Both statements read like he's talking about past changes and development strategies, not Valve looking to the future.
He's talking about when a group of people get together to make a game, the group will value one talent over another so it's best for every employee to have multiple talents instead of just one. Because you never know if the group will need a multiplayer developer or a single player level designer.
While I don't totally disagree with you, I get a different read from the full article.
Basically the question beforehand is him answering that "things change very quickly in the video gameindustry so we don't give people titles that make them overly attached to what they do now".
This example then follows up talking about how consumers have moved toward multiplayer, so these days being good at Half-Life level design is no where nearly as useful as being good at making social multiplayer titles.
Since the person isn't "director of campaign level design", Gabe feels they're much less likely to argue against this move since it doesn't make their current job look irrelevant if Valve sits down and makes a social multiplayer game.
Or in other words he's saying being adaptable and either multi-talented or capable of being multi-talented is very useful, and gives Valve's shift toward social multiplayer over campaign games while retaining the same staff as an example of why.
He certainly never calls the level design capability a useless talent, as Valve has clearly made some campaign titles recently, just that the multiplayer abilities are notably more useful, which suggests where their emphasis is. They also suggest that emphasis is driven by their consumers, which looking at the Steam Stats page, suggests they've had success.
