I think they are both important in predicting medical school success, however at the same time... typically (not always, I agree school difficulties vary) you would consider a 4.0 student over the course of 4 years pretty intelligent/knows how to study/apply information/etc, and would expect them to do well to some degree. The MCAT on the other hand, a standardized test that you take one day in your life - how can you compare a test asking you ambiguous questions about 16th century artwork to 4 years of good grades? Especially when the difference between a 6 and 10 in VR can be as little as 7 questions total? In the end ONE passage on the MCAT can determine whether you will have to retake your MCAT/get accepted - which I don't think should be the case (especially if that 1 passage is some ridiculous/pointless VR passage) when compared to 4 years of ~120 hours of coursework.
Never understood how an MCAT score has a huge correlation with board scores when 1/3 of the test is VR, something you'll never see again in your life.
(PS. I did well on my MCAT, I got very lucky on VR, I still hate VR more than anything.)
Regardless, according to AAMC, so far GPA wins for a reason:
Someone with a 3.7 GPA only needs around a 31 to have a ~70% chance of getting in.
Someone with a 3.6 GPA needs roughly a 34 to have a ~70% chance of getting in.