There's really not enough information in the article to judge its validity... Some comparisons between devices, as but one example, could support the idea that

There's really not enough information in the article to judge its validity... Some comparisons between devices, as but one example, could support the idea that AMP affects conversion rate (AMP is only shown on mobile). As it is, I have my doubts that the mechanisms he speculates could have such an impact. URLs, for example, are really de-emphasised on mobile.