All You Need Is All You Need Is Kill
I wake up. I write a book. It’s good, but there’s some issues with the prose and characterization – what if we got a big-budget cast to give it that grandiose Hollywood feel? I wake up. There's a movie now. Tom Cruise is here. The American secondary market seems to like it, but the nuance of the original was lost – what if we did a Japanese production, and switched up the protagonist? I wake up. This time, there’s an anime…
All You Need Is Kill (2025)
Well, this is maybe my least favorite Studio 4°C joint, but still a solid movie.
Craft-wise, it’s good. The colorwork is incredible, and there’s lots of beautiful environment art. The sound design was great; I particularly dug the whirly tube used to indicate wind whistling through Darol’s core, and the bassy hits of the axes in the final battle were a nice contrast to the typical treble shing of movie swordfights.
The 3D here I actually don’t mind, and surprisingly, it’s at its best in the slower, more character-driven scenes; they did a great job of capturing the rough, oblong charm of 4°C’s chara-designs without them feeling too tamed by the consistency of 3D modeling. The character animation also shines in capturing Rita’s slow, steady increase in dexterity. It’s frustrating to watch the characters pilot the jackets at first, their movements sluggish and clumsy, which only makes it all the more satisfying a dozen loops later when Rita and Keiji lithely prowl the battlefield like predators on their suits’ digitigrade legs.
The action boarding is where it lost me, and, well, there’s a lot of action in this movie. It’s constant and frenetic in a way where your eyes never have a good spot to rest – there’s not enough strong, lingering keyframes to ground the action. A lot of the fights use this zoomed-in, first-person view where all you can see is the overwhelming radial symmetry of a Darol and some splatter effects, which doesn’t help, and even in the fights that don’t, the camera is still going crazy, with the animators presumably drunk on the power and freedom CG has given them.
Script-wise, I found it to be a smart adaptation that stumbled at the finish line…
…but after watching, I wanted to make sure that statement was qualified, and being the person that I am, I went back to the other versions of the story to make sure my opinion was properly informed.
All You Need Is Kill (2004 LN)
oops, this pic shows how long it took me to write this
Well, reading this again for the first time in over a decade certainly made me less precious about its adaptations straying from the source material, because honestly? It’s messy as hell, and there’s quite a few elements that could use a touchup. The world has enough military scifi as it is, and that goes double for military scifi that’s played this straight. Our protagonist is a generic, sexist meathead whose internal monologue is composed entirely of cliches and hamfisted similes, and the rest of the book’s prose isn’t much better.
—
“The squad was full of badasses who had been through Hell and back so many times they were on a first-name basis with the Devil. When shit started flying, they started telling jokes. Not the kind of joke you told your mother over dinner, either.”
—
that’s a real quote from the book. jesus christ.
I think the success of All You Need is Kill might have been a right time, right place sorta thing. Stories about time loops are perennially popular, of course, and Japan is no exception. Before this era, time loop stories tended to be character studies: tales about finding love, about the desire to fix past mistakes, or about using repetition to overcome one’s personal failings. All You Need is Kill was a harbinger of a new breed: the time loop as a power fantasy, where even the most hapless schmuck can level up and grind XP, videogame-style, to become exceptional in a world where they were previously a nobody. This conceit is everywhere in anime these days: Re: Zero, Solo Leveling, Reincarnated as a Slime… All You Need is Kill just arrived early, and the concept resonated with people.
Interestingly enough, though, after re-reading it, All You Need is Kill doesn’t actually show all that much repetition. After Keiji finally answers the call to action, we quickly cut to dozens of loops later, where he’s already achieved prowess in battle and figured out quite a bit about how to combat Mimics. For a story whose legacy and adaptations focus so much on the leveling aspect, there’s shockingly little sense of incremental progress in the original book itself. Mind you, showing that sort of repetition is inherently tricky to pull off in a book; it’s way easier to do a snappy, well-paced montage sequence in a visual medium.
—
“Anyone who has spent any time in a kitchen is familiar with the inherent risks, such as they are, and know what can be done safely and what can’t. Never throw water on an oil fire, keep the knife pointed away from your carotid artery, don’t use rat poison when the recipe calls for parmesan cheese.”
—
what are we doing here?
For all my issues with its prose, All You Need is Kill does land its emotional beats, and the ending is its saving grace. Once you put all the sci-fi aside, when you ignore the lizard brain satisfaction of watching someone get stronger, the emotional meat of the story is about human connection. It’s about the catharsis Keiji experiences when he finally finds someone who understands what he’s going through, and about the relief he feels with Rita at his side, knowing he doesn’t have to do it all alone any longer. And that’s why the book’s ending – where Keiji’s forced to kill Rita in order to exit the time loop – hits so damn hard.
What makes this ending work is that it’s not an edgy grimdark anime twist just for the sake of it – it’s tied in rather well with the story’s themes and structure. Unlike future works in the genre, Keiji’s steadily improving combat skills aren’t just a mindless power fantasy – every step he takes to become a more efficient killer erodes his humanity. The power of trial & error sets him apart not only in increased skill and results, but in an atrophied ability to empathize and form meaningful relationships. As foreshadowed by the rank-and-file soldiers’ perception of the Full Metal Bitch, Keiji finds that becoming the world’s savior does not result in happiness and fulfillment, but infamy and scorn. To become the ultimate warrior is to become Atlas, bearing the Earth upon his shoulders, or Jesus on the cross – a sacrificial figure that forsakes their own life in service of mankind – and killing Rita is the final act that pushes him across that tragic threshold.
For all the bullshit you have to deal with to get to this conclusion, it is, at least, a strong one. If there’s anything in this story worth iterating on and retreading, it’s making that ending land properly.
Now this is a truly awful adaptation. It’s not that they changed the details – Rita’s axe is now a sword, Farrell is a strait-laced religious nut instead of a brotherly training partner, Shasta’s been changed from a shy cutie to a bearded physics major. It’s not even about the fact that they abandon the jackets halfway through, so Tom Cruise can run around in a black jumpsuit with an assault rifle like every other Tom Cruise movie*. No, the problem is that Edge of Tomorrow fundamentally misunderstands what’s interesting about this story.
*It is sort of about the fact that the very first time we see Rita in combat, it’s her taking a stupid hit and getting laid the fuck out. That’s not a minor aesthetic detail, that’s full-on character assassination.
When you put aside all the set dressing, the pillars of All You Need is Kill’s story are threefold:
- Pillar One: the satisfaction of self-improvement through repetition
- Pillar Two: the Sisyphean torment of being trapped alone in an unending nightmare, and the emotional catharsis of finally meeting someone else who understands what you’re going through
- Pillar Three: the gut-wrenching pain of having to fucking kill that person
Edge of Tomorrow manages to bungle all three of these.
Instead of keeping the source material’s horror angle of slowly learning to fight back in between being violently killed by monsters day in and day out, Tom Cruise spends his time inside in a gym, training the way anyone else could, with only the minor inconvenience of occasionally being shot by Emily Blunt. Worse, that training barely even makes him more competent; instead, Edge of Tomorrow seems to think the advantage of the time loop is just that you can memorize everything that’s going to happen ahead of time. Our lead advances confidently across the battlefield not because he’s improved himself and become a lethal warrior, but because he’s memorized each twist and turn like a rat in a maze. This is why Rita, who’s not presently in the loop, has no remaining advantage and gets killed constantly without his help. L.
The second pillar is a total wash, because the protagonist meets his allies a mere quarter of the way into the movie. He barely has time to understand he’s in a time loop, let alone get existential about it, before he’s spirited into a secret planning room with Rita and He-Shasta, so there’s no sense of relief. Hell, he’s not even convinced to help until after linking up with the others! I initially didn’t mind the changes to Cruise’s character: “talking head military PR guy forced onto the front lines for being an impudent dickbag” is a lot more fun than “generic MC who joined the army for unspecified reasons and has no interesting backstory to speak of”, after all. But when you take away the self-determination and crushing isolation? You’re taking away the whole story. Another L.
Obviously, since things are so off the rails by this point anyway, they don’t manage the ending, either. I shouldn’t be surprised that they wouldn’t let Hollywood darling Tom Cruise do something as objectionable as killing the female lead. They try a softball attempt at this theme: that Cruise has to learn to let her go, and accept that she can’t make it all the way through the loop with him. It’s weak as hell, though, and she ends up surviving anyway. Three for three on Ls.
I’ve focused entirely on the writing here, by the way, because visually Edge of Tomorrow is simply dreadful, from framing to design. Who would’ve thought you could make powered exoskeletons feel this drab and boring?
All You Need Is Kill (2025), again
Alright, so, let’s compare for real now that we’re informed.
This version nails the time loop schtick easily the best of the three. We hit the ground running, establish the loop’s key grounding moments efficiently, and get into our second loop within like, ten minutes. Equally commendable is that All You Need is Kill 2025 doesn’t shy away from having short or unceremonious loops for fear of moving the plot along too slowly: Rita spends a good ten minutes and a few dozen loops just absolutely beefing it before anything like a plan or improvement makes itself known. I really appreciate this in a time loop movie – it really sells the appeal of incremental improvement when someone does so bad for so long.
So, we’ve retained the horror of the original scenario, nailed the self-improvement idea, and added visual flair in a way that’s uniquely suited to the medium. Great job, Pillar One has been achieved handily.
Like Edge of Tomorrow, Kill 2025 makes pretty drastic changes to the lore, setting, and characters. Most notably, while our protagonists are still named Keiji and Rita, everything about them has changed. In the promo material, the movie’s creators made a big deal about the fact that they ostensibly swapped the story’s protagonist by choosing to follow Rita instead of Keiji. The thing is, though, both characters are so fundamentally different from the source material that it doesn’t really matter who’s named what. If the characters are soldiers, they sure don’t look like it, act like it, or discuss it, and neither Rita nor Keiji is a decorated, near-mythological war hero.
None of this matters, however, because the most important part of the narrative – Pillar Two, the trauma bonding – is the primary focus of the movie, and it’s delivered excellently. Through its increased focus on the grinding hopelessness of Rita’s initial loops, Kill 2025 sells her lonely anguish hard, leaving the viewer well and properly primed for that feeling of catharsis once Keiji shows up. What’s more, this adaptation takes things a step further with a subtle but significant change to the source material: both Rita and Keiji’s memories stay intact loop to loop.
The power imbalance inherent to single character time loop stories has long been an issue in the genre, particularly those featuring romance. To its credit, the original Kill novel actually engages with it more than most. Keiji is devastated when he realizes Rita has forgotten all the time they spent together, and struggles to reconcile his feelings with the reality in front of him; this loss is one of the final straws leading to his actions in the book’s finale. Since Studio 4°C is taking a less cynical approach to the characters, though, eliminating this pitfall essentially allows them to turn the second act into a feel-good scifi romcom while sidestepping the ethical problem altogether.
Pillar Two: handled with aplomb!
Much like Edge of Tomorrow, Kill 2025 strays from the source material’s pitch-dark ending, in that Keiji and Rita both walk out alive. Sure, it’s closer: they have their climactic final battle, and Rita ostensibly kills Keiji… but he comes back. Unlike Edge of Tomorrow, I think Kill 2025 has a justification for this: that with its increased focus on its characters’ backstories, resolving those characters arcs is enough to satisfyingly conclude the story without needing to resort to tragedy*. Unfortunately, the conclusion of those arcs doesn’t really feel earned.
*The director’s quote that the story is “about killing someone you love” notwithstanding
There’s certainly some great, subtle characterization to be had in this movie. Without much dialogue, we’re shown how Rita makes her way through life: how invisible she is in the lunchroom, how no one really knows her, how shocked they are when she talks to them. Hell, we see how she’s literally willing to be repeatedly torn apart by aliens rather than directly ask someone for help. So when she trauma dumps her tragic past on Keiji at the end and explains what made her this way, sure, it’s satisfying enough. Unfortunately, I can’t say the same for Keiji, who basically starts and finishes his arc within a span of 15 minutes.
Moreover, what’s hardest for me to believe is that anyone could achieve any meaningful character growth in this particular situation. In a typical time loop movie, the protagonist is stuck reliving a normal day: it’s over the course of hundreds or thousands of human interactions that a character is forced to look within, to think about how their reactions to social situations shape their lives and their own relationships, and what they can do to change. This makes an intuitive sort of sense to me.
Keiji and Rita, on the other hand, are in an action movie. They spend their days programming robo-dogs, upgrading mech suits, and killing, and killing, and killing. How is someone supposed to reflect on past trauma, examine their own behavioral shortcomings and grow as a person amidst all of that violence?
And then we have the ending proper, with Rita finding Keiji safe, sound, and alive when all’s said and done. Look, change is fine, and I’m not beholden to the original novel’s nihilism – there’s nothing wrong with a happy ending. But when the ending of the story is the thing that ties your source material together, you better have a damn good idea if you want to replace it, and they simply don’t here. The climactic finale battle doesn’t really have any stakes if whoever’s killed will just come back to life, so why build it up so much? Even worse, by portraying this outcome as a happy ending, Kill 2025 blunders right back into the genre trope it so deftly avoided earlier: Keiji’s lost all his memories, yet the narrative wants us to think the two still share something. From Rita’s perspective the two have gone to hell and back together, sure, but Keiji has no idea who she is, and the chances the two will form that same meaningful bond together again are extraordinarily unlikely.
So ultimately, I’m left cold. We’ve thrown away the thematically coherent ending of the original, replaced it with muddy character arcs and an ambiguous ending whose text and tone are at odds with each other. We were so close this time…
But hey, if there’s any property it makes sense to constantly reboot, it’s All You Need is Kill, right? Maybe fourth time’s the charm…
I wake up. I write a book. It’s good, but there’s some issues with the prose and characterization – what if we got a big-budget cast to give it that grandiose Hollywood feel? I wake up. There's a movie now. Tom Cruise is here. The American secondary market seems to like it, but the nuance of the original was lost – what if we did a Japanese production, and switched up the protagonist? I wake up. This time, there’s an anime. The director seems to understand the building blocks that make the story compelling, but fumbles things at the finish line. What if, next time–