Solver drones, as a rule, do not self-modify.
Worker drones have some basic capacity for self-repair. Now, they don’t heal, they’re still robots. Cracks and degradation of components can be sealed and reinforced with the nanites contained in oil, but this process is unable to actually restore missing pieces.
Instead, workers have a internal fabrication. They can intake raw matter, metal and plastic and sand, and mechanically transform them into new components. All that’s left is maintenance; the parts then need to be taken out of the mini-factory in their stomach and secured in their rightful place.
For worker drones, then, healing and self-modification are essentially the same act of mechanical repair.
This isn’t the case for solver drones. Every solver quickly masters the restorate command, and transcends beyond the need for manual repair. They still need to intake matter, but it’s stored in a deep extradimensional pocket. Their command-augmented bodies can turn their matter reserves into replacement parts in a fraction of the time that workers require, and commands can summon the newly manufactured replacements in place.
But of course, how does the restore command know what pieces to create and where to put them? The drone’s specs need to be understood programmatically; and it’s far too much data to write and configure manually. Thus, each solver has a stored blueprint, an ideal state to restore their bodies to.
But the most important thing to know is that solvers restore instictively. This means that a solver who tries to upgrade their bodies will either have to suppress their restoration, or inevitably “heal” back to normal. In theory, you could change the blueprint, but solvers’ most potent defenses exist to protect their blueprint; it must be protected from corruption, because it’s what protects the rest of the drone from corruption.
It’s rare to find a driver with the ability to mess with blueprints, the precision to do so in a controlled way, and even if you could make any change to the blueprint that you want, what would you actually change? You’d need to understand deeply how it all works, model the ramifications of every change on the complex system of drone mechanism. If you mess this up, it might kill you, or it’ll disable to you such that you can only be healed back to this disabled state. It’s far too risky for most solvers to dare.
This, incidentally, is part of why you don’t have more drones like Beau walking around. A functional disassembly drone arm has a vast and complicated blueprint, and that blueprint is stored on the disassembler’s hard-drive, not their arm. You would have to succeed in reverse engineering a blueprint, and then once you did, you’d then need to figure out how to patch a drone’s blueprint and then finally find someone willing to undergo that procedure.
And yet, Beau is the exception. Alice attached a disassembler’s arm to him, and his blueprint incorporates it. Most worker drones couldn’t sustain being wired up to a disassembly drones (the oil and power draw would be dehibilating if not outright intractable for a worker drone.) Alice, though, is so adept at working on drone internals that she could jury-rig a setup to make it work.
(Those especially fluent in the deep lore of corrupt combustion will know that part of why Alice knows so much about disassembly drone specs is due to her relationship with T.)
It’s theorized that because Beau already had the disassembly drone arm attached to him when his core ignited, the arm was stored as part of his innate blueprint.
Obviously, it would be deeply unethical to test this theory, so for now it’s just a theory.
However, Yeva scanning Beau’s unique blueprint is part of what leads to the design of her monad prosthetics. Of course, whenever the gauntlets break, she needs Alice to repair them, because when they’re broken, her ability to execute restoration goes with them.
Gauntlet transformation operates on the same principle as the restore function; material is pulled from the dimensional pocket. Through Beau, they discover another feature: if he holds something in his hand when he transforms, that material is taken to the pocket dimension. It’s a convenient way to dispose of trash.
As Beau starts operating on drones, a curious interaction is discovered. Say a drone has a damaged internal part that needs to be replaced. Beau removes it, and disposes of it with a gauntlet transformation. He installs a replacement.
If that drone was a solver, he has damaged their blueprint. This is a difficult fact to notice at first — necessarily, it only happens with especially small parts. Maybe by chance, Yeva and her highly perceptive eyes are supervising Beau while he works, or he happens to dispose of a crucial yet still small component.
When Beau sends a part of a drone to his subspace, that part disappears from the target’s blueprint — as if corrupted. Not permanently; it registers as an enemy command, and it can be countered with data recovery
Disassembly drones have so many features custom built for outmatching and disassembling solver drones. It’s theorized that this is yet another, rarely observed for its niche requirements and effects. How many tricks did disassembly drones have up their sleeves?
Thus, going forward, Beau is careful not to use this ability on his allies, and tends to be pretty difficult to set up a situation where he can rip out a foe’s parts and transform them away.
Now, to talk about some retcons.
After arc two (“Warpath to the Corpse Spire”), the Wheel Group notices a new type of zombie infesting copper-9. Slithernots are small ring 4 annoyances, like a bundle of wires crawling across the ground. They’re poorly understood because of their passive ability: they can trigger the corruption-censoring circuitry even in ignited solver drones that otherwise have those safeguards disabled.
This means that you could walk into a room teeming with dozens of slithernots and not register their presence. If they obscure something you’re looking at, you’ll think nothing of it, or nudge them aside on autopilot, as perceptively invisible as the frames of your glasses.
Of course, this effect is achieved through corrupt commands, and so it’s possible to resist; and it’s possible for it to simply fail. If they are sufficiently distracting, if one is alone (and thus without other slithernots compounding the forget-me aura), it’s possible to notice them.
When you become aware of a slithernot, it hisses and attacks, as if offended at being perceived. And once you realize they’re there, they present no danger. Remember, ring 4. Most cadets could (literally!) stomp on these things to kill it.
There’s just one last trick up their sleeve: when a slithernot dies, they explode with data. It’s like a flashbang or EMP composed of corruption, run through the same function that grants slithernots the forget-me aura. Only now, rather than awareness of the slithernot slipping from your mind, everything slips from your mind. You’re briefly dazed, the entire content of your working memory wiped, and the last few moments are blurry if not outright forgotten. What were you doing here?
Slithernots are thus pretty annoying to deal with, and the Wheel Group quickly has to deal with a lot of them. As you can imagine, for every slithernot you become aware of by chance, there must be several you walked by without thinking about.
The Wheel Group develop’s strategies to scan for slithernots and task teams with clear out the densest infestations of them. There’s enough of that every active solver becomes familiar with the tedium. It’s unclear how they’re breeding so fast.
Slithernots are captured and experimented on, and studies conclude that the command that cloaks slithernots is keyed to their size. The bigger and more data-heavy they get, the more their cloaking wanes in effectively. The curve is a dramatic slump.
So, there’s good news: worries that slithernots would develop or mutate into a ring 3 or greater antimemetic threat seem unlikely to materialize; there can’t be any hulking monsters with this function hiding from them.
The bad news: the corollary to the above that the function becomes substantially more effective when they’re smaller; they can breed so fast because baby slithernot are invisible.
Now, the Wheel Group has another route to understanding and countering this threat: Adam. His innate function is [Zombie Process Management], after all. Unfortunately, he snaps a few pics of slithernots and it doesn’t seem to do anything.
Normally, one photo is enough to catch a ring 4, but there are a few reasons one might not be enough. Stronger zombies require more photos, of course, which is the immediate and most concerning conclusion to draw, but there are other reasons. Bad photos have neglible effect, and slithernot’s stealth isn’t entirely corruption-based. And, fundamentally, Adam’s photos must be of the zombie; photographing part of the zombie, or a construct or creation of the zombie won’t work.
Clearly, some assumption they’ve made about slithernots is incorrect.
For all that slithernots pose little danger, there’s clearly further mysteries to investigate. Research committees and strategies outline plans to get to the root of their origin, and find a way cut it off at the head.
Unfortunately, no one remembers to follow up on that.
Speaking of Adam, I’m considering some major retcons happen to the Adam plotline.
First is that Cyn is not a zombie anymore, so Adam’s driver simply can’t work on her; the battle against him thus cannot happen as described. Instead, what might happen is that when Beau lets Cyn out, she detaches one of her threads and lets it get consumed by a zombie, And through that zombie, takes over Adam’s system.
I also don’t know what to do about the Null Sync. It’s a very cool moment, but it undeniably would make my life easier if it doesn’t happen.
I’m also rethinking how the arc is set up. To understand why, consider this: why does the arc happen? When must it happen? What does it amount to in the end?
Now, for reasons I won’t get into, Adam’s defeat is crucial to setting up the Citadel Incident arc, but overall, it feels like a plot detour.
How can we fix it? Here’s one possibility:
What if Beau isn’t just the first person to figure out who’s behind the core snatching — no, he’s the only person who can figure it out. Whenever anyone else begins to suspect that Adam is doing something shady, they just… forget about it. The reasoning slips away.
Except for Beau. He’s able to hold on to the thought, and when he tries to convince others, they all think he’s crazy — with one exception. Nori isn’t able to remember, either, but she trusts Beau, so when he says some crazy stuff is going on, she entertains it.
After Adam’s taken out, Nori goes to the detective, and explains the memory issue. This is clearly a tricky problem, and the detective’s the only drone who might have a chance of cracking the case.
(You aren’t able to appreciate the full significance of this until the Anonymous Group makes their proper debut, which is chronologically next after Compress & Extraction.)
But there’s a suggestive outline to draw around Adam’s slipperiness, the Anonymous Group’s anonimity, and the slithernots. The obvious theory would be that Adam controlled the slithernots all along, but they don’t disappear when he’s dead. Did he finally achieve partial control of them without telling the Wheel Group? Is something weirder is going on?
So yeah, that turns Adam’s defeat into setup for the final pre-citadel arc, where Beau and the Detective (and probably Lizzy & Doll) work together to uncover what’s wiping everyone’s memory and why.