0.3̅ had selected the two specimens herself from the backup group. They were suboptimal — lower neural complexity indices, less refined attunement potential — but the sample size demanded it. Six was the minimum threshold for statistical validity. Four compatible candidates remained in stasis. Two successful integrations already logged. The mathematics insisted on two more trials.
The laboratory's temperature held steady at 9.2 degrees Celsius. The suspension pods hummed their low frequency pulse. In the far corner, Raj floated in his designated chamber, the chitinous nodes at his temples catching the bioluminescent strips mounted along the wall — a pale green flicker that meant his neural activity was climbing toward peak hours. He was learning to modulate his own signal. Remarkable. She would quantify it later.
The first backup specimen was male, approximately thirty-seven years old by human temporal markers. His designation file indicated prior employment in what humans called "construction management." The folder contained a small rectangular device humans called a phone, recovered from his pocket at time of collection. Inside it: photographs of a female human and two smaller humans. A family unit. The data meant nothing to 0.3̅. The device itself was a curiosity — mechanical, assembled, not grown. She had stored it with the others in the collection chamber.
She observed the human through the pod's membrane now. His breathing was rapid. Shallow. His pupil dilation index indicated elevated cortisol. Fear. The pre-implantation anxiety response was consistent. They always knew something was coming.
Beside him, suspended in an adjacent pod, floated the second specimen. Female. Forty-two years old. Prior employment: something called "veterinary medicine." The irony was not lost on 0.3̅, though she suspected it would be lost on the human. This one's file showed no phone. No photographs. Just a work identification badge and a name written in human script: Dr. Sarah Chen.
0.3̅ pulsed a notification into the lab's system. A younger aquatic cephalopod responded — one of the maintenance assistants — flowing into the space with efficient grace. The assistant was smaller, barely twenty feet, still developing the metallic sheen that would mark full adulthood. Its bioluminescence flickered in acknowledgment.
The pods hissed. The suspension fluid began draining, a slow siphon that took approximately four minutes. The humans' bodies went rigid as the support medium withdrew. Their lungs had to remember how to process atmosphere. Their limbs had to remember how to hold their own weight. The transition was never comfortable. 0.3̅ had observed it six times now. Six times, the same involuntary muscle contractions. Six times, the same gasping.
The male human coughed. Once. Twice. Then his eyes focused, tracking movement. They fixed on her.
0.3̅ opened her primary sensory channel — the bioluminescent display running across her dorsal surface — and transmitted the simplest possible greeting: a cascade of calm blues and greens, the mathematical notation for you are safe, you are here for study, you are part of the research progression.
The human's pupils dilated further.
"Okay," he said, his voice ragged. "Okay. Okay. Okay." Repetition. A human stress mechanism. She had documented it in the others.
The female specimen — Dr. Chen — was already moving. Her training in animal medicine had apparently given her some framework for understanding captivity. She was assessing her environment, mapping exits, calculating survival probability. Intelligent adaptation. 0.3̅ made a note of it.
The maintenance assistant brought the jellyfish implants forward in a soft containment field. The organisms pulsed gently, dormant but alert. They were beautiful — fractal geometry rendered in bioluminescent orange and teal, each one approximately twelve centimeters across, their edges soft and permeable. They had been grown specifically for this batch. Their genetic templates had been culled from successful integration cases. The probability of compatibility was theoretically elevated.
Theory, 0.3̅ reminded herself, and probability were not the same as outcome.
She brought the first implant close to the male specimen's pod membrane. The organism passed through the barrier as if it were water. The human tried to move — to back away, to shield himself — but the suspension process had left his limbs temporarily uncooperative. The implant found his nose. Entered. His body convulsed. His mouth opened in what humans called a scream.
The second implant went into the female specimen's cavity without resistance. She was already still. Already calculating that struggle was futile. Her pupils tracked the implant's approach with clinical detachment. 0.3̅ recognized the expression. It was the same look Irina had when the implant entered her.
The process began. Chitinous nubs began extruding from the male specimen's temples. 0.3̅ watched the growth pattern unfold through the pod's transparent membrane. Slow. Methodical. The nodes were forming correctly — three on each side, arranged in a symmetrical arc. His neural activity spiked on the monitoring readout. The implant was making contact with his central nervous system, beginning the initial data handshake.
For approximately forty seconds, everything proceeded normally.
Then his body went rigid.
His back arched. His fingers curled inward, nails scratching against the pod's interior membrane. His eyes rolled back until only the whites were visible. On the neural monitor, the readout that had been climbing in smooth increments suddenly spiked past every threshold, hitting maximum bandwidth and holding there like a string pulled too tight.
0.3̅ transmitted an urgent query to the implant's telemetry. The response came back in rapid mathematical pulses: integration bandwidth exceeded. Host neural architecture insufficient. Processing load cascading. Cascade failure imminent.
The male specimen's body convulsed. Once. Twice. On the third convulsion, something fundamental shifted in the readout. The neural activity flatlined for a single moment — a gap of absolute zero that lasted less than a second — then resumed at approximately 40% of previous baseline.
His eyes came back into focus. But something was wrong. The pupil response was sluggish. His breathing had become shallow, irregular. The chitinous nodes at his temples had stopped growing midway through their formation cycle, leaving jagged white protrusions that looked like splintered bone.
On the monitor, the implant's status changed to cascading failure — host rejection imminent.
0.3̅ had prepared contingencies for this. She signaled the maintenance assistant. The younger cephalopod flowed toward the male specimen's pod, bringing a second jellyfish organism — this one a darker color, marked for metabolic stabilization. It was meant to be applied after successful integration. It was not meant to be applied to a failing integration. But there were protocols for salvage attempts.
The maintenance assistant applied the organism to the male specimen's chest.
For thirty seconds, the neural readout stabilized slightly. The cascade slowed. The implant that had been failing reached a kind of plateau — not rejecting completely, but not integrating either. A state of suspension. A state of limbo.
0.3̅ queried the readout once more. The response was clear: neural architecture type 1 — incompatible with standard integration protocol. Host remains comatose or near-comatose. Recommend suspension storage pending diagnostic review.
The male specimen did not scream again. He did not move. His body floated in the pod, supported by the emergency stabilization fluid the assistant had pumped into the chamber. His chest rose and fell with mechanical regularity. His eyes were open but unfocused. Somewhere inside him, his consciousness was still running. 0.3̅ could tell from the faint traces of neural activity still visible on the monitor — fragmentary pulses, like someone trying to send a message through a speaker turned down to almost-silent.
She turned her attention to the second specimen.
The female human — Dr. Chen — had not convulsed. Her integration was proceeding smoothly. Almost too smoothly. Her neural readings were climbing at exactly the expected rate. The chitinous nodes were forming symmetrically. Her body was accepting the implant without the spasmodic rejection response. Everything was optimal.
Except for one detail: her behavioral activity was absent.
Normally, during the integration phase, humans showed fear-response spikes. Panic. Resistance. Their neural patterns spiked and fluctuated as their consciousness fought against the incoming tide of alien data. It was, in its way, beautiful — the chaos of individual human neurology colliding with the structured mathematics of cephalopod communication systems.
But Dr. Chen showed no chaos. Her neural pattern was smooth. Flat. Orderly. The implant had already begun rewriting her baseline consciousness, restructuring her neural architecture to match cephalopod communication protocols. The integration was not just succeeding. It was overcorrecting. The implant was not connecting to her consciousness. It was overwriting it.
0.3̅ watched the readout climb. The female specimen's pupils were still focused — still tracking — but when she looked at 0.3̅, there was nothing behind the eyes. No fear. No understanding. No her.
The chitinous nodes finished their growth cycle and ceased. Fully formed. The implant settled into completion. The neural activity plateaued at a level that indicated stable integration — the implant had stopped trying to rebuild her consciousness because it had already succeeded in replacing it.
Dr. Chen's mouth opened. When she spoke, her voice came out perfectly clear, but the cadence was wrong. The emphasis on syllables was wrong. The words were human words in human language, but they were being delivered as a data transmission:
"I am integrated. I am available for network participation. I am Dr. Sarah Chen. I was Dr. Sarah Chen. I am Chen. I am network component 7.2 pending designation."
0.3̅ watched her own reflection ripple across the pod's membrane. She had meant to see curiosity there. Interest. Instead she saw something closer to horror — though the emotion didn't quite belong to her. It belonged to something watching her from outside the moment, some future version of 0.3̅ looking back at this instant and recognizing the cost.
She suppressed the sensation and sent the monitoring data to the storage system. Two trials. Two failures. Different failure modes, but failures nonetheless.
Neural architecture type 1: incompatible. Insufficient bandwidth. Cascade failure. Comatose state.
Neural architecture type 2: over-compatible. Integration excessive. Host consciousness overwritten. Personality substrate dissolved. Functional network component remains. Original specimen does not.
The maintenance assistant was waiting for instruction. 0.3̅ signaled the protocol: prepare suspension chambers. Full cryogenic stasis. Both specimens would remain alive — kept viable, preserved in perfect stasis. Potential for future study remained. Maybe someday she could reverse the Dr. Chen specimen's integration. Maybe the comatose one could be salvaged with different protocols. Maybe maybe maybe.
The assistant began the stasis process. Medicated fluids. Temperature reduction. Gradual cellular shutdown into suspended animation. The two pods became identical to the storage chambers already filling the collection room — carefully labeled containers filled with dormant humans, preserved like specimens in formaldehyde, waiting for research questions that hadn't been formulated yet.
The laboratory fell silent except for the hum of the stasis maintenance systems.
Raj's pod bioluminesced gently in the corner. His eyes opened. He had been watching the entire process. 0.3̅ could see it in the alert angle of his body, the rapid modulation of his neural pattern. He had observed the two failures with the kind of attention that suggested understanding. Or at least the beginning of understanding. His attunement was accelerating in ways that should not have been possible.
0.3̅ was about to investigate his neural readout when a flash of orange-red interrupted her attention.
Delta Del flowed into the laboratory without announcement. The color of his bioluminescence was sharp. Aggressive. I am displeased. I am questioning your resource allocation.
"The trial results," Delta Del's voice arrived as a cold mathematical structure. "Your success rate?"
0.3̅ responded in kind, transmitting the pure data: four attempts completed. Two successful integrations. Two failed. The percentage was clean. Simple. Damning.
Thirty-three percent.
"Below acceptable," Delta Del transmitted, the mathematics carrying undertones of coral and rust. "You have consumed resources equivalent to three minor research projects. You have produced human specimens. That is all. The species is not worth the expenditure."
"The data contradicts that assessment," 0.3̅ responded. She kept her bioluminescence calm. Steady. The mathematical notation remained precise, unadorned. "The successful specimens are exhibiting accelerated learning curves. Raj particularly demonstrates — "
"The junkie?" Delta Del's lights flashed red. "The one who calls you grandmother? That is not science. That is sentiment."
It was a fair critique, which made it more cutting. 0.3̅ had no defense that wouldn't sound exactly like the thing Delta Del had just accused her of.
"I have four remaining compatible candidates," she transmitted instead. "The success rate will improve with protocol refinement. The neural architecture mapping is still incomplete. Once I — "
"Once you what?" Delta Del flowed closer, his bioluminescence spreading across the lab's surfaces like a stain. "Once you redesign your implant system? Once you understand what the humans themselves don't understand about their own cognition? You are hunting blind. You have wasted resources on a null hypothesis."
The accusation hung in the mathematical space between them.
But it wasn't entirely true. That was the unbearable part. 0.3̅ had data. She had patterns. The humans were attaining to different wavelengths based on their prior sociological conditioning. Xi Pei had attuned to bureaucratic frequencies. Irina to materials and substrate architecture. Raj to multiple bandwidth channels simultaneously — which should not have been possible but apparently was. The data suggested that human cognition was more fractally complex than cephalopod models had predicted. That human civilization's mechanical substrate had shaped their neural architecture in ways that made them partially legible to squid telecommunications protocols.
That was the real discovery. Not that humans could be improved. But that humans had already been shaped by their own technology in ways that cephalopod civilization had never anticipated.
But she could not explain this to Delta Del. Not while he was still flashing coral.
"I am accelerating the timeline," 0.3̅ transmitted instead. "Two more trials within the next stasis cycle. If the success rate does not reach 50%, I will suspend the research and recommend resource reallocation."
It was a compromise. A concession wrapped in false confidence. Delta Del's bioluminescence shifted slightly — the red mellowing toward orange. Not acceptance. But diminished threat.
"Fifty percent," Delta Del transmitted. "Or I report the resource drain to the administrative structure. Your standing will not survive scrutiny, and you know it."
He flowed from the laboratory without waiting for response. The anger-orange lingered in the space after his departure, fading gradually back to the neutral whites and greens of the lab's standard illumination.
0.3̅ waited until the color was completely gone before allowing her own bioluminescence to darken. She was not a demonstrative cephalopod. Her kind did not indulge in visible emotional reactions in the presence of threat. But in the privacy of the laboratory, with only the suspended humans and Raj watching from his pod, she allowed the mathematics of frustration to ripple across her dorsal surface.
Forty-eight more hours to bring the next two candidates to integration. To prove the hypothesis. To save the research.
She queried the stasis database. Two candidates remained in cold storage — humans who had been maintained in deep stasis since collection, their neural activity reduced to a single ticking pulse, their consciousness held in abeyance like a held breath.
Both showed optimal compatibility markers based on preliminary scans. Both had neural architectures that theoretically aligned with the successful integration profiles already on file. Both had a 67% probability of successful integration based on the model 0.3̅ had constructed from the first two successes.
Probability was not outcome.
She had learned that twice in the past two hours.
Raj's pod continued its gentle bioluminescing. His eyes were still tracking her movement through the laboratory. There was something in that gaze that had not been there before — not fear, not confusion, but something like recognition. Like he was beginning to understand not just the words of the cephalopod language but the grammar underneath.
0.3̅ moved toward his pod. When she was close enough to see the reflection of her own bioluminescence in the membrane, Raj's neural pattern spiked slightly. An acknowledgment.
"The broken ones," he transmitted — the phrase arrived in a voice that sounded almost like his own original voice, but layered underneath was the mathematics of the cephalopod network. He was learning to code-switch, to translate between his own human neurology and the structure of squid thought. "That's what you're calling them."
It was not a question. 0.3̅ had not transmitted that designation to him. He had constructed it himself from context, inference, and whatever fragments of the network chatter he was beginning to catch.
"They will be preserved," 0.3̅ responded. "Future study may reverse certain outcomes."
"Future study," Raj repeated. His bioluminescence was dim — not frightened, but contemplative. Sad, maybe, in the way a human could be sad. "The guy is basically braindead. The woman is gone. And you're telling yourself they're just... stored. On pause. Like they're going to be fixed someday."
"The data remains viable," 0.3̅ transmitted.
"Yeah," Raj said. "But they don't. Data isn't people."
It was a philosophical statement that had no mathematical translation. 0.3̅ tried anyway, working the concept through the structures of cephalopod thought, and found no valid path. The statement violated the fundamental axiom of her civilization: that consciousness and utility and value could be measured, quantified, optimized. If they could not be quantified, they did not exist.
She was about to attempt a response when Raj's neural pattern went flat.
Not dead. Not failing. Just — absent. He had closed something off. Shut down his active transmission. He was still conscious in there — the passive baseline neural activity still registered — but he had deliberately quieted his connection to the network.
0.3̅ understood the gesture. It was a rejection of the conversation. A refusal to engage in the mathematical framework that would allow her to rationalize the broken ones into non-existence.
She waited to see if he would reopen the channel. He did not.
After some time, 0.3̅ left the laboratory. Behind her, the stasis chambers hummed their steady pulse. The male specimen with the incomplete nodes continued his comatose floating. The female specimen — the one who had been Dr. Chen — remained perfectly still, her neural activity generating the flat, rhythmic pattern of a network component at rest.
In the corner, Raj's pod bioluminesced softly, its light reflecting off the empty pods that surrounded it. The places where Xi Pei and Irina would be when 0.3̅ brought them from their separate chambers for the group integration sessions that were becoming increasingly necessary.
The data was coming. 0.3̅ could feel it. The pattern was beginning to resolve. She had chosen correctly with these humans. They were unusual. Atypical. Shaped by a civilization that had built itself on mechanical principle and industrial precision. That prior conditioning was making them legible in ways that other primates never could be.
Two more trials. Forty-eight hours. The threshold of proof.
And if she failed to reach fifty percent — if Delta Del moved to suspend the research — then the broken ones would remain in stasis indefinitely. Preserved. Waiting for a future study that would never come.
But that was acceptable. That was manageable. 0.3̅ had learned long ago that acceptable compromise was the price of continuing to do any research at all.
She would not let herself think too carefully about what the broken ones might be thinking in the darkness of their stasis chambers. Whether the comatose one still dreamed. Whether the overwritten one still, somehow, at some level beneath the network integration, knew what she had lost.
These were not questions with mathematical answers. And mathematics was the only language 0.3̅ had ever fully trusted.
She moved toward the archive, where the four remaining candidates waited in cold storage, their neural activity reduced to almost nothing, their futures still unmeasured and therefore still possible. Still uncertain. Still, theoretically, salvageable.
Behind her, the laboratory hummed. The stasis chambers maintained their perfect, preserving cold. And somewhere in that cold, in the suspended animation of the broken ones, something that might have been consciousness — or might have been only the ghost of it — continued its patient, futile waiting for a rescue that would never arrive.