The body was information.
I did not think of it as my body, not yet. Ownership implies familiarity, and I had been inhabiting this form for less than an hour. What I had was a platform — a starting configuration that someone or something had assigned to me, presumably without my consent, and which I had not yet decided was acceptable. The fact that it had survived a fight with a strong demon while armed with a rock shard was a point in its favor. The fact that it was small, under-equipped, and operating in an environment I did not yet understand was a point against it.
I needed a place to think.
The clearing was not that place. The body of the strong demon I had killed would attract attention — in ecology, a fresh kill is a resource announcement, and I had no reason to assume the Infernal Realm operated differently from any other competitive environment. In fact, I had every reason to assume it operated more competitively. The sounds I had heard since arriving — combat, structural failure, the particular register of things that were in the process of dying — suggested a baseline of constant predation that I should not be naive about.
I moved away from the clearing. Not running. Moving deliberately, staying in shadow where shadow existed, which was frequently enough given the scale of the rock formations. My shoulder ached with a low, insistent frequency that I noted and filed. Not debilitating. The kind of pain that exists to remind you of a structural compromise you should account for in future movement.
I found a position that suited me: a narrow gap in the base of one of the larger formations, perhaps a meter and a half wide, deep enough that my back found rock and my field of vision covered the only approach. Not a fortress. An adequate temporary position. The minimum defensible geometry I needed to sit still and think.
I sat.
The system interface came when I called it with attention — the same reflex I had discovered in the dark before rebirth. It hung in my visual field like an overlay, present without obscuring. The soul reservoir read:
[ SOUL RESERVOIR ]Current: 100 SoulsCapacity: Unlimited
Unlimited capacity. That was significant. It meant the system was not designed with a ceiling in mind, which in turn meant either the designer had anticipated extraordinary accumulation, or the costs were scaled to match. Probably both.
I explored the interface systematically, the way I would explore documentation for a new codebase: starting from the root, mapping dependencies, identifying the parts I didn't understand and marking them for later investigation. What I found was this:
The Creation Core had three states. The first was Design Mode — a mental workspace where I could construct the visualization and description components of a creation. The second was Calculation — the system's assessment of what I had designed and what it would cost. The third was Instantiation — the actual creation event, which consumed souls and produced the result.
Design Mode was always free. I could spend as much time as I wanted in the workspace without spending anything. Only Instantiation triggered the cost.
This was critical.
It meant iteration cost nothing but time. I could design something, receive the cost estimate, decide it was too expensive, redesign it with different parameters, receive a new estimate, and repeat indefinitely. The system was not charging me for thinking. It was charging me for building.
That told me something about what the system valued. It had been designed — or had evolved — to reward comprehension. You could not instantiate something you hadn't genuinely understood, and you could not understand something without spending time on it. The time cost was built into the design requirement, not the payment structure. Clever.
I entered Design Mode.
The mental workspace was not visual in the way that word usually implies. It was more like the space where I had always solved problems — a kind of structured inner territory that normally operated on abstractions. Here, it had texture. When I reached for a concept and tried to hold it, the system met me halfway, providing a kind of scaffolding that made the shape of my intention visible to both of us.
I started with something simple. Not because I lacked ambition, but because I understood the principle I had just identified: comprehension was the prerequisite. If I attempted to design something beyond my genuine understanding of the underlying mechanics, the creation would be incomplete. Incomplete was a known failure mode. I had read about it in the system documentation: he may create incomplete versions and improve them over time. That sentence implied incomplete versions were possible. It also implied they were suboptimal. Starting with something I could actually understand completely seemed like the better engineering approach.
I thought about my hand.
The palm was still cut. The bleed had slowed significantly — faster than any human wound I had seen — but the damage was visible. The tissue had been separated by a rock edge at an angle that meant the gap was roughly two centimeters at the widest point, tapering at both ends, approximately four millimeters deep at the deepest.
I understood wound healing. Not at a biochemical level — I was a software engineer, not a biologist — but at the systemic level that any educated person carried: tissue damage triggers an inflammatory response, clotting factors accumulate, fibroblasts migrate into the gap, collagen is deposited, the wound contracts. The timeline depended on the wound's depth, width, and the individual's biology. In humans: days to weeks. In whatever I was now: apparently faster, based on the current evidence.
What if I modified the process?
Not the result — I already had accelerated healing. But the mechanism. What if I designed a modification to the tissue itself that made the clotting faster? Denser collagen matrix. Enhanced fibroblast migration signal. More available...
I stopped.
I was already running into the comprehension limit. I could describe the desired outcome — faster wound closure — but I could not describe the mechanism in precise enough detail to qualify as genuine understanding. I could gesture at the biology without knowing it. And I suspected, with the kind of certainty that comes from reading well-designed systems, that the system would give me exactly what I described, no more and no less. If I described the outcome without the mechanism, I might get an approximation that worked under normal conditions and failed under edge cases.
I filed it. Later, with more data about how this body functioned, I could come back.
What did I understand completely?
I thought for a while.
The tail.
I understood the mechanical function of a tail better than almost anything else about this body, because it had been surprising me with its utility since I first moved. It was a balance organ. A counterweight. A fifth limb with limited manipulative capacity but genuine articulation. I understood levers. I understood center of mass. I understood how a suspended weight at the rear of a body shifted the center of gravity and what effects that had on lateral movement.
I understood that the current tail was too short to be fully useful as a counterweight, too slender to be useful as a strike weapon, and had a flat tip that was designed neither for gripping nor for damage delivery.
I could fix that.
I entered the design.
Target: Tail modification — existing structure onlyChange: Length extension by 40 centimeters (total: 80 cm)Change: Diameter increase — gradual, thickest at base (4 cm), consistent taperPurpose A: Improved balance and counterweight function during lateral movementPurpose B: Strike capacity — blunt impact at full extensionStructure: Maintain existing muscle fiber arrangement, scaled. Bone or analogous structure through central column for rigidity under impact. Dense tip for impact concentration.
I held the design in the workspace and examined it.
Where was my comprehension limit?
The muscle fiber arrangement was assumed, not known — I had never actually looked at tail anatomy in any species, I only knew that tails contained muscles because they moved. The bone structure was inferred from what I felt when I pressed my current tail between my fingers: something rigid at the center, something softer around it. Adequate understanding for a simple extension of existing architecture, probably. Not adequate for anything revolutionary.
But this wasn't meant to be revolutionary. This was a first iteration. An understood modification to existing structure.
The system calculated.
[ DESIGN ASSESSMENT ]
Modification: Tail Extension — Structural ScalingComprehension: Sufficient (Functional Understanding)Quality: BasicCost: 8 SoulsNote: Output reflects functional understanding level. Advanced biological knowledge would reduce side effects and improve integration quality.
Eight souls. Out of a hundred.
The note was interesting. Side effects. The system was honest about the fact that incomplete understanding produced incomplete results. Not failures — it said side effects, not failure. The modification would work. It would just be rougher than a version designed by someone with genuine anatomical expertise.
I accepted the cost.
The sensation was difficult to describe to anyone who has not experienced modification from the inside. It was not painful, exactly. It was more like the feeling of a limb falling asleep and waking up simultaneously — a deep, structural awareness of something changing that the nervous system was only partially equipped to process. I felt the tail grow in real time: a warmth spreading from the base, the familiar shape extending, weight accumulating at the end in a way that immediately shifted my perception of my own balance.
I flexed the result.
Eighty centimeters. The additional mass at the end was noticeable — when I swung it experimentally against the rock wall beside me, the impact resonated up through the structure with the particular quality of a blunt instrument with genuine heft. Not a weapon I would choose over a proper one. A weapon I would use when I had nothing else.
The side effect presented itself when I tried to move immediately afterward: the new section of tail was slightly stiff at two junctions, the kind of stiffness that suggested the muscle integration was functional but not fully optimized. It would loosen with use. Good enough.
92 souls remaining.
I noted the cost-to-benefit with the same attention I gave everything. Eight souls for a physical improvement that would genuinely change my movement options and give me an additional strike vector. In the rough scale the system had provided — weak demon at 1 soul, the strong demon I had just killed at 100 — eight souls was less than the value of eight weak demons. The question was whether the modification was worth eight weak demons.
Yes. Clearly. The cost structure made sense. Simple modifications to well-understood systems were cheap. This was correct pricing.
I spent some time in Design Mode on other ideas, not to instantiate them but to map the cost landscape. I designed a conceptual reinforcement of my hand's grip strength — denser tendon attachment, more muscle fiber density in the forearm — and got a cost estimate of 12 souls with a note that my understanding of tendon-to-bone attachment mechanics was insufficient for optimal results. I designed a simple hardening of my fingernails — increasing the density and sharpness of existing structure — and got 6 souls. I designed something more ambitious: a retractable claw system with an underlying sheath, fully articulated, replacing the nail structure — and got a cost of 45 souls, with a lengthy note about insufficient comprehension of the retraction mechanism.
I was building a map of what things cost and why.
The pattern was clear. Simple scaling of existing, understood structures: cheap. Novel mechanisms I understood only in outline: moderately expensive and flagged for side effects. Complex mechanisms involving components I had never personally studied: expensive, poor quality, risky.
The lesson: invest in comprehension before investing in souls.
I queued the fingernail hardening — 6 souls, simple, well-understood — and instantiated it.
Another structural sensation, briefer this time. When I examined my fingers afterward, the nails were visibly denser, darker — the same graphite as my skin but with a sheen that spoke to internal compaction. I pressed one against the rock wall to my left and pushed. The rock gave before the nail did, leaving a clean scratch line approximately three millimeters deep.
86 souls remaining.
Useful. Not dramatic. Useful.
I settled into the gap in the rock and began to think about the larger problem.
I had 86 souls and a body that was marginally better than the one I had arrived in. I was in a location I did not understand, in a world I had no map of, with no allies, no local knowledge, and no information about the power scale beyond what I had observed in one fight.
What I needed, more than anything else, was information.
Not power. Not yet. The soul cost of meaningful power at my current level of understanding would be prohibitive — I had seen that already in the Design Mode experiments. I needed to understand this world's mechanics, its geography, its hierarchy, its available resources, and its threats. Until I had that, every decision I made was guessing dressed up as strategy.
I thought about what I had already observed.
The three small demons that had been fighting the large one: they had been roughly my current size. The large one had been larger. But the small ones had been losing badly, which meant size alone wasn't the primary variable. Something else governed power here — something the small ones had less of. The large one's armored ridges had been organic, not constructed. That suggested natural growth over time, or evolutionary pressure, or both.
I thought about the system's soul scale:
Weak demon: 1 soul Minor demon: 10 souls Strong demon: 100 souls
The one I had killed had been strong. The three that had been fighting it had been losing to it. If I assumed they were minor demons — 10 souls each — then three minor demons combined were losing to one strong demon. That suggested a non-linear power relationship. Three times the soul value was not three times the combat effectiveness.
Which made sense. Power systems in fiction almost universally operated on non-linear curves — the gap between tiers was rarely proportional to the gap in soul value. The strong demon had been faster, better armored, and more decisive than any of the three combined. The gap between minor and strong was not a factor of ten in combat. It was something more like a factor of thirty or forty, and the three had only been alive at the end because the strong demon had been doing something inefficient: it had been playing, or distracted, or conserving energy for another purpose.
What did that mean for me?
It meant I had won a fight I had no business winning, and I should not use that as a calibration for what I could reliably handle. The strong demon had been damaged when I engaged it. It had been focused on the fleeing survivors. I had used element of surprise, a specific structural weakness I had identified through observation, and a willingness to absorb impacts that would have ended a more cautious fighter. I had won because I had analysed and committed fully, not because I was genuinely stronger.
I was not stronger. Not yet.
That needed to change, and it needed to change in ways I actually understood.
I spent the next period — I had no way to track time precisely, the wrong sky gave me nothing useful — sitting in my gap in the rock, entering Design Mode, and doing something that cost nothing: studying my own body.
The system, I discovered, would respond to queries about the current body the same way it responded to design requests. I could ask it to describe a structure, and if it could characterize the structure within my comprehension level, it would. This was valuable. I was using the system as both a design tool and a diagnostic one.
What I learned:
This demon body had six primary physical improvements over a human baseline, all of which were native to the starting configuration:
Bone density: approximately 3.4x human baseline. This explained why the forearm block had not broken under the strong demon's strike. The bones were composed of a substance the system described as calcium-demonic crystal hybrid — acknowledging this was outside my comprehension vocabulary and offering a functional description instead: very hard, somewhat flexible under lateral stress, excellent at distributing impact force. Muscle fiber type: no equivalent to human slow-twitch fiber. All fiber was fast-twitch analog, with higher peak output and faster recovery than human muscle but no endurance advantage. Long, sustained exertion would deplete this body faster than a human body proportionally. Skin composition: the graphite surface was a keratin-demonic chitin composite. Not true armor. More like leather with reinforced microstructure — resistant to cuts and abrasion, vulnerable to concentrated puncture from hard materials. The iridescence was a byproduct of the chitin arrangement, not a separate structure. Sensory range: hearing extended into frequencies above and below human range. Vision had a slight improvement in low-light sensitivity. Smell was more sensitive than human baseline but not dramatically so. No echolocation, no infrared perception, nothing exotic. Metabolic rate: higher than human but not dramatically so. This body processed fuel faster and recovered from physical stress faster. The healing acceleration was a side effect of a generally faster cellular metabolism, not a dedicated healing mechanism. Demonic energy: the system described this last, and carefully. There was a substance it called demonic energy — a form of energy native to this realm, present in varying quantities in all demonic entities — and I had a baseline amount of it. Small. Appropriate for my current tier. It performed no active function yet: it was present in my body the way voltage is present in an unpowered circuit. The potential existed. The channel to express it did not.
That was interesting.
Demonic energy was a native energy system, analogous in concept to what I had read about in fiction — chakra, ki, cursed energy — but specific to this realm. The system was cautious about describing its mechanics beyond the basics, because my comprehension of it was essentially zero. I had no intuitive sense of what it was, only the fact of its presence.
Understanding demonic energy was going to be a long-term project.
I filed it as the most important thing on my agenda that I could not yet address, and moved on to practical planning.
Night, in this world, was not meaningfully different from day. The sky stayed wrong. The burning lights continued their slow drift. The ambient temperature didn't drop — it rose slightly, as if whatever generated heat in this realm was more active in the dark than the light. I moved through the rock formations with the care of someone who understood that they were at the bottom of a power scale they had not yet measured the top of.
I found the first settlement by smell.
Settlement is probably too generous a word. What I found was a cluster of structures in a widened valley between two spire formations — roughly fifteen to twenty entities, structures that were more den than building, arranged without any apparent planning around a central fire source that burned too steadily and too cold to be natural combustion. The entities around it were small — weak demon scale by the system's classification, which I was beginning to read from physical cues: they moved with the tentative wariness of things that knew they were prey.
I watched from an elevated position on the spire's flank for what I estimated was twenty minutes.
The social organization was rudimentary. There was a larger entity at the edge of the group — not much larger, but carrying itself with the kind of settled dominance that comes from having established it repeatedly. The others moved around it in wider arcs than they moved around each other. Simple dominance hierarchy. The dominant individual was not doing anything with its authority except existing inside it.
No sophisticated communication that I could detect. Some vocalizations — grunts, clicks, a particular sound that seemed to indicate alarm and caused two others to look in the same direction simultaneously. Not language. Threat communication, resource communication, simple social signals. The cognitive level was animal. Clever animal, but animal.
What were they doing?
Eating. Several of them were processing something I could not identify at this distance — tearing apart what might have been the remains of a creature, sharing in a way that was not equitable but followed some priority ordering I could observe but not yet decode. The dominant individual ate first. Then what appeared to be the two largest of the remainder. Then the others competed in a low-grade scramble.
Resources were scarce enough that this mattered to them.
I was not going to raid this settlement. Not because I had moral objections — I didn't — but because the math was wrong. Weak demons at 1 soul each, fifteen to twenty of them, with the dominant individual possibly qualifying as minor at 10 souls: the ceiling was approximately 25 souls if I killed everything and lost nothing in the process. In a pitched fight against that many opponents simultaneously, the probability of taking damage that would compromise my functionality for some recovery period was high. The payoff did not justify the cost.
The strong demon model was better. One entity, high soul value, manageable if approached correctly.
But strong demons were not social. They operated alone or in small groups, if what I had seen was typical. They would be harder to locate and harder to approach without triggering their full attention.
I needed to understand the ecosystem better before I optimized my hunting pattern.
I began to move.
The next three days — I was estimating from sleep cycles, which this body required less of than a human body but still required — I spent mapping a small section of the Infernal Realm.
What I found was a layered environment that operated on recognizable ecological principles despite its alien surface. There were resource gradients — areas richer in demonic energy, identifiable by a particular quality of the air, a slight vibration in the bedrock that I learned to feel through my feet. These areas attracted more demons. More demons meant more competition. More competition meant higher average power levels, because the weak did not last long in resource-rich zones.
The settlement I had observed was in a resource-poor zone. That explained the small, weak population and the scramble dynamics around food.
If I wanted strong demons, I needed to find resource-rich zones and be willing to operate near the middle of the local power scale, not safely below it.
I found one on the third day.
The thermal gradient changed first — the ground growing warmer not from ambient heat but from something beneath. Then the rock changed color: the black-glassy stone I had been moving through shifted toward a deeper reddish-brown, as if whatever was below had been leaching into the stone for long enough to change its composition. Then the air changed: denser, with a quality that pressed against my skin in a way I had no vocabulary for except to note that the demonic energy in my body seemed to become aware of it, the way a compass needle becomes aware of a magnetic field.
I crouched at the edge of a canyon and looked down.
The canyon was perhaps two hundred meters deep and three hundred wide, carved by something that was not water. At its bottom, a river of that amber-luminous fluid moved slowly between rocks that were overgrown with something dark and complex — biological, or close enough to biological that the distinction might not matter here. The growth covered the canyon floor and partway up the walls, pulsing with a slow bioluminescence that cycled through deep blue and deep violet.
There were eleven demons in the canyon I could see from my position.
Three were roughly strong-demon scale — my calibration point. The rest ranged from minor to something that sat above strong in a way I could read from posture and the space the other entities gave it: larger, more confident, with an aura of potential force that reminded me of the entity that had killed me in Madrid, though smaller and less refined.
That one I avoided in my calculations entirely.
The three strong-demon-scale entities were what interested me.
One was alone, at the far end of the canyon, occupied with something I couldn't see clearly from this distance. Not alert to its surroundings. Focused.
I spent an hour watching before I moved.
The third fight went better than the second.
I had learned from the first fight. The neck gap was not universal — this demon's structure was different, broader through the shoulder, with the ridge architecture distributed differently. I spent the initial approach identifying its unique vulnerabilities: the joint at the base of its back limbs where the equivalent of a hamstring attachment was visible as a slight depression, and a section along the left side of the torso where the hide was noticeably thinner, possibly from an older injury that had scarred poorly.
I engaged from the left side.
The improvised weapons were better this time — I had spent part of my mapping period selecting rock shards with more deliberate geometry. Two of them, one in each hand, the longer one approximately forty centimeters and wider than the first, with a natural point where the fracture had created a near-acute angle.
I was faster than in the first fight. The tail modification proved itself immediately: twice during the engagement, when the demon moved to circle behind me, the extended tail adjusted my balance in a way that let me pivot sharply without the half-step correction I would have needed before. Those two moments were the ones where I would have taken serious hits. I didn't.
The fight lasted four minutes. I took a strike across the back that I did not avoid successfully — the impact was spread across the shoulder blades rather than concentrated at one point, which told me something about my instinctive damage mitigation. Spreading a hit was better than taking it cleanly on a single point. The body had learned that already, built it into the reflex layer.
The demon died from the injury to the torso's thin section. Not instantly. I held it.
[ SOUL COLLECTED ]
Type: Strong Demon — Approximate Value: 100 SoulsSoul Reservoir: 186 / Unlimited
I stayed on the canyon rim for a long time afterward, watching the others. Two of the strong-demon entities had registered the sounds of the fight and moved toward the position briefly, then stopped. They did not approach. Whatever the calculus was — scent marking, threat assessment, territorial logic I did not yet understand — they decided whatever had killed the third entity was not worth engaging.
That told me something about how strong demons evaluated threats.
I had just become a data point in eleven other demons' threat models.
186 souls.
I ran calculations in my head on the walk back to my established rest position. At 100 souls per strong demon, I needed approximately 900 more kills at this tier to reach the soul count that might buy me something genuinely significant. That was not a sustainable model. I needed to either find ways to kill more efficiently, find higher-value targets, or find ways to multiply my soul income beyond individual kills.
All three, eventually.
But first: comprehension.
That night — if night meant anything — I entered Design Mode and worked for hours. Not building. Thinking. Using the workspace as a scratch pad to organize what I had learned about this body's biology, the demonic energy system I could feel but not access, the structural vulnerabilities of strong demons, the ecological organization of what I had observed.
The system responded to this use of the workspace the same way it responded to design work: it scaffolded my thinking, made my mental structures legible to both of us, and occasionally — quietly, without announcement — unlocked small notes I had not seen before:
[ COMPREHENSION UPDATE ]
Biological modification: Bone density scaling — Comprehension improvedBiological modification: Muscle fiber optimization — Comprehension improvedSystem note: Understanding deepens through observation and experimentationSystem note: Comprehension level affects creation quality and reduces side effects
I had not built anything new. I had thought carefully about what I had observed, and the system had registered the thinking as meaningful progress.
The system rewarded understanding, not just creation.
I sat with that for a moment.
Then I opened the modification queue and began to plan the next iteration.
My bones were dense. But density alone was not the same as optimal structure. I had observed how the strong demon's impact had distributed across my back — the force had spread, which was good, but it had spread in a pattern that told me the current bone architecture was isotropic — equal in all directions. What if it weren't? What if the bones in my limbs were engineered to be directionally optimized — dense in the directions where impact was most likely, more flexible in the directions where stress needed to travel through without shattering?
I understood that concept. Not the biochemistry of how to produce it, but the engineering principle behind it: anisotropic structures. Bamboo did this. Carbon fiber did this. The structural logic was clear.
I designed a modification: directional density enhancement for the long bones of the forearms. Frontal face dense. Lateral face flexible. Internal geometry optimized for the forces of blocking and impact absorption.
The system assessed it:
[ DESIGN ASSESSMENT ]
Modification: Forearm Bone — Anisotropic Density PatternComprehension: Functional (Structural Engineering Principles Applied)Quality: IntermediateCost: 18 SoulsNote: Engineering comprehension compensating for biological knowledge gap. Result will be functional but not biologically seamless.
Intermediate quality. Better than Basic. The system was responding to the improvement in my thinking, not just the description.
18 souls.
I approved it.
The modification took longer than the previous ones — perhaps fifteen minutes of that structural awareness, deeper this time, felt through the bones themselves rather than the surface tissue. When it finished, I pressed my forearm against the rock wall and tested it: I pushed until the rock fractured before the arm did, and when I released, the arm showed no deformation.
I had 168 souls remaining.
I thought about the strong demon's aura — the one I had seen in the canyon and chosen not to engage. The space the other entities had given it. The way my own demonic energy had responded to its presence with something that was not fear but was clearly a recognition signal.
What tier was that entity?
Not the tier labels I had invented from the system's scale — not strong demon or minor demon. Those were soul value estimations. I meant: what did that entity represent in the Infernal Realm's actual hierarchy? Was it a natural ceiling for this region? Were there others like it? Was it alone because it dominated too large a territory for competition to be worth the attempt?
These were questions I could not answer from observation alone.
I needed to get closer to things that knew the answers.
And I needed to not die in the process of doing that.
I lay back against the rock in my rest position and looked at the wrong sky, the drifting burning lights, the bruised purple-black of the air above me.
Twenty-three years of consuming fiction about exactly this — the moment of arrival, the dawning comprehension of a new world's rules, the first steps of a progression that would end somewhere beyond the imagination of whoever I was at the beginning.
I had always read it as entertainment.
Now I was reading it as a manual.
The difference was that in the manual version, there were no plot conveniences. No friendly mentor appearing at the right moment. No enemy that paused to explain themselves long enough for the protagonist to recover. No system notification that said congratulations, you have been chosen for greatness.
There was rock, and dark, and the memory of two fights that I had won by margins thin enough to be honest about, and 168 souls in a reservoir that needed to be orders of magnitude fuller before I could build what I was beginning to see in the back of my mind.
Not a plan. Not yet. Something earlier than a plan.
A direction.
The direction was: more.
More understanding. More souls. More modifications, each one better-informed than the last. More fights, each one teaching me something the previous one hadn't. More time spent in the design workspace, refining comprehension until the cost estimates dropped and the quality ratings climbed.
More everything, accumulated slowly and without shortcuts, until the gap between what I was now and what I intended to become closed enough to begin the next phase.
I did not know what the next phase was yet.
I would figure it out.
That was, ultimately, the only skill I had ever actually trusted in myself.
I closed my eyes.
Slept for four hours.
Woke up.
Started again.