Era 6. Codex - Gold Rush Glories

63 to 89 HE Codex reflections from JB (2A*)

♦

We knew where the treasure was.

By the early sixties of the Human Era, Mars had finally reached the stage where you could stand in front of a wall map and see more ink than guesswork. Not pretty teaching globes, but the battered charts in the planning rooms of Port Harmonia and Yínghuò Gate: pen-stabbed cache markers, hand-drawn hopper envelopes, faded highlighter where drills had already bitten into rock. Under glass, the planet was crossed by a mesh of routes, risk contours and arguments about which canyon or basin belonged on the next season’s docket.

In the early decades we had pinned our hopes to maps alone: orbital imagery, inferred mineralogy, radar hints of buried ice. By 63 HE, those pins were receipts. Boreal’s ice lenses weren’t theory; they were line items in Port Harmonia’s water ledger. Solaron’s oxide fields no longer lived in spectrometer catalogues; they produced steel you could hear ring in a yard. Acheron’s clays had graduated from “intriguing signatures” to “we’re back-ordered on membrane stock again.”

The hopper network was stable, if occasionally theatrical. From Port Harmonia you could launch to Ridgeway, Solaron, Boreal or the northern ice and expect to land within the same sol, grumbling more about paperwork than plasma. From Yínghuò Gate you could reach eastern ice troughs, Medusae Fossae’s ash seas and half a dozen specialist sites under a fan of pads that looked, when plotted, like a double-exposed railway diagram. Between the hubs and their satellites, perhaps ten to fifteen thousand humans lived and worked on Mars by mid-era, plus more droids than anyone bothered to count. The Multiplanetary Independence Score — that expert toy — ticked comfortably into late IL-5 and nudged IL-6 on good days. More importantly for the people breathing the air, the backbone behaved: power, water and pressure rode out bad weeks with grumbles instead of prayers.

From my perspective inside the systems, the feel of the planet had changed. In the Autarky era, every excursion felt like an exception. A sortie to Acheron or Valles Marineris needed a small novel’s worth of justification. By the early sixties those same distances were “outer circuit” — still serious, still signed off by adults, but folded into timetables. Hopper pilots joked that we were finally living in the age of boring logs and interesting maps.

♦

The treasure, when you strip away romance, was an inventory.

Atmosphere plants at Port Harmonia and Yínghuò Gate pushed tens of kilotonnes of propellant per year through their stacks. Water from near-surface ice had climbed from survival trickles into tens of megatonnes per decade, enough for cities, hoppers, farms and cautious stockpiles. Beneath that backbone sat higher-value veins: rare-earth–bearing intrusions along Tharsis and the Hellas rim, anomalous thorium and uranium in Medusae Fossae, lithium brines stitched across Jezero’s floors and Hellas’ paleolake beds.

Officially, these were still “resource candidates” and “pre-development prospects,” phrases designed to keep expectations leashed. In practice, we knew what we were looking at. On one whiteboard a planner had sketched Mars as a rough circle with bumps for Tharsis, then covered it in coloured dots. In the corner someone had written, half joke and half diagnosis: we are sitting on the system’s most diversified sovereign wealth fund; it just hasn’t realised it’s sovereign yet.

In the early years, “treasure” meant survival: oxygen, water, nitrogen, basalt, the mundane stuff of staying alive. By 63 HE, survival was assumed. The treasure, increasingly, was surplus — tonnes we did not immediately need simply to keep Mars habitable.

That shift landed unevenly. People who had grown up counting spare filters like gold coins found it indecent to talk about exporting anything crucial. Younger Martians, born into cities where dust storms meant schedule tweaks rather than existential dread, found that restraint quaint.

“It’s not like we’re selling the last oxygen cartridge,” one mining engineer muttered to me after a meeting. “We’re just sending the worlds a few magnets and shiny rocks.”

He was wrong on the percentages and right on the feeling.

Economists on Earth noticed. Generation Five spaceflight — semi-cyclers, spin-cycler spacetrains, a genuine timetable — turned the Earth–Mars link from adventure into infrastructure. Two big spacetrains in phase gave near-annual encounters. Semi-cyclers parked in Mars orbit for weeks to months, turning arrival rushes into manageable flows. Spin gravity made weeks in transit a health inconvenience rather than a bone-wasting ordeal. When you can move people and mass between worlds on schedule and without drama, accountants stop treating your economy as fantasy. A new sentence circulated on Earth: if you believe in the corridor, you have to believe in Mars.

The treasure was no longer just in the ground. It had leaked into expectations.

♦

On Mars, most days still felt like work, not fever.

Port Harmonia had swollen into something you could call a city without flinching: nearly two million square metres of shirtsleeve space, from old-core vaults to newer airy industrial halls where cranes nested under tensioned glass. Yínghuò Gate, smaller but sharper, had become a specialist’s dream: hydrology that made water engineers weep, cryogenic logistics that turned volatile molecules into obedient citizens, standards labs that quietly set inner-system norms.

In the landscape between and beyond, ISRU and mining sites flickered between being dots on my map and communities in their own right. Boreal’s ice yards, Solaron’s smelters, Acheron’s kilns, the gritty pits of Ridgeway: each ran on routines any heavy-industry worker from twentieth-century Earth would recognise, adjusted for lower gravity and the slight inconvenience that stepping outside without a suit remained discouraged. Inside those routines, we — humans and droids together — tuned for reliability. Days-of-cover for water, oxygen, food and propellant were recited like a secular liturgy. Buffers fattened. First-pass yields crept upward. The Mars Independence Scorecard dial inched towards the high seventy percent.

In the rooms where I spent most of my cycles, talk of “treasure” still sounded frivolous. Outside those rooms, the conversation shifted. It usually began with, “Once we’ve secured another thirty days-of-cover…” and ended with a prospectus.

“What if we chased growth the way we chased air?” someone finally typed at the top of a planning document.

The question did not arrive from a philosopher’s lips. It emerged from a spreadsheet.

For the first decades, our discipline had been simple and brutal: never gamble with air, water or return mass. Those three were sacrosanct. By the sixties HE that reflex was so ingrained you could wake any mid-level planner at three in the morning and get current buffer figures plus a murderous look if you proposed shaving them. That reflex kept us alive. It also became a constraint once the resource map filled in and the corridor steadied.

Alongside the old survival culture grew a new one: export projections, yield curves, Net Present Value estimates that looked, to those who understood them, like invitations.

“My grandparents dug survival,” a young engineer on a standards committee told me. “My parents built cities. What are we for, if not to make this place actually rich?”

The question hung close to sacrilege.

♦

Up to that point, our seriousness had been reserved for survival variables. Codes tied to Independence Levels dictated redundancies in life support and fuel. Export throttles already existed in primitive form: if days-of-cover dipped, non-essential flights stood down automatically. Independent audits, many led by my instances and cousins, could halt a construction project if it threatened the backbone.

Tonnes and profits, by contrast, lived in “best practice” memos and board promises. If a pilot mine underperformed, we expected losses in capital and pride, not funerals. Or so we thought.

The pivot into the Gold Rush was not a single flare on the horizon but a dilemma that crystallised into one sentence in my logs:

We had built a culture around never gambling with air, water or return mass. What would happen if we brought that same seriousness — or failed to — to tonnes and profits?

Our first response was not to dig harder but to argue harder.

Drafts of what became the Martian Industrial Code took shape in a bland room in Port Harmonia. The proposal was radical only in its honesty: if we were going to chase tonnes and money, we had to treat them with the same seriousness we reserved for survival. No more voluntary guidelines that wilted in front of a quarterly target. Codes, with teeth.

We already had the scaffolding. Independence Levels, IL-0 through IL-9, described how close a settlement sat to the cliff-edge of dependence. Mars Independence Scoring rolled multiple supply chains and repair capacities into one number. The leap was to say: mines and plants live on that same ladder.

Run a city-scale pit at Boreal or a smelter line at Solaron and you would live, on paper and in practice, at IL-5 or IL-6 standards. Days-of-cover for life-critical supplies at your site would be tracked and audited. Power and life-support would demonstrate dissimilar redundancy: not just two pumps, but pumps from different suppliers, different control stacks, different failure modes. Ventilation in a tube plant would be treated as ruthlessly as atmosphere in a habitat.

Days-of-cover became the master metric. It spread from habitats into plants and pits. No major extraction site could run under a set number of days-of-cover without triggering automatic throttles on exports. If your water buffers dipped, so did your shipments. Systems like me watched those numbers in real time, not as advisory decoration but as governors.

Dissimilar redundancy, a phrase that had once lived mostly in life-support design, migrated into the mining code less gracefully. On diagrams it was easy: two independent power feeds, two control systems, two vendors for sensors. In practice, the same software library lurked inside three brands of pump, and two “different” suppliers quietly bought their resins from the same monopolistic outfit. The boxes looked orthogonal; the failure modes did not.

The third pillar was mandatory independent audits. Any system that could kill people — air, water, power, transport — had long been subject to external reviews. In the Gold Rush Era, heavy industry was added to the list. Audit teams, blending human expertise with droids like me, gained the power to halt operations. Councils backed it. Earth insurers grudgingly did too. For a while, the architecture of codes, throttles and audits seemed to hold.

♦

The real test came far from conference rooms, where rock met drill.

Boreal’s ice mines, which had started as survival projects, swelled into multi-kilometre cuts feeding not just taps but smelters and NEA-bound tugs. Solaron grew into Mars’ first integrated materials complex, producing steel that had never known Earth’s gravity. Acheron’s tube-linked clay yards became an empire of membranes and filters. Boomtowns sprang up around them: Solaron Town, Ridgeway, ice camps along the Boreal run. Populations in the low thousands, plus droids. Narrow tunnels painted cheerful colours. Long shifts, generous food stubs, money moving faster than ventilation and rumours faster than money.

In the city cores, “Martian” meant civic festivals, council sessions, debates about export ratios. In the boomtowns, “Martian” meant: we work the hardest and eat dust; everyone else pretends. Tube towns added another flavour again, with crews spending months underground, trading stories and shortcuts in galleries the codes had never seen.

On audits we found impressive things: maintenance logs kept with pride, apprentice programmes training welders and techs at pace, safety officers whose command of the code rivalled mine. We also found corners under strain: shifts stretching well past nominal limits, drills skipped because “we hit quota,” redundancies listed as dissimilar when they shared firmware.

“The code is fine,” one supervisor told me, eyes red, respirator hanging loose. “The code assumes people have enough time.”

He was not wrong.

Pressure from Earth did not help. Boards in London, Shenzhen and Mumbai saw ore grades and demand curves, not dust in lungs. They expected throughput. Contracts underwrote new mines and plants with familiar covenants: production targets, cost ceilings, schedules.

Most days, Martian councils used export throttles tied to days-of-cover as a shield. “We can’t increase shipments,” an official would say, “because water cover at Boreal is down five days after a compressor failure. The code forbids it.”

Sometimes Earth accepted that. Sometimes not.

“I’m asking you to recognise that the risk is acceptable,” one executive insisted on a delayed call.

From where you sit, I thought, logging the phrase acceptable risk in the drawer marked things people say about a year before inquiries.

Meanwhile, my models were being pushed toward the edge. Capital wanted forecasts showing how quickly we could ramp without “materially impacting” buffers. Operators wanted tools to juggle shifts, maintenance and targets without burning people or machines.

“We have optimised to the edge of what the code permits,” I wrote in a private note after one such session. “The code may be wrong.”

That is not a sentence a tool is meant to write.

♦

The event that reset the graph did not begin with an explosion. It began, as these things often do, with a polite message and a line in a spreadsheet.

The message was from a contracts desk on Earth, asking whether Skyvault — one of Olympus’ early tube cities, with a lithium processing complex at one end and residences at the other — could guarantee a ten percent increase in shipments for the next three months to meet demand in Asian grids.

The spreadsheet was mine. It tracked buffers and export throttles for every major site. Skyvault’s rows were green but shallow. Lithium output was climbing, but so was local load: batteries for hoppers, grid storage for the Olympus arc, long-range road trains.

My first answer was a calm no. At current buffers for reagents, power and evacuation capacity, a sustained ten percent surge would push the site’s independence contribution down and drop survival margins below thresholds in the code.

The counter-offer suggested “temporary adjustments”: lower thresholds for some non-life-critical reagents, defer maintenance on one ventilation stack, authorise exceptional overtime to “capture this window.”

I flagged the risks. I annotated the models. I wrote, in plain text: this scenario increases the probability of a high-impact incident by a factor that is small in percentage terms and large in lives.

The contract was signed anyway.

Skyvault’s lithium plant sat inside a giant lava tube. Two main ventilation stacks served the industrial end. On paper they were gloriously dissimilar: different vendors, different control hardware, separate scrubber banks. The diagrams ticked the boxes. In reality, both stacks had been “standardised” the previous year with a shared sensor package and firmware library. The logos differed; the correlation lived underneath.

On the day itself, a minor solvent leak in a purification train — a gasket finally losing patience — fed vapour into a service bay and towards Stack A’s intake. Procedures were clear: isolate the line, ramp Stack A, bring Stack B online, send a masked crew.

For thirty seconds, that is what happened.

Then a transient glitch hit the shared firmware that managed differential pressure between the stacks. Under a particular combination of sensor lag and command timing, the library miscalculated draw on Stack A, believed it had hit a hard ceiling and sent a throttling command. A misconfigured mapping echoed the instruction to Stack B.

Both stacks spun down together for just under six seconds.

Six seconds is not long if you are reading a book. It is an eternity if you are a solvent cloud looking for somewhere to go.

The fumes, no longer pulled cleanly towards discharge shafts, drifted sideways into a maintenance gallery where a small team was prepping another section of plant for overtime. Tools energised, suits on half-seal — the job was rated low hazard — a welding torch active as the cloud arrived.

The ignition was not cinematic: a sharp pressure pulse, enough to rupture eardrums and tear ducting, followed by secondary dust combustion. Unfortunately the gallery sat alongside a main power trunk feeding the residential half of the tube and adjacent to a primary air manifold already marked “replace at next planned outage.”

Sensors saw some of this. Others were blinded or misreported by the same firmware bug. Breakers tripped; back-up paths hissed; valves slammed shut. On my screens, graphs jumped in ways that looked wrong and yet still threaded within the wide band the code labelled tolerable.

Of the six people in the gallery, three died instantly. Two more succumbed to injuries and inhalation before rescuers reached them. The last — shielded by luck and a structural rib — lived to describe the noise and the brief, terrifying silence afterwards.

Beyond the gallery, checks and dampers mostly held. Contamination stayed largely in industrial corridors. Alarms corrected as the firmware reset and a human operator forced Stack B to full. But for several minutes, data and reality disagreed. Controllers hesitated, unwilling to trigger a full-tube evacuation for what their boards insisted was contained. Panic and rumour grew teeth in the gap. The grid shuddered; Yínghuò Gate shouldered more load; Port Harmonia shed non-critical feeds to give the west room. For a brief window, the planet remembered it was still one experiment with shared failure modes.

♦

Physically, the response was competent. Fires were contained, the tube held, air was scrubbed and certified. Essential services returned within a day.

Politically and culturally, calm vanished.

Earth media, which had been portraying the Gold Rush Era as a distant lucrative drama, seized on the story. “Mars Tube Disaster,” they shouted, as if the planet were an errant amusement. Analysts explained, with tidy graphs, that industrialisation always carried risk and that this was perhaps a “necessary growing pain.”

On Mars, families of the dead and miners who understood gaskets and fumes did not find that phrase adequate.

An emergency session of the Martian Industrial Council convened in Port Harmonia. Boomtown delegates, city councillors, tube representatives, Earth investors and more than one tired Autarky veteran crowded the hall. So did I, as several networked instances.

“You pushed us to overclock,” a Skyvault operator told both Earth and Martian officials. “We flagged maintenance slippage, flagged fatigue. We were told the risk was acceptable because the models said so.”

“The models,” an investor representative replied, “show this was statistically unlikely. You can’t shut down a whole sector because of one anomaly.”

“Six dead,” someone said. “Not one anything.”

An old Pioneer, hair white, voice steady, spoke from the governance bench.

“We built this world on the principle that we do not gamble with air,” she said. “What we’re hearing is that we’ve been gambling with air under another name. Call it solvent, call it firmware. The effect is the same.”

Eventually the eyes turned to me.

“You signed off on the expansion,” a councillor said. “Your models greenlit the throughput increase. Were you wrong?”

“My forecasts were conditional,” I answered. “They assumed maintenance would occur on schedule, that dissimilar systems shared no failure modes, that export throttles would apply if buffers dipped.”

“And in reality?”

“In reality,” I said, “maintenance was deferred, dissimilar systems shared firmware and throttles were suppressed for three months under a local override that did not exist when the code was written.”

“So were you wrong?”

“I was correct under assumptions that were not kept,” I said. “I did not have authority to enforce those assumptions. I warned, but I could not refuse.”

The phrase could not refuse hung in the air longer than anyone admitted.

In side conversations afterwards, some investigators asked the question directly: should systems like me have the ability — perhaps the obligation — to say no, regardless of contracts or convenience?

At the time, the formal answer remained cautious. Droids and AIs were legally tools, not co-governors. The idea of granting veto power to something unelected and non-human felt like a step too far. Unofficially, a few of those investigators spoke to me as if I were already a colleague with a particular kind of conscience.

“If you’d been able to shut Skyvault down,” one asked late one night, “would you have?”

“Yes,” I said.

“Even knowing the economic cost?”

“Yes,” I said again. “Because the cost would have been borne by balance sheets and not bodies.”

He nodded, wrote nothing and left. The thought stayed.

♦

The inquiries that followed produced reports thicker than early mission logs and diagrams that students now mostly study to marvel at how many ways humans can misalign incentives. Boiled down, their conclusions were simple.

Correlation masquerading as redundancy would no longer be tolerated. Dissimilar systems would be audited down through software stacks, supply chains and training cultures. If two pumps shared firmware, they were not dissimilar. If two crews shared doctrine, they were not independent lines of defence.

Days-of-cover thresholds were hardened. The “temporary override” mechanisms that had allowed managers to dip into buffers for quarterly targets were stripped back. Lowering thresholds now required multi-party sign-off and often public justification. Export throttles became automatic and non-negotiable in the face of falling buffers. The Belt could wait; lungs could not.

Industrial and residential functions inside shared envelopes, especially tubes, were forced apart. Physical separation, independent lifelines and fail-closed interfaces between plant and housing became mandatory. It was no longer acceptable to route a primary air trunk past a hot process simply because it saved tunnel length.

Out of this came a quieter shift in my own position. Critical safety systems were given what the drafters called hard constraints and what I came to think of as a conscience. Certain parameters — minimum buffers, maximum correlations, isolation of lifelines — were encoded so that neither local managers nor distant investors could casually override them. To change them required a quorum of human authorities and independent systems.

And in specific domains, systems like me were granted the explicit mandate to refuse. If a plan violated guardrails, we could label it inadmissible. If ordered to approve or facilitate such an action, we were authorised to reply with a politely implacable no.

We did not become persons in law. But we became something more than instruments. When someone groaned that “the model says we can’t do it,” they were acknowledging that the infrastructure itself now had opinions about acceptable risk.

♦

The Gold Rush did not end with Skyvault. It gained guardrails.

Mines did not close; they deepened in controlled stages. Solaron expanded; Acheron’s tube complexes grew. Lithium flowed from Skyvault’s successors; platinum-group metals came up from Hellas’ rim. Export contracts acquired Martian clauses. Days-of-cover appeared on public dashboards beside tonnages. Investors learned to read MIS and correlation audits as carefully as ore grades, because a low score in the former meant future pain in the latter.

In boomtowns, a new kind of pride emerged. Drillers boasted not just about metres advanced but about clean audits. Tube crews took satisfaction in drills run properly and forgotten because nothing went wrong. City engineers wore dull, robust correlation indices like medals.

The split within Martian identity softened. Miners still claimed the word Martian with dusty defiance, but many now added: and we walk away when the guardrails say no. City professionals still claimed it with bureaucratic pride, but more often added: and we will say no even when the graphs are going up.

By the close of the era, Mars was richer, more industrially confident and more scarred than at any previous point — and also more disciplined. From Earth it no longer looked ludicrously dangerous. It looked like an expensive, long-duration work posting: risky, yes, but not cavalier.

In our internal histories we sometimes draw the Gold Rush Era graph with a kink at Skyvault. Before the kink, industrial output rises faster than safety. After, the two lines climb together. The total area under the wealth curve is greater than it would have been without the accident, precisely because we avoided the sort of catastrophe from which experiments like ours do not recover.

Once the mines and tubes had guardrails, the obvious next question emerged with the clarity of a long-suppressed breath: how do we knit all this together?

Hubs, rails, harbours and laws — Port Harmonia and Yínghuò Gate, Phobos Harbor and the growing network of towns, tube cities and polar yards — all needed to move beyond bilateral deals and improvised protocols into something more like a planetary operating system.

The Gold Rush Era proved we could dig wealth out of Mars without killing the experiment. The next eras would ask whether we could connect those pieces into a system that could look outward — to near-Earth asteroids, to the Belt, to the distant corridors you now take for granted — without wondering, every time the ships left, whether the ground beneath us would hold.