Everything's still better than being silent. Luckily Valve's slowly breaking that habit, but it needs to happen quicker. Valve's been silent for too long.
Community Management Science Part II:
Back in the day, they didn't need to talk a whole lot. It was kind of a normal thing in the AAA industry to be quiet, and let your marketing and PR guys do the talking. Now, in the age of Youtube DevLogs, Twitch Streams and generally a lot of openness from indie devs, it's becoming a lot more common, and a lot more companies are being pressured into doing it. (where it is being more interactive with the community, being more open with the community via twitter/twitch/youtube, etc).
Now, they don't necessarily have to do any of this, there are still plenty of companies who don't, and they get along fine. Should Valve do it? Thats totally up in the air. Again, this goes back to the whole "Who would read it, and would it be worth it?" type of thing. If they write a post and it only reaches a third of the population, is that worth the time? They're already super-duper busy, is it worth writing up a monthly thing "Hey, this is what we've done"? With Valve and their work structure/flow (or game dev in general) it's possible that they spend a month on something and realize they don't want to go for it. Like CC or Asteroid, they could start something early, because they have the time/resources to do so, before getting hit by another wave of work and they'll just sit there for a while. So what would they write about? As I understand it, Roadmaps are a thing at Valve, but they're deadlineless.
They could say "We're working on this new map this quarter!" then not use the map for 2 years. Saying "We're doing this right now for this bigger thing we don't want to announce" has a high possibility of working against them at the current moment/state of 'video game community mentalities' where everyone thinks that making a quality level/game/update takes only a couple months with 5 people. To do open development right/proper community engagement and as I've seen, it requires a LOT of work. For Valves 3 most active games, CSGO, TF2 and DotA2, you would need a large team of community managers to engage, plan out engagement events/activities and most importantly stay in deep touch with each of the games development cycles. ONTOP OF THAT ALL, they need to start to embed themselves in the communities of the games. If there are any fan-communities (like ours, for example) they would need to reach out there too. It's a LOT of work.
Additionally, on top of it all since there are already a lot of communities (in TF2 for instance) who do a LOT of their own community interaction/engagement events (ie: stuff like the contest we have, the 72hr Jam, April fools, etc) does Valve REALLY need to do the same thing that we're already doing with a lot of success? Them just posting about it on their blog is already more than enough. Could they do better, oh yea, totally. They still could start doing more engagement, and if done correctly, would be helpful, and wouldn't hurt them, but thats just based on what I know. I don't know the TFTeams workflow, nor their team structure or whatever, so all that I'm saying could be right or wrong... Won't know more unless someone like Jill, Dave or Eric come in and flat out tell us.
I could probably go on and on about this for another 6 pages, but I'll leave it at that. I don't think I really resolved my argument, but I think I got a point across.
Sorry for the essays. I should probably just write this all up or make a video series or something.
TL;DR: Sure, they could, but would it be worth it? For TF2, could it be done in such a way that isn't harmful? The answer is yes, but then it goes back to the first question: Would it be worth it?