So Google I/O happened a couple weeks back — the 15th, if you're keeping track — and out of everything they threw at the wall during that keynote (new Nexus 4 storage tiers, the Galaxy S4 and HTC One "Google Play editions," a preview of this Android Studio thing that's supposed to replace Eclipse eventually), the only announcement I actually acted on was Play Music All Access. I signed up for the trial that same afternoon, mostly because I'm the kind of person who panics about missing a pricing window. If you lock in before June 30th you get it for $7.99 a month forever instead of the standard $9.99. I don't even know if that's a good deal yet. I just didn't want to find out in July that I'd missed it.
Two and a half weeks in, I've got opinions.
The pitch is simple enough: it merges Google's whole streaming catalog with whatever you've already uploaded to your Google Music library (up to 20,000 tracks, which covers basically my entire ripped-CD collection from college, RIP my old Dell that used to hold all that in iTunes). So the radio stations it builds pull from both places at once. Start a station off some random Spoon b-side I uploaded myself back in 2009 and it'll happily slot in tracks from the actual catalog around it. That part genuinely works better than I expected.
What doesn't work as well is the radio itself, honestly. I've been a Spotify subscriber for close to two years now and its algorithm just gets me better. All Access radio has a habit of circling back to the same six or seven "safe" songs per station way faster than I'd like, and after day four of every Wilco station devolving into the same three tracks I started just building playlists by hand again, which defeats half the point of paying for a discovery tool.
The interface is fine. It's the same Google Music web player skin, just with a little more stuff bolted onto it, and the Android app is smooth, no complaints there. But it's US-only right now, and I've got a friend up in Toronto who's been asking me to just email him the login, which, no, absolutely not, but also I get why he's asking, because there's no announced timeline for it going anywhere else. Spotify's been in like a dozen countries for a while now. Feels like a weirdly small-scale launch for a company that size.
I keep going back and forth on whether I'll actually pay for both this and Spotify past the trial. Realistically the answer is probably not, since paying twice for the same category of thing is exactly the kind of subscription creep I complain about on here constantly and then do anyway. But if I had to bet today, on May 31st, with the trial clock ticking, I think All Access loses. The catalog-merge trick is neat and I like that it doesn't shove a separate app on me, it's just a feature bolted onto something I already had installed. But Spotify's social stuff (seeing what people are listening to, the collaborative playlists my roommate and I have been building since last year) is sticky in a way a slightly-smarter radio station isn't.
Small unrelated complaint while I'm here: my apartment building switched trash pickup days without telling anyone and I found out because three different neighbors left bags out on the wrong morning and the raccoons had a field day. Nothing to do with any of this. Just been a week.
Anyway. If you got an I/O invite or know someone who did and got a free extended trial out of it, worth poking at for the upload-merge feature alone even if you end up canceling before the $7.99 window closes like I might. If you're starting fresh with no existing library to merge in, I don't think there's much reason to pick this over what you're probably already using.