., one paper replicating it successfully and one paper proving the results transfer to different tasks. At this point wasn't it logically correct to say that's scientifically proven? Now we know there are papers which prove those proofs questionable. That's what I mean.

If only we had a bullet-proof test to measure fluid intelligence (classic IQ tests have quite a well-known number of problems, solving math problems and puzzles is known to be much more of a skill, working memory is working memory - not intelligence itself) objectively...