Indeed kodos... ehm, kudos to Mike for this great article!
I really like most of the changes suggested here, and I can live with the rest.
I am not a huge fan of increasing ability scores in general, so I actually welcome the fact that this happens regularly only in Basic, while in Standard (where certainly most min-maxers will gravitate around) you will have to give up feats if you want to indiscriminately boost your ability scores.
I am not worried about even ability scores. Feats will follow the tradition of having odd-numbered ability score requirements, so increasing something from 14 to 15 will still have the benefit of qualifying for new feats (although part of me would prefer to see feats prerequisite go away, but this is hardly going to happen anyway). Second, I do find it fair enough that always boosting your primary score is not strictly better than spreading your bonus around.
The +10 to knowledge checks first gave me a wtf? reaction. That's a
huge number. But there is actually a hidden problem with knowledge checks, and that problem is that unless you restrict the check to "trained only", everybody should try knowledge checks on everything all the time. It is not the same as other skills really... the essential benefit of lore checks is
getting clues that can be used to your advantage, but the check has no penalty and doesn't even take time. Therefore you either say "trained only" (or something more complicated like "trained only when the DC is at least XX") or you give the trained characters a huge bonus so that it actually makes a difference.
The part about the DM adjusting checks DC depending on whether the group uses skills doesn't bother me much, but I can understand that a lot of gamers will hate the idea because it makes for an inconsistent world where jumping over a chasm is more difficult if the protagonists are better at it. In fact I think this should not be suggested at all... rather, the DMG should suggest to use wider chasms in your adventures
OTOH I think it's totally fine for one group to be better at jumping chasms and another group to be worse, just because the first uses skills and the second doesn't. This is not different from group #A in any edition using default rules vs group #B granting some extras (like more feats, more HP or more generous starting ability scores) which frankly happens the majority of the time.
If even this is not widely accepted, they can always toss in a simple "if the group does not use skills, every PC chooses one single ability score and get +N to all relevant checks".
Overall the article reveals changes that really improve the modularity of 5e!