Week five required a change in method.
Not a change in objective. The objective had remained stable since the first soul entered the reservoir: acquire resources, increase comprehension, reduce uncertainty, build toward a state in which the local hierarchy no longer defined my ceiling.
The method was the inefficient part.
Hunting strong demons individually had carried me through the first month because it was simple, measurable, and survivable. Locate an isolated target. Observe. Identify vulnerability. Engage. Collect. Repeat.
It had worked.
That did not make it optimal.
By the twenty-ninth day, the local ecosystem had begun responding to me.
This was not paranoia. It was pattern recognition. Strong demons that would have ignored distant movement two weeks earlier now reacted to disturbances in the terrain faster. Packs moved in wider formations. Lone entities spent less time exposed in the secondary convergence zones. Twice, I observed a strong demon abandon a feeding position after detecting energy residue from one of my previous fights.
I had become a pressure in the system.
Predators change prey behavior. Competitors change territory use. A repeated anomaly becomes part of the environment, and the environment adapts.
The conclusion was straightforward: if I continued hunting the same way, the return curve would flatten.
I needed to stop chasing targets.
I needed targets to come to me.
The idea came from the dissection table.
There was no table, obviously. There was stone, a dead strong demon, and my hands inside its torso while the body still held enough residual heat and energy structure to remain informative. But the function was equivalent.
Over the previous week, I had learned that demonic bodies did not lose energetic coherence immediately after death. The soul could be collected at once, but the body's energy channels remained structurally present for a short period afterward, degrading over hours rather than seconds. They no longer circulated energy actively, but they retained direction. Architecture remained after function ceased.
That mattered.
If energy channels still had direction, then they could be manipulated before collapse.
Not restored. Not animated. Manipulated.
I killed the first strong demon of week five in the secondary canyon and did not collect immediately. Instead, I opened its torso and studied the energy channels around the upper abdomen, where several large routes converged before distributing into the chest and limbs.
The channels were not blood vessels. They did not contain fluid. They were more like preferred pathways through tissue where demonic energy moved with lower resistance. Death disrupted the pressure gradient, but the pathways remained.
I placed my right palm against the exposed channel cluster and expressed a controlled amount of demonic energy into it.
The result was immediate and wrong.
The corpse convulsed once. Not life. Mechanical contraction from residual energy discharge. The channel cluster overloaded, ruptured in three places, and released a pulse into the surrounding air strong enough that I stopped the experiment and moved backward.
Useful failure.
The pulse carried the energetic signature of a wounded strong demon.
I watched the air for thirty seconds, then opened Design Mode and mapped what had happened.
Input: low-volume energy expression through palm channel.
Medium: dead strong demon energy-channel cluster.
Failure mode: uncontrolled dispersal due to absence of living regulation.
Output: wounded-demon signature pulse.
Application: lure.
The system did not need to instantiate anything. This was not a modification. This was technique.
That distinction was becoming more important. The system was powerful, but relying on purchased structure for every improvement was inefficient. Souls were capital. Technique was labor. Labor could be repeated for free.
I spent the next two hours refining the process.
Too much input ruptured the channels instantly and produced a loud pulse that would attract attention broadly but without control. Too little input produced nothing detectable beyond a few meters. Angled input through the side channels rather than the central cluster produced a narrower signature, directional enough to draw from one side of the canyon rather than every side simultaneously.
By the fifth attempt, I had a functional lure.
A dead strong demon, opened at the torso, with residual channels stimulated by controlled low-output projections at intervals of forty seconds, produced the impression of an injured but living entity bleeding energy into the environment.
An injured strong demon was a resource.
A resource attracted predators.
I chose the kill site carefully: a narrow basin with three approach routes and one elevated exit, seventy meters from the secondary convergence point. The basin walls were irregular, which gave me cover but did not trap me. The strongest approach route came from the south; I placed the body so that anything entering from that angle would expose its left side before seeing the entire basin.
Then I waited.
The first target arrived after twelve minutes.
Strong demon. Alone. Cautious, but not cautious enough. It entered the basin low to the ground, tracking the corpse's residual signature, attention fixed on the body rather than the wall above it.
I killed it with two focused projections.
One to the rear leg joint, dropping its forward momentum into the ground. One to the neck analog while it tried to recover.
Fight duration: eleven seconds.
Physical damage taken: none.
Energy expenditure: moderate.
Soul gain: 100.
[ SOUL COLLECTED ]
Type: Strong Demon — Approximate Value: 100 Souls
Soul Reservoir: 623 / Unlimited
Six hundred twenty-three.
That crossed the threshold.
I did not stop.
Crossing a threshold was not the same as being ready to spend. A minimum cost meant the cheapest possible version. Cheapest possible versions had side effects, quality limitations, and unknown consequences. I had learned that lesson. The system priced capability, but comprehension determined efficiency.
More buffer first.
The second target did not arrive alone.
Two strong demons approached from opposite sides of the basin after the lure's fifth pulse. Not coordinated in the way the pack had been coordinated, but aware of each other. They did not charge the corpse immediately. They circled.
Good.
That meant they were intelligent enough to compete without immediately committing. It also meant they were giving me time.
I let them circle until one crossed the southern approach angle and exposed the thinner tissue under its left shoulder plate. I opened with an arc projection rather than a focused strike.
The arc did what it had been designed to do. It did not kill. It interrupted both at once.
The closer demon recoiled from the energy sweep, not injured enough to be disabled but forced to reset its movement. The farther demon hesitated, and that hesitation created separation.
Separation created sequence.
I dropped from the wall onto the closer one, surface layer active, right palm charged. Contact-range discharge into the exposed shoulder gap. The interaction between my energy and its internal channels was cleaner than it had been before the musculature enhancement. The output traveled deeper, not because the projection was stronger in absolute terms, but because the delivery path through my body was more stable.
The demon collapsed.
The second attacked during the collection window.
I did not collect.
I moved.
That was the correct decision. Souls could wait. Teeth could not.
The second demon was faster than the first and had enough awareness to avoid the corpse, the obvious bait point, and the direct line of my right palm. It had seen the projection. It adjusted.
Good. Better targets taught more.
I used the tail to change direction mid-step, letting its strike pass close enough that the air pressure registered along my ribs. Then I shifted the surface layer into the hands only, reducing total energy drain while increasing contact intensity at the fingers.
The next exchange lasted twenty-two seconds.
It tried to keep distance from the palm.
I used the nails.
The hardened nail structure had been one of my earliest modifications, cheap and simple. I had underestimated its long-term scaling. As my musculature and energy conduction improved, the nails had become more than cutting tools. They were contact points. Small, dense, precise contact points where energy could be expressed into tissue through narrow paths.
Not claws. Needles.
The demon died with five punctures under the jaw and a burned-out channel cluster in the throat.
[ SOUL COLLECTED ]
Type: Strong Demon — Approximate Value: 100 Souls
Soul Reservoir: 723 / Unlimited
[ SOUL COLLECTED ]
Type: Strong Demon — Approximate Value: 100 Souls
Soul Reservoir: 823 / Unlimited
I reset the lure.
The body degraded too far after the third pulse series to remain useful. The energy channels collapsed into noise, no longer carrying the wounded signature with enough clarity to be convincing. I replaced it with one of the new corpses and improved the cut pattern.
Wider exposure at the channel convergence.
Less input.
Cleaner leak.
By the end of the second day, I had killed five strong demons using three lure sites and one relocation after a pack began approaching from outside my acceptable engagement geometry.
Five hundred souls in two days.
Not because I had grown five times stronger.
Because I had stopped using my strength as the only variable.
[ SOUL RESERVOIR ]
Current: 1,023 Souls
Capacity: Unlimited
One thousand twenty-three.
The number looked different from the inside.
At one hundred souls, the system had been possibility.
At five hundred, it had been a tool.
At one thousand, it became infrastructure.
I returned to the secondary convergence zone and opened the design I had been deliberately avoiding until the reservoir justified it.
Energy-matter interface.
The phrase had bothered me because it was too broad. Surface expression was simple: energy outside the body. Projection was simple: energy directed away from the body. Absorption was simple: energy moving inward.
Interface was not movement.
Interface was relationship.
Energy and matter already interacted. My projections struck rock. My surface layer reduced physical force. Contact discharge damaged tissue. But those were crude interactions, mostly output-driven. Energy hit matter. Matter reacted.
An interface implied control at the boundary.
The question was not: how do I make demonic energy touch matter?
The question was: what happens at the exact point where demonic energy and matter negotiate which one changes?
I started with what I knew.
Physical matter had structure. Stress lines. Density gradients. Fracture points. Tissue planes. Demonic energy had pressure, direction, density, frequency, and compatibility signatures. When energy contacted matter, the result depended on the matching or mismatch between the energy's behavior and the matter's internal structure.
Against rock, unfocused projection produced impact.
Against demonic tissue, contact projection produced disruption.
Against the portal anomaly, the advanced entity's output produced stabilization.
Three interactions. Same broad category. Different mechanisms.
I designed the modification as a sensory-control layer rather than a weapon.
Target: Palm channels, fingertip channels, surface expression layer, and tail-tip contact node.
Function: Establish controlled boundary interaction between demonic energy output and contacted physical matter.
Mode A: Read — detect stress lines, density gradients, and energy-conductive pathways at contact range.
Mode B: Grip — bind expressed demonic energy to surface structure temporarily, increasing traction and allowing force transfer.
Mode C: Shear — apply directional energy pressure along identified structural weakness lines.
Restrictions: Contact or near-contact range only. No transmutation. No creation of matter. No remote manipulation. Effectiveness scales with comprehension of target material and energy output control.
Unknowns: interaction with non-demonic matter, portal topology, living advanced-tier tissue.
The system took longer than usual to calculate.
That interested me.
[ DESIGN ASSESSMENT ]
Modification: Energy-Matter Interface — Boundary Contact Layer
Comprehension: Partial-Intermediate
Quality: Intermediate
Cost: 612 Souls
Expected Functions:
— Contact-range structural reading: functional
— Surface grip: functional
— Directional shear: limited by target comprehension and energy output
— Portal/anomaly interaction: possible but unstable; insufficient topology comprehension
Warnings:
Interface operation increases cognitive load. Misreading structural feedback may cause inefficient output or recoil through the host's own channels.
Shear mode against living targets may trigger unpredictable energy feedback if target possesses superior energy integration.
Modification does not grant matter control. It improves boundary interaction only.
612 souls.
Expensive.
Honest.
The warnings were acceptable because they described failure modes I could manage. Cognitive load could be trained. Recoil could be tested at low output. Superior energy integration was already a reason not to touch the advanced entity directly.
1,023 minus 612 left 411.
I approved it.
The installation did not feel like a door opening.
It felt like a second skin learning to think.
The change began in the palms, spread into the fingertips, then propagated through the surface expression layer in thin branching paths. The tail-tip altered last, densifying slightly at the end, not enough to change its shape but enough that I could feel it as a distinct contact point.
Then the feedback arrived.
Rock beneath my left hand stopped being rock in the simple sense.
It became planes and tensions. Pressure history. Microfractures. Density variation. Heat distortion. A set of relationships held in temporary stability.
I pulled my hand away.
The information vanished.
Contact range. The system had been precise.
I touched the rock again and entered Read mode deliberately.
The same structure returned, clearer this time because I was expecting it. I shifted to Grip mode. Demonic energy spread from my palm into the surface irregularities and held. When I pulled, my hand did not slide.
Useful.
I shifted to Shear mode at the lowest output I could manage and directed pressure along a fracture line the interface had identified.
The rock split.
Not exploded. Not crushed. Split.
A clean break eight centimeters long opened beneath my palm with less energy expenditure than a focused projection would have required to damage the same surface.
I stared at the fracture for several seconds.
This was not more force.
It was better force.
That distinction mattered more than the number of souls spent.
For the next six hours, I tested the interface on rock, dead tissue, living growth from the canyon floor, and the remains of strong demon armor plates. The results were consistent enough to begin classification.
Against simple rock: excellent efficiency if a fracture line existed; poor efficiency if the structure was uniform.
Against armor plates: moderate efficiency; the plates distributed force well, but Shear mode could separate the boundary between plate and softer tissue if I had contact long enough.
Against muscle tissue: high risk, high reward; energy feedback was significant, but disruption was precise.
Against bioluminescent growth: extremely strong reaction. The growth's energy system was shallow and distributed, making it easy to interfere with.
I did not test it against a living strong demon that day.
New tools should not be field-tested for the first time in uncontrolled combat unless there is no alternative.
There was an alternative.
The portal.
I returned to the primary canyon on the thirty-third day.
The advanced entity was at the convergence point when I arrived, absorbing. Its maintenance sessions had become more frequent over the previous week. That pattern mattered. Either the anomaly was degrading faster, or something on the other side was applying pressure, or the entity had detected instability after my close observation and adjusted its schedule.
I did not like the third possibility.
It implied my presence had already influenced its behavior.
I settled into the southern cleft, ambient matching active at minimum, absorption running just below net-positive threshold. The interface increased sensory noise even when idle, which forced me to reduce other active processes. Surface layer too high interfered with Read mode. Absorption too aggressive distorted the ambient matching boundary. Projection channels held charge more cleanly than before but produced background pressure in the palms that could be detected if I was careless.
The system had warned me about cognitive load.
It was correct.
I waited through one full circuit.
When the advanced entity moved to the southern wall, I knew before seeing it because the anomaly responded. The interface, even at forty meters and without direct contact, could not read the portal, but it could read the stone around it. The rock carried tension differently when the anomaly pulsed. Tiny shifts. Structural pressure without physical force.
The portal was anchored into the canyon wall.
Not metaphorically.
The anomaly's outer boundary attached to matter through a lattice of energy-matter contact points embedded in the stone. The advanced entity's maintenance work was not stabilizing empty space. It was maintaining anchors.
That changed everything.
Anchors could be studied.
Anchors could be copied.
Anchors could be damaged.
Anchors could be stolen.
I moved closer.
Not to eight meters this time. That had been aggressive before. With the interface active, the feedback risk was higher and my masking less clean.
Twelve meters.
The advanced entity worked with both hands against the anomaly. Its energy output followed the sequence I had observed before, but the interface let me understand the physical side of it now. Every time the entity applied pressure to the portal boundary, the stone anchors took load, flexed, and settled. The process resembled tightening a set of tension cables around a membrane.
Maintenance.
Definitely maintenance.
But not perfect maintenance.
One anchor near the lower left edge of the anomaly was degraded.
I could see it in the rock response. The anchor accepted energy from the entity, held for three seconds, then leaked pressure into the surrounding stone. The advanced entity corrected for it each cycle by applying more energy to that point than to the others.
A weak point.
Not an attack point. Not yet.
A learning point.
I waited until the entity moved through the upper anchor sequence, attention away from the lower left. Then I extended my tail slowly along the rock behind me, using the tip as the contact node. The tail-tip touched the wall twelve meters from the anomaly, far from the active boundary but close enough to the stone network that the interface picked up a faint version of the anchor lattice's structural pattern.
Read mode.
The information was incomplete.
Still useful.
The anchor lattice was not made of demonic energy alone. It used demonic energy as pressure, but the shape of the anchor was imposed through something else. A foreign signature. The same cleaner, higher-frequency energy I had sensed beyond the portal. It threaded through the anchor points like a second material woven into the first.
Two systems meeting.
Energy-matter interface.
The name made more sense now.
The portal was not merely an opening between worlds. It was a negotiated boundary between two power systems, held in place by anchors that forced both systems to agree, locally, that passage was possible.
I wanted that pattern.
Not the portal. Not yet.
The pattern.
I copied what I could into Design Mode while maintaining Read contact through the tail-tip. The workspace accepted the input as structural data, not comprehension. That was fine. Data came before understanding.
The system updated without prompting.
[ COMPREHENSION UPDATE ]
Energy-Matter Interface: Boundary Anchoring — Initial Exposure
Portal Topology: Outer Anchor Lattice — Minimal Structural Data Acquired
Foreign Energy Signature: Detected
Classification: Unknown World-System Energy
World Signature Recording: Incomplete
Recording.
That word held my attention.
The system could record a world signature.
Not identify it. Not open a path. Record.
Which meant world signatures were usable variables.
I kept the tail-tip in contact for another eight minutes.
Too long.
The advanced entity stopped mid-sequence.
No dramatic motion. No roar. No sudden attack. It simply stopped applying energy to the anomaly and turned its head slightly toward the southern wall.
Toward the stone.
Toward the path my interface contact was using.
I killed Read mode instantly.
Ambient matching stabilized half a second later.
The entity spoke.
I understood three words.
Not through fluent comprehension. Through the accumulated seventeen words, the later additions, the cadence, the context, and the system's quiet pattern matching.
[ LINGUISTIC COMPREHENSION UPDATE ]
Infernal Tongue — Vocabulary: 31 words/concepts identified with moderate confidence
New recognized phrase: "small watcher"
New recognized concept: "anchor"
Probable sentence meaning: "Small watcher touches the anchor."
I did not move.
The entity turned fully now.
Its amber eyes did not find me. Not directly.
But it had found the action.
That was worse in some ways. It meant the interface use was more detectable than my body if applied to connected structures. I could hide myself. I could not hide the fact that the anchor had been touched.
The entity raised one hand.
The energy pressure in the canyon changed.
Not attack pressure. Lockdown pressure. The ambient field thickened, and the anomaly's outer anchors brightened in my energy sensitivity as the entity forced the lattice into a more rigid state.
It was securing the portal.
Against me.
Correct response: leave.
I began withdrawing, one meter at a time.
The entity spoke again.
This time the system gave me a clearer structure.
[ LINGUISTIC COMPREHENSION UPDATE ]
Probable meaning: "Not prey. Not servant. Not named."
Not named.
That phrase mattered.
Names existed as a category in the local hierarchy. Or identity markers. Or formal classifications. The entity had determined I was not prey, not servant, and not named.
Which meant named things had status.
I filed it while retreating.
At twenty meters, the entity stopped tracking the wall and returned to the anomaly. It did not resume normal maintenance. It began a different sequence — shorter, sharper, more forceful. The degraded lower-left anchor tightened until the leak disappeared.
I had cost it effort.
Small effort, probably.
Enough to matter.
I left the canyon and did not return for two days.
During those two days, I hunted.
Not because I needed the souls immediately, but because the correct response to being detected by a higher-tier entity was not to sit still and think indefinitely. Thought mattered. So did increasing the cost of killing me.
The lure technique worked twice more before the local strong demons stopped responding to it in the same way. Adaptation again. Expected. I killed three strong demons and one pack coordinator across two controlled engagements.
[ SOUL COLLECTED ]
Type: Strong Demon — Approximate Value: 100 Souls
Soul Reservoir: 511 / Unlimited
[ SOUL COLLECTED ]
Type: Strong Demon — Approximate Value: 100 Souls
Soul Reservoir: 611 / Unlimited
[ SOUL COLLECTED ]
Type: Strong Demon — Approximate Value: 100 Souls
Soul Reservoir: 711 / Unlimited
[ SOUL COLLECTED ]
Type: Strong Demon Pack Coordinator — Approximate Value: 135 Souls
Soul Reservoir: 846 / Unlimited
The coordinator was worth slightly more than the previous one.
Not a fixed class. Individual development mattered. Behavioral complexity mattered. Possibly leadership experience mattered. The soul system valued accumulated structure, not just raw power.
That was worth remembering for eventual creature design.
On the thirty-sixth day, I returned to the primary canyon.
Not to touch the anchor again.
To listen.
The advanced entity was waiting at the convergence point.
That was my first impression, and I tested it carefully before accepting it. It sat in the same absorption posture as before, but the energy pattern was different. Less inward flow. More ambient awareness. The body was still; the field was not.
It was watching for the watcher.
Good.
That meant I had become significant enough to occupy attention.
Bad.
Same reason.
I stayed at sixty meters instead of forty, ambient matching low, interface fully inactive. For three hours, neither of us did anything that would look meaningful to an observer without energy sensitivity.
Then the entity spoke.
This time, I was certain it was speaking to me.
[ LINGUISTIC COMPREHENSION UPDATE ]
Infernal Tongue — Vocabulary: 44 words/concepts identified with moderate confidence
Probable meaning: "Small watcher. Learn far. Touch not."
I considered answering.
The calculation was not simple.
Silence had preserved ambiguity until now. But ambiguity was decaying. The entity knew there was a watcher. It knew I had touched an anchor. It knew I was learning. Continuing to provide no response might maintain the unknown, or it might reclassify me from cautious anomaly to invasive threat.
A minimal response could establish a different model.
Not enemy.
Not servant.
Not prey.
Learner.
I had forty-four words. Not enough for grammar. Enough for one crude signal.
I spoke in Infernal Tongue for the first time.
The pronunciation was imperfect. I heard the errors as I made them. The registers did not sit naturally in my throat; this body could produce them, but I had not trained the muscular control.
One phrase.
"Learn. No break."
The canyon became extremely still.
The advanced entity turned its head toward me.
This time, its eyes found my position.
Not approximately.
Exactly.
So speech broke the masking.
Important.
It stood.
I prepared every active system without visibly moving. Surface layer at minimum. Projection channels charged but not expressed. Arc mode available. Interface inactive. Absorption off. Reservoir full.
The entity did not attack.
It spoke.
The system translated fragments, not the whole.
[ LINGUISTIC COMPREHENSION UPDATE ]
Recognized concepts: "learn," "break," "anchor," "named," "below," "gate," "permission."
Probable meaning: "Learning without name breaks gates. Below gives permission."
Below.
The thing below.
The source.
The entity was not an owner.
It was a caretaker.
And it answered to something beneath the surface.
I gave no second response. One had been enough. More would expose ignorance faster than it would build communication.
The entity watched me for nineteen seconds.
Then it sat back down.
Not dismissal.
Permission was too strong a word.
Tolerance.
Temporary tolerance under observation.
I left before that tolerance could change.
That night, in the secondary convergence zone, I opened the system workspace and reviewed the world signature recording.
It was still incomplete. A fragmentary set of energetic intervals, anchor-shape relationships, and foreign-system resonance patterns captured during my tail-tip contact with the portal lattice.
But the system could now display it as a variable.
[ WORLD SIGNATURE RECORD ]
Source: Unknown Connected World
Status: Incomplete
Usability: Insufficient for portal creation
Usability: Sufficient for recognition if encountered again
Energy Character: Structured foreign system; non-demonic; compatible with negative-emotional residue classification pending confirmation
Anchor Pattern: Outer lattice partial
Risk: Existing gate maintained by higher authority; unauthorized interaction may trigger enforcement
Negative-emotional residue.
I read that line several times.
The system was not naming the world. It did not have enough data.
But it had inferred a category.
A world where power interacted with negative emotional residue.
That narrowed possibilities dramatically.
I thought of the fiction I had consumed in my previous life. Power systems built on fear, hatred, grief, and death. Energies produced by human negativity. Invisible structures layered over ordinary civilization. Monsters born from emotion.
Interesting.
Very interesting.
I closed the record and looked at my hands.
Thirty-six days.
846 souls.
Energy-matter interface installed.
A world signature fragment acquired.
First communication with the advanced entity established.
The thing below confirmed as a governing variable.
The portal no longer represented an eventual curiosity. It was now a strategic direction with enough data to become a plan.
Not enter yet.
Entering another world through a gate maintained by an entity above me, with permission apparently controlled by something below the surface of the Infernal Realm, would be idiotic at current comprehension.
But preparing to enter?
That was different.
Preparation was what I did.
I opened a new design workspace.
Title: Gate Interaction — Unauthorized Micro-Aperture Feasibility.
The system did not reject the title.
That was enough to make me continue.
I began planning week six.