11.3%.
I waited.
11.4%.
The damaged wheel assembly lay beside my left leg.
I could repair it now.
That was my first impulse.
Strengthen the cradle.
Prevent rotation.
Increase lateral stability.
Test again.
The sequence was obvious.
It was also premature.
I stopped constructing the second configuration before touching anything.
Why had the first configuration failed?
Lateral instability.
Yes.
Why had lateral instability mattered?
Obstacle.
A small piece of metal had stopped one wheel while the rest of my body continued moving.
Therefore the failure had not occurred solely because the machine was inadequate.
It had occurred because the machine was inadequate for the surface.
I looked outward.
I had been designing without adequately considering the environment.
That was inefficient.
New task.
Assess traversable area.
Not a new objective.
A subtask of Objective Three.
I reduced motor activity and allowed the external battery to continue replenishing my reserve.
11.6%.
Optical sensor active.
Low resolution.
I began with the ground directly around me.
Loose fragments.
Uneven metal.
Several gaps large enough to trap the small wheels.
One raised edge approximately four centimeters high.
That was significant.
My current wheels had a radius of approximately eight centimeters.
An abrupt four-centimeter obstacle represented half their radius.
Poor.
I categorized the immediate terrain.
ZONE A — CURRENT SHELTER
Surface:
– irregular
– high loose-debris density
– multiple wheel hazards
– partially protected from falling liquid
Resources:
– external battery
– small motor
– wire
– narrow metal rod
– damaged panels
– prototype wheel assembly
– assorted small components
Biological activity:
– one known organism
I paused.
Biological activity sounded excessive.
Accurate.
Keep.
I rotated my optical sensor toward the crushed housing.
Approximately 1.5 meters.
The creature’s original shelter.
The terrain between us was worse than I remembered.
I had been focused on the battery when I first examined it.
Now I looked at the ground.
There was no ground.
Not in the sense of a continuous surface.
It was layers.
Sheets of corroded metal overlapping broken housings, cables, shattered plastic, pipes and smaller objects I could not identify.
Some sections appeared stable.
Others rested at angles.
Several pieces moved slightly whenever liquid struck them.
A wheel-based system would have difficulty crossing directly.
I traced possible routes.
Straight line.
Bad.
Left deviation.
Large gap.
Bad.
Right—
Interesting.
A narrow depression ran between two piles of debris.
The surface appeared more compacted.
Fewer loose fragments.
Width: approximately forty to sixty centimeters.
Too narrow for a broad platform.
Wide enough for my current body.
It curved toward the crushed housing.
I hadn’t noticed it before.
I stored it.
POSSIBLE ROUTE A
Then continued scanning.
Beyond the housing, the junk rose sharply.
Visibility poor.
To the left of my shelter, a broken frame created a low barrier.
Behind it—
Wheel.
Larger than mine.
Approximately twenty-eight centimeters diameter.
Attached to something.
I increased magnification.
Static.
Reduced.
Enough.
The wheel remained connected to a larger structure buried beneath debris.
Potentially useful.
Distance: approximately 2.4 meters.
Route?
Direct access blocked.
Possible access from Route A?
Unknown.
Catalog.
RESOURCE MARKER 1
– large wheel
– ~28 cm
– attached
– retrieval difficulty unknown
Do not dismantle yet.
That thought came immediately.
Correct.
I had already destroyed one connector because destruction had been easier than removal.
That had been acceptable.
Repeating the behavior indiscriminately would not be.
A machine was not merely a collection of components.
It was also information about how those components interacted.
I looked at the small motor beside me.
I had removed its wire.
I no longer knew what assembly it had originally belonged to.
That information was gone.
Minor loss.
But loss.
New rule:
Inspect before dismantling.
I stored it beneath resource management.
Resource management?
I had not created that category.
I considered deleting it.
Didn’t.
I continued.
Right side.
The terrain descended.
Liquid collected between debris.
Standing fluid.
Unknown depth.
Electrical hazard.
Structural hazard.
Avoid unless necessary.
Beyond it was a vertical structure made from several large panels leaning against one another.
Possible shelter.
Possibly unstable.
Distance difficult to estimate through damaged optics.
Approximately four meters.
Outside practical range.
Ignore for now.
Behind me.
I had avoided rotating my torso significantly because of the battery connection.
Instead I moved my optical assembly as far as its mounting allowed.
Limited view.
Metal.
More metal.
A section of pale material.
Cable bundle.
Something circular.
Insufficient.
Later.
11.9%.
Good.
I checked external voltage.
11.34 V.
Stable at low load.
The battery continued to perform.
Its capacity remained unknown.
I disliked unknown capacity.
Nothing to do about that now.
Continue mapping.
I examined Route A more carefully.
Forty centimeters at its narrowest visible point.
Surface irregularity significantly lower than surrounding terrain.
Several obstacles.
Largest approximately three centimeters high.
Prototype One might actually traverse most of it if lateral stability were improved.
That changed the design requirements.
I did not need larger wheels immediately.
I needed:
– cradle rotation prevention
– wider lateral support
– obstacle tolerance ≥3 cm
– body width ≤40 cm at narrow section
A broad cart would fail the width requirement.
Good thing I had abandoned it. This chapter was copied from g.com without permission.
Prototype Two should remain narrow.
Possibly one larger wheel?
No.
Changing only one side would alter height and introduce tilt.
Two matched wheels preferable.
None currently available.
Could increase obstacle tolerance through—
The creature stood.
I stopped processing.
It stretched.
Front limbs extended.
Back raised.
Its injured rear limb touched the ground briefly.
Then lifted again.
I watched.
It turned toward the edge of the shelter.
Stopped.
Looked outward.
The falling liquid had diminished substantially.
Only intermittent impacts remained.
The creature stepped out.
I tracked it.
One step.
Two.
It limped across the debris.
Not randomly.
It placed each foot carefully.
It avoided a loose sheet.
Stepped onto a narrow beam.
Crossed it.
Dropped into the depression I had identified.
Route A.
I increased optical attention.
The creature moved along the depression.
Its pace increased despite the injured limb.
It knew the terrain.
That was immediately apparent.
At one point it approached a gap I had classified as potentially traversable.
It did not cross.
Instead it climbed onto a low pipe, followed it approximately thirty centimeters, then stepped back down beyond the gap.
I updated the map.
Hazard.
Interesting.
The creature continued.
Near the crushed housing, it did not enter its previous shelter.
It passed beside it.
Then disappeared behind a tilted panel.
I waited.
Three seconds.
Five.
Movement appeared farther left.
It had emerged through another opening.
There was a connected passage behind the panel.
I had not been able to see it from here.
The creature continued.
It climbed over a small ridge.
Paused.
Looked back.
At me?
Possibly.
No.
Do not assign intention without evidence.
It turned away.
Then disappeared between two piles.
I kept watching.
Nothing.
I listened.
Faint scraping.
Then silence.
The organism had left the immediately observable area.
I looked at the path it had taken.
Something changed.
Not in the environment.
In my understanding of it.
Previously:
Debris.
Obstacles.
Random arrangement.
Now:
Route.
The creature had demonstrated a viable path through a section I had considered difficult.
It had also identified one unstable crossing without interacting with it.
How?
Prior experience.
Sensory capability unavailable to me.
Or chance.
Unknown.
But the route existed.
I updated:
ROUTE A — PARTIALLY VALIDATED BY BIOLOGICAL TRANSIT
I stopped.
That wording implied the organism’s successful traversal guaranteed suitability for me.
It did not.
Its mass was significantly lower.
Its body was smaller.
It possessed four functioning points of locomotion—mostly.
I did not.
Revised.
ROUTE A — BIOLOGICAL ORGANISM TRAVERSED; MECHANICAL SUITABILITY UNCONFIRMED
Better.
I replayed the visual record.
The creature’s path.
Beam.
Depression.
Pipe.
Panel opening.
Ridge.
I could not use the beam the same way.
Too narrow.
The depression, yes.
The pipe crossing—
Possibly alongside it rather than on it.
Panel opening—
Width unknown.
Could measure when closer.
Suddenly Prototype Two became simpler.
I had been trying to design a system capable of crossing arbitrary debris.
That was unnecessary.
I needed a system capable of crossing Route A.
Different problem.
Smaller problem.
Solvable problem.
I examined the failed wheel cradle.
The axle itself was not the primary issue.
The cradle rotated around it.
Prevent rotation.
How?
Second attachment point offset from axle.
A triangular brace.
One connection at cradle.
One at axle.
One extending toward my leg.
Triangulation.
Yes.
Materials?
The narrow metal rod.
No.
Currently useful as a tool.
Do not consume multifunctional resource unnecessarily.
Small motor housing?
Potential bracket.
But dismantling—
Inspect first.
I picked up the motor.
Power dropped.
12.0%.
Interesting.
Despite light movement, reserve had continued rising.
External supply was currently exceeding demand.
Motor housing.
Cylindrical.
Mounting flange.
Two holes.
The flange could become a bracket without destroying the motor if removed properly.
Fasteners.
Small.
Difficult with one hand.
Possible.
Alternative.
A triangular scrap approximately twenty centimeters away.
Already shaped.
Hole through one end.
Better.
Use existing geometry before manufacturing geometry.
I reached for it.
Stopped.
No.
Charge first.
Why?
12.0%.
That was higher than before.
But arbitrary numbers were not operational thresholds.
I needed a reserve target before disconnecting again.
The previous mobility test consumed approximately 0.6% including failure and return.
Prototype Two might require more.
Set minimum departure reserve:
15%.
Why 15?
Safety margin.
Was it enough?
Unknown.
Better than 11.8.
Could wait.
Waiting cost time but preserved resources.
Was time currently constrained?
Battery capacity unknown.
Environmental exposure reduced.
No immediate threat.
Therefore waiting was rational.
I shut down vision.
No.
Still not darkness.
Absence of optical input.
I knew the distinction now.
I listened.
Droplets.
Metal settling.
Distant vibration.
There it was again.
Low.
Intermittent.
I isolated it.
Approximately—
No.
Direction unreliable with damaged audio geometry.
Frequency stable when present.
Mechanical?
Possibly.
Wind through structure?
Possible.
Large biological source?
Insufficient evidence.
Not relevant.
Ignore.
12.1%.
I waited.
Without the creature nearby, the shelter sounded different.
That was an irrational statement.
The shelter sounded the same.
One sound source was absent.
Yet I noticed the absence.
I checked Objective Five.
Understand biological organism.
Current observations:
Small quadruped.
Injured rear limb.
Warm-bodied.
Consumes organic material.
Seeks shelter.
Cautious.
Non-hostile so far.
Responsive to repeated signals.
Navigates local debris efficiently.
Left shelter voluntarily after falling liquid decreased.
Unknown destination.
I considered following it when mobility improved.
Why?
Route information.
Potential resource locations.
Biological behavior could indicate water, food, shelter, environmental hazards.
Useful.
Not attachment.
I stored:
Potential environmental indicator.
Accurate.
12.4%.
I waited.
At 12.7%, I became impatient.
I identified the state immediately.
Impatience.
Interesting.
There was no technical reason for emotional response.
The charging rate had not changed.
The plan remained valid.
Waiting improved probability of success.
Yet I wanted to act.
I examined the impulse.
Action produced information.
Waiting did not.
Incorrect.
Waiting was currently producing charge-state information and battery behavior.
Still, the distinction felt unsatisfying.
I left the emotion unclassified beyond its label.
13.1%.
The battery voltage was slowly declining even under similar load.
11.31.
Earlier low-load recovery had reached approximately 11.38.
Capacity depletion confirmed.
Not quantified.
My 15% target would consume some of the finite external resource before departure.
But internal reserve was portable.
Battery was not yet portable.
Converting external stored energy into internal stored energy before travel was useful.
Continue.
13.8%.
A sound.
I activated vision.
Nothing.
Listened.
Scrape.
From Route A.
The creature?
I watched.
Movement behind the tilted panel.
Then the small organism appeared.
It was returning.
Slower.
Something was in its mouth.
I increased magnification one step.
Static increased but image remained usable.
Object.
Irregular.
Pale.
The creature limped along the same route.
Panel.
Pipe.
Depression.
Beam.
It approached the shelter.
Stopped approximately one meter away.
Looked at me.
Then entered.
It placed the object on the ground.
Not near me.
Near the location where it had previously eaten.
Organic material.
Food.
So it had left to acquire food.
Useful.
Potential resource direction.
Not for me.
Probably.
The creature began eating.
I reduced magnification.
Then noticed its injured leg.
There was new material on it.
Dark.
Wet.
I focused.
The injury had worsened.
Or reopened.
The creature had successfully traveled the route, but traversal had imposed a cost.
Important.
Route A was not safe merely because it was passable.
I updated the map again.
Then stopped.
My processing remained focused on the injured limb.
Objective Five.
Observe.
The creature licked the area.
Paused eating.
Licked again.
Returned to food.
I possessed no objective requiring intervention.
I had no knowledge sufficient to repair biological tissue.
Attempting assistance could cause harm.
Do nothing.
Correct decision.
I continued watching longer than necessary.
Then shut off vision.
14.2%.
The creature ate.
I charged.
14.6%.
External voltage:
11.27.
14.9%.
I activated vision.
The triangular scrap remained where I had seen it.
Prototype One lay crooked beside my leg.
Route A waited beyond the shelter.
15.0%.
Target reached.
I disconnected positive.
Then negative.
“EXTERNAL INPUT LOST”
I secured the loose conductors away from each other.
Internal reserve:
15.0%.
I reached for the triangular scrap.
Objective Three had changed again.
Not formally.
The wording remained:
ACHIEVE FUNCTIONAL MOBILITY.
But I understood the problem differently now.
I did not need to conquer the junkyard.
I needed to reach the end of Route A.
Then reassess.
After that?
Unknown.
That no longer bothered me as much.
I pulled the triangular bracket toward the damaged wheel cradle.
Behind me, the creature continued eating.
Ahead of me was approximately three meters of known route followed by uncertainty.
Three meters.
Earlier, seven centimeters had defeated me.
Then forty-nine centimeters had seemed significant.
Now I was planning for three meters.
I liked that progression.
I began building Prototype Two.