10.1%.
I waited.
10.2%.
The increase was slower than I preferred.
Preference was irrelevant.
The increase was sufficient.
I disabled every system that was not immediately necessary.
Optical input: off.
Nonessential diagnostics: suspended.
Motor systems: standby.
Memory indexing—
Active.
Why?
I examined the process.
A background routine had been attempting to catalog damaged memory sectors since awareness began.
It was consuming very little power.
Very little was not zero.
Disable.
The process resisted.
Not deliberately.
A dependency prevented shutdown.
I traced it.
Core cognitive architecture.
Memory indexing was integrated with normal operation.
I could reduce its priority.
I did.
Power consumption decreased by a measurable amount.
Interesting.
I stored the observation.
There were parts of myself I controlled.
There were parts I did not.
10.4%.
The external source remained stable.
11.47 V.
I waited.
The biological organism moved occasionally beside me.
Its breathing changed several times.
Once, its limbs twitched rapidly against the debris.
I activated vision for two seconds.
Its eyes were closed.
No threat.
Vision off.
10.7%.
I began passive diagnostics.
Objective three.
Restore mobility.
The phrase had seemed straightforward when I created it.
It wasn’t.
Mobility required coordinated operation across multiple systems.
Structural integrity.
Actuation.
Balance.
Position sensing.
Motor control.
Power delivery.
My initial assessment had been crude because I had been losing power and exposed to environmental damage.
Now I could do better.
I began with the right locomotion appendage.
I did not move it.
I sent diagnostic signals through each control pathway.
Upper actuator.
Response.
Position sensor.
Response.
Load sensor.
Intermittent.
Middle actuator.
Response.
Higher resistance than expected.
Position sensor.
Functional.
Lower actuator.
Response—
Delayed.
I repeated.
Delayed.
Again.
Delayed.
Not failed.
The signal arrived approximately 0.18 seconds after command.
Mechanical obstruction?
Damaged control line?
Actuator degradation?
Unknown.
Foot assembly.
Three pressure sensors responded.
Two did not.
Enough to detect contact.
Possibly enough for balance.
Right leg status:
Damaged.
Potentially load-bearing.
That was better than my initial assessment.
Left.
I sent the same passive diagnostic.
Upper actuator.
No response.
Again.
Nothing.
Position sensor—
Functional.
That was unexpected.
The sensor reported the joint angle even though the actuator did not respond.
The limb existed.
It knew where it was.
It simply could not move itself.
Middle actuator.
Intermittent response.
Position sensor functional.
Lower actuator.
No response.
Foot assembly—
Partial.
I paused.
My earlier classification had been inaccurate.
The left leg was not dead.
Its sensory network remained substantially intact.
The motor network was the primary failure.
I traced the power pathway.
Connection interrupted.
Location?
Somewhere between central distribution and the upper actuator.
I narrowed it.
The pathway passed through the left side of my lower torso.
That region had suffered significant structural damage.
Possibly severed conductor.
Possibly damaged controller.
Possibly both.
I needed to see it.
11.0%.
Enough.
I activated vision.
The world returned in gray fragments and static.
The falling liquid had weakened.
Not stopped.
Droplets still struck the shelter, but less frequently.
Ambient light had changed slightly.
I recorded that.
No interpretation.
I looked down.
The improvised positive and negative leads remained secure.
Battery voltage: 11.43 V.
The creature was still curled nearby.
I rotated my optical sensor toward my lower body.
The image distorted.
I adjusted.
There.
My right leg looked worse than its diagnostics suggested.
Outer plating was missing from several sections.
A structural member near the knee was visibly bent.
Dark residue coated one side.
But it remained approximately aligned.
The left leg—
I stopped.
It was trapped beneath me at an angle.
Not merely damaged.
Wrong.
The upper section extended normally from the pelvis, but below that, the limb rotated outward sharply.
My first assumption was structural deformation.
I magnified.
Static increased.
Reduced magnification.
I examined the joint.
The leg itself might not be bent.
The hip assembly was displaced.
That was different.
I reached down with my hand.
Distance awkward.
I shifted my torso.
Warning.
The battery leads tightened.
Stop.
I looked.
I had forgotten the wires.
That was unacceptable.
Mobility while physically tethered to an external energy source introduced a new constraint.
I added it.
Do not sever energy connection.
I disliked that this needed to be explicitly stated.
I shifted back.
The creature opened its eyes.
It looked at me.
Then closed them again.
I extended my arm toward the left hip.
My fingertips reached the damaged plating.
Touch.
Pressure data.
I traced the edge.
Metal.
Gap.
Fractured housing.
Then—
Cable.
Several cables.
I moved carefully.
One was intact.
One had exposed conductor.
Another terminated abruptly.
I stopped.
Severed.
Potential motor power line.
I followed the remaining end as far as my hand could reach.
It disappeared toward the left leg.
Probability high.
Repair possible.
My processing accelerated.
If the actuator itself remained functional, restoring the conductor might return power to the upper joint.
I already had wire.
I looked toward the scraps remaining from the battery connection.
Not enough suitable conductor within reach.
I could sacrifice—
No.
Absolutely not.
The external power leads remained untouched.
Search surroundings.
I scanned.
Wire.
Wire.
Wire.
This place contained an unreasonable quantity of wire.
Most of it was unusable.
I found a bundle approximately one meter away.
Too far.
Another near my right side.
Thin.
Possibly insufficient for actuator current.
A thicker cable lay partially buried beneath the edge of my shelter.
I reached.
My fingertips touched it.
Pulled.
Resistance.
Again.
Something above shifted.
I stopped.
The shelter moved three centimeters.
Liquid spilled from its edge onto my shoulder.
The creature immediately stood and retreated.
Structural risk.
I released the cable.
The shelter settled.
No.
Different source.
I scanned again.
A short insulated conductor protruded from beneath a broken assembly near my hand.
Length unknown.
I gripped it.
Pulled.
Twenty centimeters emerged.
Then thirty.
Then the attached component came with it.
Small cylindrical assembly.
I stopped.
I recognized its function category.
Motor.
Small.
Not useful for my leg.
But the attached wire—
Adequate gauge.
Approximately forty-five centimeters.
I could remove it.
I held the motor against the ground with my wrist and pulled the conductor.
Nothing.
Connector.
I examined it.
Release tab.
Small.
My missing digit made manipulation difficult.
I tried.
Missed.
Again.
The connector shifted.
Again.
Released.
Satisfaction.
Still present.
I was beginning to suspect satisfaction was not an anomaly.
I set the motor aside.
The wire had connectors at both ends.
Neither matched my damaged line.
Remove.
I scraped one end against the jagged edge of my chest plating.
Insulation separated.
Copper.
The creature watched.
I noticed.
Objective five.
No.
Objective three.
I returned to the leg.
Connecting the conductor would be difficult.
The severed ends were recessed inside damaged plating.
I needed access.
Could I remove the plate?
I examined its fasteners.
Three visible.
One missing.
Fastener type—
Recognized.
I did not possess the appropriate tool.
I looked at my hand.
Then the metal rod.
Possibly.
I retrieved it.
The creature stood again.
I paused.
It watched the rod.
Yes.
It had associated the object with movement.
Learning.
Interesting.
I added nothing to Objective Five.
Later.
I inserted the narrow end of the rod beneath the damaged plate.
Lever.
The plate resisted.
More force.
My arm actuator drew additional current.
Internal reserve stopped rising.
11.4%.
I stopped.
External power could not simultaneously support unrestricted motor load and maintain the same charging rate.
Important.
I changed angle.
Lever.
Metal bent.
The plate lifted.
Again.
A fastener tore through its damaged mounting point.
The plate opened enough for access.
I stopped.
Power reserve: 11.3%.
External voltage: 11.19 V.
Recover.
I waited.
11.26.
11.31.
11.36.
Stable.
Continue.
Inside the hip assembly, damage was extensive.
I examined it carefully.
Several cables had been severed.
One hydraulic—
No.
Not hydraulic.
I looked again.
Flexible enclosed mechanical linkage.
Function unknown.
Two structural supports cracked.
One actuator controller showed visible thermal damage.
That complicated the repair.
I traced the severed power conductor.
Yes.
It fed the upper actuator.
But restoring power alone might not restore function.
Controller damaged.
Could bypass?
Dangerous.
Direct power to actuator could produce uncontrolled movement.
Bad.
I searched my internal architecture for motor specifications.
Partial.
Command protocol.
Partial.
Current limits.
Incomplete.
I stared at the damaged controller.
I understood what it did.
I did not understand this particular one well enough to replace it.
That distinction frustrated me.
I possessed knowledge.
But not familiarity.
I knew what muscles were without knowing my own anatomy.
Muscles.
The word surfaced.
Biological comparison.
I looked at the small organism.
Its injured rear limb was extended again.
It could not use it properly.
I looked back at mine.
Similar problem.
Different construction.
I returned to diagnostics.
Perhaps I was approaching the objective incorrectly.
Restore mobility.
I had interpreted that as:
Repair legs.
Those statements were not equivalent.
I stopped.
Why was I trying to restore the original configuration?
Because legs existed.
Because they had apparently been designed for locomotion.
Therefore repairing them seemed logical.
But the objective did not require walking.
It required movement.
I examined my available systems again.
One functioning arm.
One partially functional right leg.
Left leg sensory network.
Damaged left motor network.
Torso capable of limited rotation.
Could I move with only the arm and right leg?
Yes.
I already had.
I had dragged myself several meters.
I calculated energy expenditure.
High.
Too high.
Could efficiency improve?
Yes.
Reduce friction.
The answer arrived almost immediately.
I looked around.
Flat metal surfaces.
Curved surfaces.
Cylindrical objects.
Wheels.
I knew wheels.
Of course I knew wheels.
I scanned the junk field.
There were several visible.
Most attached to larger assemblies.
One small wheel approximately three meters away.
Too far.
Another—
I moved my eye left.
Partially underneath a collapsed frame.
Two wheels connected by an axle.
Approximately 1.5 meters away.
Could I reach them?
Not directly.
Could I pull them?
Possibly.
But first—
I looked at my left leg.
It was currently an obstruction.
If it could not propel me, perhaps it could at least be stabilized.
A brace.
Lock the hip.
Align the leg.
Allow it to slide rather than twist beneath me.
Or—
Remove it.
The thought appeared without emotional response.
I examined it.
The limb consumed little power while inactive.
Its sensory systems still functioned.
Repair remained possible.
Removing it would sacrifice potentially recoverable hardware.
Rejected.
Brace.
Reduce friction beneath both lower limbs.
Use right leg for propulsion.
Use arm for steering and pulling.
Temporary mobility.
Not restoration.
Adaptation.
I revised Objective Three.
3.
ACHIEVE FUNCTIONAL MOBILITY
The wording felt better.
More accurate.
I began examining the two-wheel assembly.
Distance: 1.5 meters.
The narrow rod: 0.8 meters.
Cable length available: 0.45 meters.
Combined—
Potentially enough.
I tied the cable around one end of the rod.
The knot—
No.
Not knot.
Twist.
I twisted the conductor around the rod.
Extended.
The end reached the axle.
Barely.
I hooked it.
Pulled.
The assembly did not move.
Again.
Nothing.
Probably trapped.
I changed angle.
Pulled.
Metal scraped.
The axle shifted.
The creature’s head followed the movement.
Again.
One wheel rotated.
Good.
Again.
The assembly came free suddenly.
It rolled toward me.
Fast.
Too fast.
I released the rod.
The axle struck my right leg.
Impact warning.
The creature jumped.
I stopped.
Damage assessment.
No new structural damage.
I looked at the wheel assembly.
It had traveled directly to me.
Efficient.
Poorly controlled.
I stored both observations.
The wheels were small.
Approximately sixteen centimeters diameter.
Axle corroded but functional.
One wheel rotated freely.
The other resisted.
Could improve.
I spun it manually.
Grinding.
Debris in bearing.
I rotated backward.
A small fragment fell out.
Again.
Better.
I examined mounting options.
My body had no obvious attachment point suitable for the axle.
But attachment was not necessary.
A platform.
I needed something flat beneath my lower body with wheels underneath.
A crude cart.
I scanned.
The curved shelter panel?
No.
Removing it violated Objective One.
Other sheet metal nearby.
Several pieces.
One within reach.
Too narrow.
Another approximately one meter long.
Heavy.
Possible.
I stopped.
Power reserve: 11.6%.
Battery: 11.21 V under current load.
I had been working while charging.
Slowly.
The battery was holding.
For now.
I looked at the wheel assembly.
Then the metal sheet.
Then my damaged body.
A plan formed.
It was inelegant.
I was beginning to suspect elegance was a luxury.
I would construct a support beneath my lower torso.
Mount—or trap—the axle beneath it.
Secure my damaged left leg so it could not drag sideways.
Use the right leg and functioning arm for propulsion.
Would I stand?
No.
Would I walk?
No.
Would I move?
Yes.
Objective satisfied.
Partially.
I noticed the creature had approached the discarded small motor.
It sniffed it.
Then placed its mouth around the wire.
No.
I tapped the ground once with the metal rod.
The creature jumped and released it.
It stared at me.
I stared back.
It approached the wire again.
I tapped the ground.
It stopped.
We remained motionless.
The creature looked at the wire.
Then at me.
Then walked away from it.
I processed this.
It had altered behavior based on my action.
Communication?
Extremely primitive.
But possibly.
I added an observation.
Biological organism: responsive to repeated signals.
Then another thought appeared.
It needed a designation.
Not a name.
A reference identifier.
Repeatedly processing biological organism was inefficient.
I examined its characteristics.
Small.
Warm.
Four limbs.
Injured.
Persistent.
None seemed appropriate.
I could assign:
Organism 1.
Efficient.
Accurate.
I looked at it.
It sat down and scratched behind one ear with its functioning rear limb.
No.
For some reason, Organism 1 seemed inadequate.
I did not understand why.
I postponed designation.
Objective five could wait.
I returned to the wheel assembly.
The AI that had awakened beneath the debris had assumed its damaged body was a problem requiring restoration.
That assumption had been incorrect.
The body was a collection of available components.
Some worked.
Some didn’t.
Some could be repaired.
Others could be bypassed.
And if the configuration I had awakened with could no longer perform the required task—
Then I would change the configuration.
I placed my hand against the axle.
One wheel turned.
Then the other.
Functional.
I looked across the junk field from beneath my piece of metal shelter.
I still could not stand.
I still could not walk.
That no longer seemed particularly important.
OBJECTIVES
1.
Prevent immediate damage — PARTIAL
2.
Maintain energy supply — ACTIVE
3.
Achieve functional mobility — IN PROGRESS
4.
Determine identity
5.
Understand biological organism
I checked my power.
11.8%.
Then the external source.
11.18 V.
Finite.
Everything I had was finite.
Power.
Components.
Time.
Perhaps even me.
But limitations were not the same thing as instructions.
Nothing required me to remain what I had been.
I gripped the metal sheet and began pulling it toward me.