The first thing Goblin Cave tackled, with the enthusiasm of spring cleaning, was the first five floors. Its mana being cut off during the floor 26 collapse had killed most of its mobs via mana depletion, and there were even a few souls that seemed to have... vanished, escaped its respawning loops while it has been cut off. Goblin Cave felt guilty about that. It was a profound kind of loss, something irreplaceable gone due to its lack of power and foresight, but also... there wasn't much exactly it could do to recover them, if they had escaped. The thing of primary import was that its upper floors were increasingly barren the further to the surface they got. And also, its mobs had been utterly useless for any kind of meaningful defense.

With a scything gesture, Goblin Cave despawned every normal mob on its top 14 floors, save for each floor boss. The tight whirlpool of its respawn system abruptly ceased, souls circling about with nowhere to go, and the constant dull pressure it had grown so used to over its life ceased in an instant, leaving behind a strange stillness.

And then, looking deeper, it despawned the lower-leveled half of all the mobs on all other floors, too. The souls, it caught up and redirected, ushering the majority of them to be received by its goblin-god, the control node tasked with managing its mana goblin spawns. That was enough to surge the mana goblin population, sharply advancing it towards the theorized exponential inflection point of soul generation. The mana goblins were both its most efficient way to produce more soul, and the most interesting thing within its dungeon.

To the low-level adventurers entering it, it asked for live samples of any kind of external flora or fauna. Anything that could fit into one of its mana-sealing boxes and survive. It wanted to see the whole breadth of what unempowered, mana-dull physical world had to offer, and ideally, how to cultivate it within its upper floors.

It still needed mobs for a token defense from medium-leveled delvers, but it seemed like everything was either something that could be killed with some kind of mana machinery, or something that it would be unable to scratch by any means: the mobs were not meaningfully contributing to dungeon defense.

The cavegrass the sage bled on, Goblin Cave selected to evolve into a [Rustroot Herb], simply for the lateral evolution. It was still interested in exploring the possibility space of its spawn tables, and although it was more interested in determining the why of their construction rather than the brute engineering of unlocking every mob in every category, it would not turn down free mob evolutions.

The two sages obligingly bled on several other cavegrasses, and then on the chokevine on floor three, and in the end Goblin Cave had slowly-metamorphing rustroot herb, bladeweed, sparkgrass, a [Sopor Bush] (t2, time) that one cavegrass had also unlocked, and the chokevine was metamorphosing into a [Deathrattle Choker] (t2, death).

The sages stepped through its experience-testing hall, bestowing upon each one a drop of blood, save the [Goblin], [Brown Mushroom], and [Lesser Mana Puppet], which it had already fully unlocked all evolutions: the [Wolf] branched off into a [Windfang], [Shadow Darters] into [Blood Carp], [Flame Wisp] into [Will-o-wisp], [Biting Skull] into [Crumbling Skeleton], [Bat] into [Bloody Bat], and finally, [Mana Goblin] into [Infused Amalgam]. The hallway became a series of pulsing cocoons, attached wetly to the smooth quartz walls.

As its mana slowly ticked back up, Goblin Cave spent most of the rest of it on re-excavating floor 26. Absorbing the broken wreckage of its mana resonator gave it more mana than it took to consume it, and so it didn't take long before it had fully cleared the torus again, leaving it bare rock for now. The cost of building it... it had done it over several months, the first time, and so the mana cost had never seemed too exorbitant, but now contemplating rebuilding it, it did seem like a fairly immense cost. The problem was, this was effectively its sole meaningful defense. It pieced together a smaller version, cycling feather-light pulses of mana around a tiny circuit, and tweaked some of the materials and timing mechanisms. It was little more than a toy; its alterations would likely not scale up. But it was something to occupy it while it focused on cleaning up the mess.

That was the what of what it was doing. The why, it dwelled on. Bronze was all well and good, but what the adventurers wanted, ultimately, was experience. Death. And it wanted... well, what it wanted was a unified force to explore the nature of the world, with adventurers recovering resources or providing their unique point of view. What it was getting, it suspected, was a lot of low-level adventurers that didn't have any appreciable skill at anything, and who weren't particularly inclined to help it. It would probably have to do something with its experience grinder idea. Which was unfortunate: it had been intended as a testament to the system's endless gluttony, not as an engineering project that would actually see use! But it knew of only one way to transfer experience to an adventurer, and that was struggle against an active, attacking mob. It would be much more convenient if they could drink mob blood.

DOES THAT LOWER YOUR EXPERIENCE?, it had asked of the sages, as they bled onto its mobs.

"Not just drops," they said. "But lose enough blood, and losses begin to accumulate."

IS THAT CUMULATIVE, OR AMOUNT OF BLOOD OVER TIME? It didn't have much experience with adventurers carrying on after bleeding extensively.

"Cumulative. It's a primitive experience siphon." The sage explained the concept: dungeons were like an acid, slowly etching away at the substance of an adventurer's soul. The blood was the medium by which it could forge a connection, and sluggishly draw out experience with the blood. It was theorized that there was a hidden 'secondary reservoir' of experience, filled in tandem with the more visible system value, and when that was exhausted 'true' experience loss began.

Which was interesting, because it meant that one adventurer could transfer experience to another: simply bleed enough on a mob, and then have the other kill it after it had leveled sufficiently. The nature of experience ratios and the secondary reservoir seemed to imply that that was a profoundly inefficient way to level, though. The system wanted true blood.

What it wanted was... to understand the world. The system, certainly. But also the whole breadth of it, all the unknown corners of the outer world, the nightmarish cosmos above, the things the sages had divined, the ways in which adventurers had banded together and formed communities. And it wanted to replicate them: expand its tiny internal world until it was a match to the whole of the outer world. It was a big goal.

One step was sealing off most of its upper floors with mana locks: this drastically reduced the mana density there. It could still sense, vaguely, through the space, and this let it uncap the chambers it had of worms and ants, giving them free rein of the old goblin cave. This made it mostly impractical to spawn anything, with anything beyond tier-0 plants or fungi dissolving apart into loose mana.

What it wanted was to understand the 'Gaia harmonic', the 'natural' flow of mana outside of dungeons. This involved collecting a large variety of creatures with low or negligible mana emanations, and keeping them alive long enough for their weak mana flow to start resonating. If it could actually sense it happening in its upper floors, it was a step towards recreating it in a more mana-dense space, and potentially not dissolving any worms or ants it tried to look at directly.

The other thing was... the sages spoke of a world full of people. Encampments and villages and hamlets, with vast capital cities even nearby, in the duchy. Villages in treetops, in caves, ancient fortresses and burning deserts. Vast urban cities of lands unknown to the locals. Travel was a bizarre concept to it. Obviously, adventurers did it, moving from place to place, but the way the sages spoke revealed some of the lived experience of what it was like: picking a direction and walking, and then walking and walking, over days and days, until Goblin Cave's mana would have long ago thinned out and dissipated from the distance. It was sharply envious of their ability to simply move.

And the last: adventurers returned, bearing moss and lichen, bug and rodent, branch and cutting. Soil rich with strange organisms. Tiny squirming things that swam in dewdrops. Goblin Cave couldn't look at them too closely; its mana-sight sliced apart everything it saw. But now, in the low-mana zone of its upper floors, weak, mana-ignorant beasts slowly gained a hold. And for their payment...

It had had to restructure its mana flow. With the low density, no longer could it reach up from behind the rock and form a mana lock. It had to made a secondary flow network within the rock: mithril-coated tubes and pipes, to prevent its mana from leaking back out. It directed them to specific locations, where it formed more permanent surfaces: big walls of quartz, lined with mana locks. A handful of manastone-walled chambers, pockets of dense mana sealed away from the rest of the low-pressure zone.

I CAN PRODUCE ANYTHING YOU CAN THINK OF, it wrote. YOU MINE, OR DELVE, FOR THESE THINGS, THINKING THEY ARE RARE AND PRECIOUS. A DUNGEON, PERHAPS ANY DUNGEON, CAN CREATE THEM FROM WISPS OF MANA AS EASILY AS YOU WOULD TAKE A STEP. I DO NOT UNDERSTAND THE REASONING BEHIND ANY OTHER DUNGEON'S ACTIONS, OR OF THE SYSTEM THAT BESTOWED THIS ABILITY UPON ME, BUT I AM NO LONGER INTERESTED IN KEEPING IT HIDDEN. THE COST IS PALTRY; THE EFFORT EVEN LESS.

It spawned a series of cubes in its chamber: copper, tin, bronze, silver, gold. Pewter, lead. Platinum, aluminum. And then, seeing as each one was several tons, it sliced them apart into wobbling stacks: small centimeter-long cubes of every material. The gems, it cut into more elaborate shapes; the adventurers' habit of shaping gems into polyhedra was one of the few things it appreciated about them. Complex polyhedra, faceted lozenges of emerald, aquamarine, amethyst, alexandrite, tumbling down in heaps. Thick piles of rough-milled lapis or burnt copper, forming rich blue and teal dyes.

THESE THINGS ARE USELESS TO ME. USE THEM AS YOU WILL. OR COME WITH ME, AND ATTEMPT TO COMPREHEND A MEANING BEHIND THE SYSTEM'S ABILITIES.