Vivy: Fluorite Eye's Song starts as it means to go on: a brutally direct portrayal of an AI apocalypse, with androids hunting down humans in a Disneyworld-like theme park, with a smash cut to the same theme park 100 years before where the titular early-model singing android is having a tough time pulling in an audience.
Right off the bat, this shows a few of the key aspects of the series:
Whether you like this show or not is probably going to come down to how you feel about Matsumoto, the time-traveling, hyperverbal AI who is responsible for convincing Vivy to do what needs to be done, and also for barfing up the infodumps that are responsible for conveying the details of both the world and the chaos shown in the beginning of the show. Matsumoto's voice actor, Jun Fukuyama, does a brilliant job of bringing this motormouthed cube to life, but for the first few episodes I was constantly pausing and rewinding so that I could catch all that he was saying (and not lose the thread of what was happening on screen). He also doesn't start as a particularly sympathetic character, being the one responsible for forestalling an apocalypse without disrupting history to the point that the data he carries about the sequence of events being thrown too far off. He can't afford to be too sentimental, and he can't afford to allow Vivy to be too sentimental either.
Vivy herself is a familiar kind of cool beauty archetype, and since she's the protagonist, it matters how her character works out, but while she does end up being the linchpin around which the resolution of the story revolves, she spends the first half of the show being acted upon more than acting. That said, she is a great action protagonist – there are dozens of cuts throughout the show of her looking and acting very badassed, and the animation and character design of her as a character is nearly flawless. Despite being nonhuman, she reminds me of nobody so much as Major Kusanagi in the original Ghost in the Shell anime, especially in the way in which she occasionally goes past the physical limits of her android body in very inhuman ways. She has some of the Major's reserve as well, although in the end she proves to be a more human (?) character.
I talk about the characters so much because ultimately the show is trying to answer a question: can artificial intelligences and humans coexist, and can you give autonomous machine intelligences a reason for living that will be similar enough to a human's purpose-driven life to keep them going? The fact that the show doesn't come to a really satisfying answer to this question didn't disappoint me too much, because I'm not sure there is a satisfying answer, especially one that can be fit into 13 24-minute episodes. It works just fine to keep driving the story forward and to setting real emotional stakes for why the characters are doing what they're doing.
This also helps paper over some of the unavoidable flaws that come from introducing time travel into the narrative (only Primer has ever really gotten time travel right, and even it has one or two holes), as well as (for me) carrying it through a few places where the narrative sags towards the end of the season. I give this show major credit for ditching the now-hackneyed framing of trying to install some kind of external moral governor (like Asimov's Three Laws) into the AIs and instead focusing on the consequences of making every AI devoted to carrying out a single mission, and how much room for latitude the AIs, as sentient beings, have in interpreting their mission to give themselves room for action. I also appreciate that there wasn't endless handwringing around free will, although this is probably at least partially a consequence of the show not having the runtime for that.
I don't want to go into spoilers, so I will just say that yes, the last arc has real problems (there was a lot of Rule of Cool in the last few episodes and a few scenes that were supposed to feel ominous made me laugh instead), and also that I think the show does come to a satisfying and meaningful resolution at the end. I spent about half the last episode in tears. If I had any complaint, it is that I think this could have (and probably should have) been a 24-episode anime, with the same pacing, but more time spent on some of the intricacies of the development of AI over the 100-year span of the show. The show feels a little rushed towards the end, and some parts of the end of the plot were foreshadowed earlier in the season, but with the pace of the show being so fast, they ended up feeling more like deux ex machina than they should have.
On a technical level, Studio WIT turned in a near-flawless performance. The animation is excellent throughout, the various openings and endings are great, the music is both good on its own and very well-integrated into the show, and the character designers and animators do an excellent job of differentiating visually similar characters and sometimes multiple identities for the same character. There were a lot of shows this season with superb animation (SSSS.Dynazenon, ODDTAXI, Super Cub) and Vivy more than held its own.
The upshot of all of this is that this is an extremely good show, I'm going to miss watching it every week, and any fan of science fiction and classic science fiction anime should check it out. I look forward to what Studio WIT, Tappei Nagatsuki, and Eiji Umehara do next.
67.5 out of 72 users liked this review
