As has been mentioned several times previously on this blog, a skill tax is some feature that a player feels compelled to take, not because they are interested in the skill itself, but because someone in the party should have the skill for the good of the group. Perception is such a skill, because you only need one character to be able to notice things (with perhaps one backup if the primary goes down). This is usually considered a bad thing, because the player that takes the skill is down one resource slot compared to all the other players. The same kind of dynamic can arise for things other than just skills (for example, the thief as trap finder and the cleric as healer).
I was thinking about this when reading the 5E playtest materials, which boil down skill-like checks and saving throws to ability scores. This is not a new thing, as people have been making ability checks (and even ability saving throws) for a long time. And I gather other systems (like Castles & Crusades) do something similar. But it does institutionalize and generalize a system that is already rather well understood and well liked.
What is the connection to skill taxes here? If wisdom is perception (one example from the playtest materials) then you also get all the other benefits of a high wisdom when taking it for the boost to perception (never mind for now all the problems with perception systems; my point is about features that players feel obligated to take). And, running this system using 3d6 in order (which will presumably be one of the character creation options in the final product) will help avoid excessive optimization potential.