Chương 22: Find Something harder to do.

The day produced 190 XP. He knew this because the Nexus maintained a running daily total in the corner of his HUD that he'd started consulting at the end of each day the way he used to consult the project completion tracker in Sacramento — with the specific relationship you have with a number that tells you the truth even when the truth is unrewarding.

The 190 sat there in its modest blue notation and he looked at it and then looked at the Level 6 bar at 1, 250/1, 500 and did the arithmetic that the arithmetic kept producing. At 190 XP per day, Level 6 was fourteen to eighteen days out. Depending on daily variation. Depending on whether he found something that generated more XP than the established circuit.

Depending on nothing, if the established circuit was all he ran — in which case it was sixteen days of doing what he already knew how to do, slower, with diminishing novelty and the same outputs he'd been producing for a week. He wrote it on the bark-sheet in the supply log's margin, because writing it made it real rather than something he could push to the edge of awareness and not fully examine. Day 24 XP: 190. Level 6: 14–18 days at current rate. He looked at it. Looked at it some more.

Then he wrote below it: Find something harder to do. 【Daily Summary — Day 24】 XP generated: 190 Soft cap efficiency: −38% vs. Day 10 baseline Plateau status: Active and confirmed Note: Plateau is not failure. It is the cost of the next level. The Watchers observe you accepting this. They find the acceptance less dramatic than they'd hoped. −38% efficiency. Active plateau. Confirmed. You have responded by updating a bark-sheet. Other players have responded to confirmed plateau status with crying. Actual crying.

The Watchers preferred that. The evening patrol had happened before the math, which was the correct order — activity before assessment, work before accounting, the principle that had kept the days functional through the plateau's increasingly flat returns. He'd run the full perimeter: three kilometers of established route, the biome-edge section where the iron-thorn thinned toward rust-bark, the territorial boundary check on the northwest wolf range.

Three-point-one kilometers total with the detour around the boulder cluster where the snare circuit's northern arm required a sight-line adjustment he'd been meaning to make for two days. The terrain worked him the way established routes did — not the novelty-training of new terrain, but the refinement training of known terrain run faster, the body finding efficiency in the familiar the way expertise found efficiency in repetition.

He pushed the pace on the second kilometer, let his feet make the routing decisions, and trusted the accumulated knowledge of twenty-four days in this territory to handle what his conscious attention didn't need to manage. He had become fluent in a three-kilometer loop of murder forest. This was not the professional development he would have listed on a CV but it was the professional development he had. 【Training Insight】 Evening patrol: 3. 1 km sustained movement, terrain variation. Agility: 11. 8 → 12. 0 (+0. 2) Willpower: 11. 8 → 12. 0 (+0. 2) Daily XP contribution: +47 You ran the patrol. The Watchers note the XP. Forty-seven. They would like to put that number next to the gap between your current level and the next one. They have done this math. You are attempting to outrun a plateau with the caloric equivalent of a brisk walk.

At this rate — and the Watchers want to be precise, because precision is something they have learned to appreciate from you — you will clear the plateau in approximately the same timeframe as a geological event. A slow one. You ran anyway. The Watchers note this correctly and with the exhausted respect of people watching someone bail out a lake with a soup spoon. The soup spoon is getting better. The lake is not getting smaller.

The Willpower gain alongside the Agility gain was the Nexus registering something he hadn't explicitly asked it to register — the fact that the evening patrol was doing two things simultaneously, and only one of them was physical training. The other was management.

Managing the quality of the evening, which without the patrol and without a task to put his hands to was the kind of empty that had been showing up in his awareness more frequently as the days had stabilized and the crisis density had dropped below the threshold that kept the background frequency occupied. He cooked dinner — two rabbits, the Herb-Wrapped Roast, automatic now — and ate in the shelter entrance with the fire at his back and the dark Crimson Wilds doing its nighttime register around the perimeter.

"I made 190 XP today by doing everything I'm supposed to do," he said, to the cooling embers of the cooking fire and the empty clearing and the wolf skull on the east wall.

"The Level 6 bar barely moved. This is the plateau and the plateau is mathematical and I respect mathematics and I also find this deeply unrewarding." The fire didn't offer an opinion on this. It never did. He was aware that he had started noting the fire's conversational limitations, which was either a sign of self-awareness or a sign that twenty-four days of solo operation had recalibrated his social expectations in ways worth monitoring. He opened the Ring Channel.

The channel had changed character in the three weeks since he'd first accessed it. The early panic wave had resolved into something quieter and more settled — a smaller, more stable population whose posting patterns he'd been categorizing the way he categorized everything: by content quality, by the presence or absence of requests attached to information, by what the language revealed about the situation the poster was actually in.

The group players were identifiable by their language — consistent we rather than I, references to shared shelter and coordinated hunting, the specific shorthand of people who had been talking to each other in person and carried that context into their channel posts. He'd catalogued four probable groups in the Ring Channel based on this linguistic analysis, sizes estimated at three to six members each. Unauthorized use: this story is on Amazon without permission from the author. Report any sightings.

The solo survivors were fewer and distinct: sparse posting patterns, practical content, no coordinate sharing, no asks. The posting profile of people who had learned the same lessons he'd learned and were applying similar operating principles. He'd been watching two of them with something between professional respect and the specific recognition of pattern-matching. He had been alone long enough that recognizing another person's operational discipline felt like a form of company. He noted this without examining it.

And then there was [LiraT]. He'd first noted the account in week two — a post about thorn-serpent nest distribution in the southern Crimson Wilds that had been accurate in the way that field observation was accurate: not guessed, not theorized, but derived from actual watching. He'd used the information and it had been correct. He'd filed the account under reliable source, no apparent angle alongside [KaiS] and had continued monitoring without further categorization. Tonight [LiraT] had posted again.

[LiraT]: new predator variant spotted near the Mistveil edge. moves in pairs. [LiraT]: active 0600–0900 and 1700–2000. avoids direct light exposure. [LiraT]: recommend 400m clearance from the ridgeline during those windows. He read it twice. The post had three characteristics that his threat model had been trained to look for in the inverse: it contained no request, no ask structure, no language that steered toward a meeting or a coordinate share.

It contained information that was useful regardless of whether he responded to it. And it was accurate — or at least consistent with the methodology of someone who produced accurate information, which was the best available proxy for accuracy before field verification. He adjusted his northern exploration planning accordingly. The Mistveil biome edge was north of the rust-bark transition zone, which meant any route toward Haven's Reach would pass within range of the predator window [LiraT] had documented.

He marked the activity windows on the bark-sheet alongside the northern route annotation and shifted his planned departure time two hours to avoid the 0600–0900 window. Useful information from a source with a good track record. That was the classification.

"[LiraT] posts information that is accurate and useful and asks for nothing in return," he said, to the bark-sheet and the fire and the wolf skull.

"In a chat full of people either panicking or hunting, this is extremely conspicuous." He'd read the post twice. He was aware he'd read it twice. He was aware of what it meant that he'd read it twice — not in any way that he was prepared to name precisely, but as a data point in the same category as the quiet of the evenings and the silence in the local channel's zero-message count. The category of things that were true and that he was noting without fully examining. The fire burned down to coals.

His Fatigue meter was approaching the threshold where sleep would be productive, and his meters were comfortable, and the day was complete in the sense that everything operational had been completed and the only thing left was the quality of the stillness. The stillness was fine. He looked at the bark-sheet. Find something harder to do.

He'd written it and meant it as a strategic directive — the plateau required escalating inputs, required pushing into unfamiliar territory, required the rust-bark recon extension and the boar problem and the Haven's Reach planning and all the items on the priority queue. It was the correct response to the 190 XP notification. It was also, he realized sitting with it in the late-evening quiet, the response to more than the XP notification. He didn't follow that thought further than its natural limit.

He banked the fire, settled into the shelter, and pulled the bark-sheet down from the wall to update the northern route annotation with [LiraT]'s predator windows. The practical adjustment took two minutes. He put the sheet back. He'd have more throw practice tomorrow. The count was at 44/500 and the number needed to move. He had the tannin investigation deferred again, which was becoming a habit he needed to break before it became a permanent feature of the priority queue.

The Grade B flint seam needed a harvest session. The chest armor needed the tannin before it could start. All of this was true and all of it was addressable and none of it explained why the Ring Channel was still open in his HUD at the edge of his vision when it didn't need to be. [LiraT]'s post sat at the top of recent activity — the practical timetable of when not to be on the Mistveil ridge, the activity windows she'd documented from what was clearly sustained direct observation rather than inference.

He had not posted a response. He was not going to post a response. His operating rules were written in dark charcoal and underlined twice on the bark-sheet and they were correct rules that he had arrived at through careful analysis and multiple data points and he was not revising them. He was going to close the channel. He didn't close the channel. 【Willpower — Passive Note】 Willpower 12. 0: Loneliness resistance threshold improved. Loneliness resistance threshold improved does not mean loneliness is absent.

The Watchers note the distinction. They find it more interesting than any combat result today. He ran the northern route adjustment one more time in his head. The predator windows meant the 0600 departure he'd been planning needed to shift to 0930 at the earliest. He'd lose two hours of daylight on the approach but gain the security of not walking into paired predators during their most active window on unfamiliar terrain. The tradeoff was correct.

[LiraT]'s information had made a real operational difference to a decision he was actively making. That was what good intelligence was for. He'd built an entire threat model on the principle that information quality mattered and source reliability mattered and unverified data shouldn't drive operational decisions. [LiraT]'s data was verified, once by the thorn-serpent confirmation and once by the methodology quality of tonight's post, and it had changed what he was going to do on a day that hadn't happened yet.

He would not post a response. The rules were the rules and the rules were correct. He closed the channel at 22: 14 by the Nexus clock, which was a timestamp he registered without meaning to. The camp was quiet. The fire was banked. The thorn perimeter held its rings. The wolf skull turned in the minimal air movement, patient as everything patient. He was almost certain the channel being open this long had been about the information quality. He was almost certain that was all it was.

He went to sleep leaving the question where it was — in the same holding file as all the other arithmetic that would resolve when it had enough data — and the camp was quiet, and the night ran its register, and 44/500 practice throws would become 44/500 again in the morning and then something larger.

The plateau was fourteen to eighteen days of the same work with slower returns, and at the end of it was Level 6, and at the end of Level 6 was whatever came after, and somewhere north of his territory [LiraT] had been watching predators move on a ridgeline and had written down the times. He filed this. He was very good at filing things.