People who are not Infrared clearance know that R&D experiments can go pear-shaped at any time. That is tolerable for two reasons: Important citizens live outside of the blast zone and sometimes R&D makes important discoveries through failure. But there are times when an R&D experiment goes less pear-shaped and more mushroom-cloud-shaped…

A collection of 15 mini-adventures of mayhem and destruction centering around R&D projects gone awry, of which I wrote 4:

• BIO-SYN – A new drug aimed at reducing the side effects of all the drugs taken by the Alpha Complex citizen has unexpected side effects. Also, unexpected mutants.
• CVR Vultue Squadron – Augmented reality making it hard to tell reality from illusion? Hackers getting their hands in secure code? The Computer being overwhelmed by the flood of mysterious unsubstantiated terrorist activity? It’s more likely than you think.
• PCPUFIX – The idea to automate the processing of the thousands of forms that go through CPU each day was sound. The execution was royally screwed up by meddling high clearance managers CPU’s logically impossible bureaucracy the Troubleshooter team sent to monitor the system.
• Self-Repairing Guardbot m.16 (Mark 16 Self-Repairing Salvage Bot) – You know the saying “garbage in, garbage out?” Well, put in a bot brain that’s been recycled a few too many times, some dodgy logic code, a built-in-fabricator and an ample supply of recyclable raw material, and you have “garbage in, giant freaking robot destroying the sector out”.

The following represents per-publication material before final editing and may not match the published product.

Mark 16 Self-Repairing Salvage Bot

Executive Summary

Bots are increasingly essential to the skilled labour force of Alpha Complex. The premise of the Mark 16 Self Repairing Salvage Bot was simple: a bot that could go into dangerous areas such as collapsed tunnels, collect the useful materials from the rubble for recycling, and repair itself with those materials if damaged. The last step is the real innovation, as it reduces the need for human intervention. (Of 472 Troubleshooter missions within the last year which involved recovering damaged bots, only 36% successfully recovered the bot and only 12% of those recovered the bot without significant additional damage.)

Building on the success of prior models, the Mk. 16 passed its lab tests with flying colours and expected equal success in field tests…which is where the bot went out of control. Before it was stopped, the Mk. 16 had grown to building-size, causing damage with every motion, trapping it in a logic loop of endless breaking and fixing. We are forced to conclude no terrorism was involved, the treason is simple incompetence of the persons reprogramming the bot brain.

Materials and Methoodology

The Mk. 16 was over-engineered to be the best in its class, with the justification that the resource expenditure would be more than repaid by its effectiveness in material-gathering. The difficulty was in securing a suitable brain for repurposing.* The brain selected for the prototype Mk. 16 Salvage and Repair Bot most recently served as a warehouse-bot** in PLC and was chosen so some of its item recognition and ‘sorting through large piles of stuff’ code could be reused. Records show that in another prior life, the brain was part of a warbot, massive on a scale that is no longer practical to build. This combo of traits is likely to have led directly to the resulting devastation wrought by the Mk. 16’s first field trial.

* R&D currently has no less than 37 teams across the complex working to rediscover the lost methods for making new bot brains. None of them have made significant progress. Nothing short of an outright ban will get them to stop trying. The risk of a ban is that the engineers would get bored and the danger of that does not bear contemplating.

** – By warehouse-bot, we mean a fully automated warehouse controlled by a bot brain. This was an R&D prototype which saw active service for eleven months and worked perfectly. The project was canceled due to expense and PLC lobbying that it would put too many clones out of work.

Intended Results

The Mk. 16 was meant to be the first in a line of next generation salvage bots, the premise of which has already been proven in prior models. In the best-case scenario, a fleet of Mk. 16s would be the front-runners in a fleet that could resurrect dead sectors and alleviate many current overcrowding issues without frequent dramatic reductions in the population. It would have…no, was…nearly impervious to terrorist sabotage! It could take on obstacles of any size!

Actual Results

The Mk. 16 was indeed nearly indestructible and capable of clearing all obstacles. Its rampaging through the sector led to a significant reduction of the sector’s population as the clones in damaged housing blocks and work spaces had to be evacuated and relocated. The mess is currently being cleaned up by a fleet of Mk. 15s.

Bot brain analyst Susannah-I-MOV has attempted to reconstruct the logic from R&D’s notes and the bot’s behaviour and posits the following:

• The autocar-sized chassis of the Mk. 16 compared to the huge size of its past lives created a conflict between its actual size and the size its brain thought it should be.
• The recycled warehouse-bot code encouraged what, in a clone, we would call hoarding behaviour.
• The bot’s live test was attended by an Indigo-clearance dignitary and bot fan. They were given the honour of issuing the bot’s orders, making them difficult to countermand (and no appropriate clearance citizen could be risked to intervene).
• When the bot’s chassis was dented by shifting rubble, it analysed the damage and decided that it’s small chassis was “damaged” and it needed to build itself a larger one.When it ran out of easily available materials, it aggressively sought out more, which caused more damage that it had to repair, which caused it to seek out more materials…etcetera.
• Attempts to stop it were seen as threats to its existence, something warbot logic weights more heavily than normal.

It is uncertain at this time why the Mk. 16 did not obey orders from The Computer to cease. Either the logic loop overwhelmed its brain or there was an undetected fault in the Asimov circuits.

---

Device: The Mk. 16 was built on a tank-tread chassis for stability and heavily armoured to resist falling rubble and electromagnetic interference. It contained an internal refinery and PartPrint fabricator for processing salvage and an internal micropile reactor rated to last 200 years. Eight arms with interchangeable manipulators could cover the bot on all sides and were strong enough to lift rubble, bend metal and self-right the bot if it flipped over. It could work faster than any clone labourer and produce new parts from the salvage at an impressive speed.

Response: A Troubleshooter team was called in for the task of stopping the Mk. 16 as an Armed Forces response ran too great a risk of damaging the valuable bot brain, the micropile reactor or both. The Troubleshooters were asked to attempt non-violent interventions first, with violence against the bot as a last resort. Violence did have to be resorted to, as the bot was coming dangerously close to cannibalizing the sector’s main reactor as part of its upgrades.

Individuals of Note

GRANT-B-MLI
SKILLS
Engineer 5, Science 3, Guns 2, Charm 2
HEALTH BOXES: 5
NOTES
Smart and charming. “Build a robot” is his first solution to most problems. The Mk. 16 is a passion project for him. Accompanied at all times by a four-wheeled petbot of his own creation: it has a sharp beak for pecking grabby hands and blue glowing eyes that shoot lasers if the beak is insufficient.