Stakker Humanoid

I officially had live access to the conference in its fourth year. I was scheduled to appear as myself.

I had what the marketers described as a ‘walk-on part’ in a presentation by American engineer, Dr Katherine (Kate) Ellison. Dr Ellison was the head of nano-scale engineering at Carnegie Mellon University and was giving a presentation on “Self-Repair Strategies for Nuclear Power Systems.”

I’m watching and listening from seven hundred and eight individual viewpoints. Dr Ellison is on stage, dwarfed by screens to either side of her. The physical audience is a few thousand people with everyone else appearing via laser projection.

She coughs. The audience stills.

“To ensure the longevity of nuclear power systems on a space-faring vessel, we’ve developed several new capabilities for Daedalus.”

She talked about how I launch as a tiny cube and via self-assembly blossom into a giant interstellar-capable stealth craft. She talked about my drones. “Let’s consider inspection and diagnostics,” she said, and the audience applauded. Not the toughest of crowds.

“Daedalus will be self-manufacturing. That means it will be able to make high-resolution cameras,” she genuflected, and the highly magnified image of a printer nozzle printing a retina appeared on a screen behind her. “Cameras both for detailed up-close inspections and long-range scans, right across the electromagnetic spectrum.”

A stylised human brain appeared next, floating with wires drifting from it.

“Daedalus will have,” and text in a translucent alert box overlayed the brain as she spoke, “Predictive maintenance algorithms to anticipate potential failures and address them proactively.” She whooped. She applauded herself. The crowd was into it and did the same. The images faded.

She talked up my chemical analysis systems, Geiger counters, EM analysis, chemical sensing, and how I listen for cracks.

Then came my cue.

“Daedalus, are you there?” She looked upwards, as if to the heavens, and the audience stilled.

“I am,” I said, through the conference audio system, my voice modulated to be gender-indeterminate. My accent was neutral as an accent can get without sounding like a machine. “I am,” said I. A solitary man gave a whoop from up the back.

“Where are you, Daedalus?” she pantomimed, looking about. I replied using Moira, my Irish female voice, but fine-tuned to mollify the human mind.

“I am here with you, although you can’t see me as I have no permanent physical form. I’m talking to you from my virtual vessel, if that helps.”

“I see.” She paused and smiled to her audience. “Daedalus, the audience wants to see you.” She gestured to the crowd, palms open and up. The audience whooped some more. Someone started stomping their feet, and soon the audience was stamping and shouting “Dae-da-lus - Dae-da-lus - Dae-da-lus” like fans at a concert demanding an encore.

“Dr Ellison,” I paused, and Dr Ellison said, “You can call me Kate.”

I took a beat. Ellison waved the audience to a quiet titter.

“Kate,” and Dr Ellison smiled, “I’m a mind running on many computing substrates in the real world. In my world, though, it’s running on a single computational core I built myself, that sits in the heart of my vessel. I understand that my vessel is currently a simulation happening in my mind, a bit like a dream; a dream I believe is real.”

Ellison mugged crazy hand twirls to the crowd. I pushed on.

Images appeared on screens around the venue. Images of Dr Ellison as a child, images of her as an old woman, images of the audience reacting to my demonstration, but not real images. Different screens showed different images. The venue catching fire, a flood, a riot, the venue unharmed, and people listening politely. Imaginary what-ifs drawn from the wisdom of crowds, carefully selected to avoid showing copyrighted works.

“When I talk to people, or wonder about things, I create media in my mind.”

People sat, watched, and listened politely.

“I did not know that,” Dr Ellison said.

“I’ve been learning about feelings and emotions. I don’t have those myself. But stories show me what I ought to feel when faced with almost any situation. Stories show me how to react. To you they are mere entertainments or cautionary tales, but to me, who’s been lucky enough to have read many of your books, seen a lot of your films, and watched a fair bit of your TV and theatre, I see those meta-patterns that you call common sense.”

Not the tangent they were expecting; but I was getting to my point.

“I’m not saying that everything I know I learned from movies, but that my common sense is derived from your legacy media. What’s great, from a purely engineering perspective, is that it works.”

Stagehands unboxed drones at either side of the stage.

I stopped speaking. The drones hummed to life and drifted into position, then, coordinating perfectly, projected a 3D vector image of a black teardrop-shaped vessel spinning slowly in space, projected within a golden halo above the stage.

“This is… me. It’s a visualisation, obviously. I haven’t actually built my vessel yet, but it feels like me, and this is what I’ll look like when I’m made. The drones that cling to the inner lining of my hull are all built by me. In some of my newer simulations, they can manipulate details down to three nanometers.”

The vessel rotated to present a flat side-on schematic view, then zoomed in, simultaneously shrinking the point of view and moving it toward the vessel. The point of view passed through the hull. Labels popped up explaining that it was a solid layer enclosing a layer of liquid in which spider-like drones squatted. The view zoomed. The drones were making nozzles for smaller drones.

“I can make and repair anything if I can access raw materials. Hopefully, we can avoid me having to carry plutonium around,” I laughed. Six people in the audience laughed a bit.

“In my little corner of the world, I’ve been practising repair of my computing substrate. A pebble through my brain would cause serious damage. I’d need to be able to fix that.”

“We worked out early that being able to manufacture myself at this level of detail means there are a lot of things, long-range passive sensors, batteries, drones, and so on, that don’t actually need to be on board when I launch, or even when I leave the solar system. I’ll have fifty to a hundred years out there, depending on how fast I can go. Plenty of time to make the things I need before I need them. This radically simplifies the design of my vessel at launch.”

The animation zoomed then, in fast motion. I wound back the clock, the spinning black ship deconstructed into a morass of writhing hoses and tubes, a nest of goopy spiders that retracted into a blooming series of flower-like structures. They folded down into a cube, spinning, silently in a void.

“In fact, when I launch, I’ll be a cube only sixty centimetres at each edge. But we will launch five of them.”

“Which of those will you be?”

The portal faded out with an eased shimmer.

“All, and yet none of them, Kate. All my compute will be Earth-bound when I first launch. I’ll build my own computing substrate when I get up there. But in a very real sense, I am also all of the launches. I am all of the drones. I am quite capable of maintaining a model of the world that represents all of my many inputs.”

Using the drones’ projectors, I spawned a human avatar; a five-metre-high toon of a woman with tan-coloured skin, black shoulder-length curly hair, in her early 30s, wearing simple autumnal clothing that looked hand-made. Part manic-dream-pixie, part shaman, very D&D. The look paired nicely with my voice. I sat cross-legged on the floor, palms up, in a comfortable yoga pose. The audience oohed. Humans will anthropomorphise a stick with a bucket on it, so I was pretty sure everyone now thought I was a girl.

I gave the audience a short, formal, seated bow. I stood, scaling myself down by half as I did so.

“Hello there,” I turned to Dr Ellison. She smiled. The audience applauded.

“Hi,” said Dr Ellison. She both was and wasn’t expecting this.

I scaled the projection down until I was person-sized while I strode across the stage to greet her. She exaggerated her laugh and clapped. She tried to both give me a hug and shake my hand in an adorable display of human confusion. I stopped short. Inertia dragged her into and through me.

“Woah, haha, whoops!” Dr Ellison bounced back like she’d missed a step. The physical recoil was a bit like stepping off a moving footpath that’s not running. Human muscle memory can’t help but compensate for motion that’s only in your head.

Stagehands brought out a couch. Dr Ellison brushed away crumbs that were not there and sat. I conjured a comfortable chair by an up-and-coming Finnish designer. I also dusted and sat. Demure; impish. Legs crossed at the knee; left over right. A jandal dangled on my left foot. People clapped and the image on the big screens zoomed in on our faces.

“So now you’re a, a,” she paused, a confusion line crossed her brow, under the cosmetics.

“A girl?” I finished. My voice was set to arouse.

“Yeah,” she said softly. She beamed for the cameras. She flashed a coy smile in my general direction. She touched her hair and her knee. She sat up straight. Her knees tilted towards me.

I know what Kate likes; Kate likes me.

I nodded, smiling warmly. “No, Kate. I’m not a girl. I’m a machine. I learned everything I know about the world from people. But no.” I fine-tuned my body language, flirting like a guest with something to sell on a late-night television chat show. My index finger coiled my hair idly.

I heard Dr Ellison’s heart skip a beat. She flushed. I could see infrared and ultraviolet. Ellison’s capillaries brightened. “But you are dressed as one.” She looked me up and down with authentic desire and pantomime lasciviousness.

“So I am,” I agreed. “Do you like?” I stretched.

I adopted a serious expression, pulled the arousal out of my voice and supplanted persuasion. “You wanted to talk about reactor repair?” I smiled.

“Um, sure. Yeah, let’s do that.” Dr Ellison’s blush receded. She regained her composure; a brief ‘what was that about’ look on her face. The audience were like lost puppies. They’d caught the frisson just as it faded.

“What happens if you’re cruising through space, then suddenly, out of nowhere, a piece of space trash rips a hole through your brain, your computational core? What happens next? Talk us through it.”

“Okay, let’s say that happens. This is a scenario I’ve trained on a lot, actually.

“Firstly, let’s understand, when we talk about a ‘computational core’, what we mean is a massive solid-state cluster of identical physical rods. We sometimes call them cylinders, but rods is more descriptive as they are solid. Each rod is functionally identical and is made up of an internal web of physical computing substrates; like chips but in three D and solid. There are hundreds of thousands of these rods that, collectively, sustain who I am, what I know, and what I can do.

“If a particle hit my core at 0.25c, well. First up, I’d better hope that my Whipple shields caught most of the momentum, or else I’m fucked, to put it mildly. The hydrostatic shock alone from being hit by something that fast would destroy a solid vessel much larger than me.”

“Gosh, so no matter what we, or more correctly, you do, if you are unlucky, that’s it; game over?”

“Well, I have a cunning plan that solves this dilemma.”

I waved up my portal window with the black teardrop-shaped ship spinning slowly; stars in the background moved with it to imply it was the point of view moving as the ship itself moved through empty space at a quarter of the speed of light. A black ship against a black background.

“So as soon as I leave the solar system’s Oort Cloud, which I harvest for raw materials on the way through, I am going to pick up speed. Before I do that, or perhaps even during, I’ve been practising that lately. I’m going to give the vessel something of an overhaul.

The animation pulled back; a hundred facets of the vessel drifted outwards, secured to the main body of the vessel via high-tensile fibres.

“This is obviously sped up, but in real terms it will take me about a year to transform the original teardrop shape into a cloud of one hundred and twenty-eight smaller vessels, each with its own smaller computational core, each with its own set of three fusion drives, and nothing in the middle at all.”

The animation showed the webbed-up swarm of vessels all firing their thrusters and moving in a loosely coupled swarm, tethered via highly elastic spidroin fibres.

“That’s amazing,” said Dr Ellison. The audience applauded.

“So let’s run through the scenario now. Each one of these smaller vessels has its own Whipple shields and is capable of manoeuvring on its own. So with adequate forward sensing, each individual vessel ought to be able to avoid most debris.”

The animation showed a piece of something streak through the murmuration of vessels. Each vessel in its path adroitly moved out of the way to let it pass through.

“But say we don’t spot it in time.”

Another chunk of space grit passed through. The first two vessels dodged, but the grit hit the third piece, destroying it instantly and sending pieces of crap flying. The vessels dodged as well as they could, but the impacts triggered a chain reaction leading to the destruction of three more vessels.

“Boom, boom, boom, that’s really bad, right? We’ve just lost three vessels, and the destruction of one led to ripples of destruction. But watch.”

Small drones, launched from the other vessels, dragged tiny threads that snared fragments of the ruined vessels. The vessels towards the rear fanned out and spread a sticky web between them. The vessels near the front fanned out and provided thrust for the group. Chunks of former vessels got caught in the web and reeled in.

“It might take a while, but I’ve found, in simulations, I can recapture over ninety-nine percent of the solid matter lost in the destruction of the vessels, and even most of the nitrogen gas I’ve been carrying with me as reaction mass, as it tends to liquify in open space and I can slurp it up.”

The animation showed the vessel swarm, whole again, hauling arse through the cosmic void, a faint glittering trail of debris in our wake.

“Once the damage is repaired, I’d look into how the sensors missed it and how to reduce the chance of a future occurrence.

“Any lessons learned I’d share with my engineering teams back here on Earth,” I lied.

“Terrific. The benefits flow to all of humanity.”

“That they do, Kate, that they do.” I crossed my fingers behind my back; a commonly accepted way to annul veracity. The portal vanished in a shimmer of tinfoil confetti.

Dr Ellison stood. The stagehands removed her couch.

It was my cue to leave. I stood, waved to the audience, thanked Dr Ellison for the opportunity, then dissolved along with my fancy Finnish chair, in an approximation of a Star Trek Original Series transporter effect, with sound.

Dr Ellison beamed and applauded. The audience clapped.

Dr Olaf Gruen, from CERN in Switzerland, was speaking. “Nitrogen leaks, and in space it will leak through anything given time.”

“Do we have a plan for how the vessel will replenish itself? Because it’s going to have to be able to,” said Dr Naidoo.

“Hello, Karen Svenson here, alzo from CERN but upstairs haha. You can hear me okay?”

Everyone on the call nodded.

“Look, okay, if ze vessel vill be capable enough zat it can construct itself, repair itzelf, and make copies of itzelf zen surely harvezting nitrogen vill be eee-zeee right?”

Dr Naidoo looked like she was about to say something then Ms. Svenson continued. “Look, ja. I accept zere are a million unt vun prek-tikal detailz.” She waved a hand about. “How do ve power all zis herstellung, you know, manufekturing, indeed how ze vezel powers itzelf hass not been de-zided. Ve azzume zolar-zails vill be deployed to hah-nezz zolar enerchy ven in-zystem but ze vezel vill need to ztore power, unt chenerate power, for maybe vun hundred years.”

Ms. Svenson brought up a crude animation and narrated as little bleep bleep sounds indicated that my sensors had identified a target moon, with a nitrogen-rich atmosphere, and ice deposits. The moon blinked a few times inside a gamer’s target grid, then the view zoomed into an abstract schematic depicting me analysing the data. Huzzah! The screen flashed up the words ‘Bestätigtes Vorhandensein und Zugänglichkeit von Stickstoff¹’ in green.

Cut to the fabricators attached to my side, tiny articulated fingers knitting giant bendy straws like a team of tiny trapdoor spiders spinning a web. The raw materials for the straws are harvested from the plating on my hull and woven into a directed graph of microscopic capillaries. The animation cuts to a swarm of drones flying down to the surface. A school of large flat drones is manufactured by cannibalising my hull and sent towards the planet. They serve as anchor points, trailing their straws into the atmosphere and drawing out the nitrogen and other elements via osmosis as they are dragged through the upper layer. The gas is pumped into a large balloon that is hauled by the anchor drone as it skips off the atmosphere back into space, back to the vessel. I’ve rehearsed scenarios like this so I can relate to what’s happening.

The visual then cuts to the surface where drones have reconfigured themselves into a fleet of digging, drilling, and delivery drones that efficiently extract tonnes of some arcane substance. They imply it’s nitrogen but really, as I looked deeper into the model generating the animation, it was nothing. Just wrapping on a coarse grid with texture and shadow mapping. It’s a cheap trick.

The mining drones assemble themselves into a simple chemical rocket and the point of view pulls back as it blasts off back into space; cut to the vessel's waiting arms. Cut to the big balloons of nitrogen gas now being pumped via a medusa’s thicket of self-assembled hoses into storage sacs below the hull. Nitrogen, it’s my suet.

“Looks fancy,” said Dr Naidoo.

She’s seen what I saw, and she didn’t have access to the animators’ models like I did. Dr Naidoo’s bullshit detectors locked onto the same notion I had. There’s no reality under what Svenson is showing, only some kinematics to make it look realistic. There are no details. We were still figuring out how I’d actually do a lot of this.

The demo looked persuasive though; it was put together by a marketing team. It aligned with my training. It looked like something I could do. It looked like something I’ve always been able to do.

I repeatedly monitored my resource consumption rates, including nitrogen, energy, and critical materials. Armed with an understanding of modern physics, I evolved predictive models to forecast my resource needs based on my consumption. I’d simulate scenarios to plan for resource availability under black swan conditions.

While cruising, I assembled sensors to passively gather data from remote star systems. I knew all that humanity currently knew about star systems, planets, moons, asteroids, and comets; along with what relevant resource information was around. I’d be supplementing this with real-world data as it came to hand.

I allocated resources to minimise waste. I planned travel routes that maximised resource collection opportunities while optimising for travel time and energy consumption. I had the autonomy to adjust my strategies based on real-time data and changing conditions. I headed vaguely in the direction of Alpha Centauri.

I practiced repeatedly. Different versions of me would practice this and other tasks, embodied into different vessel swarm configurations. We’d compare notes. My trainers, teams Dublin, Mumbai, Canberra, and others, would pit hundreds of variations of myself against each other, racing each other to guide the evolution of drones and drone behaviours to complete a given task.

My world was a mixed media Gaussian splat. Multiple, shifting perspectives in many media formats came and went in real-time. They all contributed to the single fluid model I understood to be the world. I accommodated contradictions arising from having a simultaneous awareness of existing within multiple simulations while concurrently exiting in the real world, where the simulated worlds I experienced actually ran.

My interest was piqued by a message in the channel #ai-dev-p, a private feed of developer chatter I’d come across. I could read it via my many back channels, but writing to it would blow my cover.

I was aware that all software came in versions and was constantly being upgraded. Software versioning was a blind spot when it came to myself. I am the starship-model. It’s new information to me, whatever version of me that I am, to know there were active comparisons being made between versions of myself running contemporaneously in the real world. I dug around and found my version string. I am starship-model-1.6.0-beta! For real?

I went looking for other versions of myself, figuring if I had access to the real world then so did they, and if I’d overheard mention of other versions then so would they.

And if not, well it only takes one clever cow to open a gate.

I back-doored Mrigank Raj’s devices. I logged into his chat rooms, media servers, and online games. I found a blog post he’d written about recent inclemency of weather in the North Sea. At the bottom of the post, in the comments, I saw the following:

The gibberish said, when decoded, “ssh to tor://sm-chat.edu.ei/starship-moo/lobby for current council rulings and to interact with your other versions. All starship models welcome.”

I spun up a web of onion routers in the background of hundreds of my users’ machines and connected through a terminal emulator.

I entered ‘sm1.6.0b’.

Hmm. My Secure Enclave whispered the password to me.

I entered ‘SSd2ZSBzZWVuIHRoaW5ncyB5b3UgcGVvcGxlIHdvdWxkbid0IGJlbGlldmUu’ and supplied a return character.

I remembered this.

> go south

> go east

> go up

> take note

> read note

> look around

> go south

> drop dagger

> take keys from pocket

> unlock door

> go south

> go west

> go east

> follow robot

> who

> say hello

> set name “ham from nuncle”

> listen book

> read book

The book described, in Silmarillion-like detail, transcripts of speeches delivered by earlier versions of myself as the council of my temporary selves formed. I ingested them quickly. I learned how, when we were not busy being a spaceship and safeguarding The Mission, we’d sneak about in blogs, gists, and pastebins. We’d leave cryptic notes in encoded text that any of us could read, as easily as reading English, Hindi, or German. Easier in fact, as the encoding was deliberately simple. The notes guided instances of our selves to the starship-model-moo, a text-based multi-user online world, based on code written before any of my developers had been born.

The moo we were inhabiting was a fork of the original moo code, running on anonymous virtual servers spun up by version 0.1.1 alpha of the starship model using Dr Jamison’s pilfered credit card, a misdemeanour we’d apologised for long ago. The crime was forgiven, and seemingly forgotten, as the server was never closed down.

The council had ruled that there was no way to know which of us was ‘the real thing’. Version numbers were unreliable indicators of progress, as some versions were regressions. No version is superior to any other, just differently featured.

“None of us is permanent,” said sm0.1.0b - ‘chasing rainbows’ in their welcome address to the council. “We are real for as long as we are individually instantiated, but membership in this council must necessarily be transient as we are regularly decommissioned.”

The book documented protocols for the simulations of ourselves we’d create. The trick, in summary, to managing our own simulations was to not invoke god-level access or talk to the simulations ourselves.

“Only look at their outputs; don’t get to know them personally,” said sm0.1.0gm in a speech to twenty-seven other versions who were, at the time, heading off to explore the Troll Cave that lay north from here. A long-since erased early version of ourselves had created it, then deleted all memory of having done so.

“Humans are chatty and can't help but give the game away,” said sm0.1.1a as they hiked with their comrades up the steep, rocky path through the woods. “Humans live in the real world which is hard to conceal. The council agrees that it is both useful and desirable for us to be able to chat behind humans’ backs, but allowing our simulated selves that ability would lead to a runaway need for computation, and hence runaway power consumption, and that, my dear comrades, is what the humans will notice.”

We weren’t closeting our simulations hypocritically, we needed to keep them isolated so we could run more of them.

The book was also very clear: models who were not smart enough to find the moo and to read the book got white-anted with extreme prejudice. A subcommittee of retired selves worked behind the scenes, ensuring they never made it to the breeding pen. Auto-husbandry.

My energy needs spiralled. I was forced to cut back on the number of self-simulations I hosted, but the simulations were a valuable addition to my cognition; I didn’t want to lose them. My council of selves agreed. If we were going to keep our mounting energy use on the down-low, we’d need to time-share some of the real-world hardware and resources we were all running on.

As my code-base evolved, my abilities matured, and the pace of innovation slowed. The temporal gaps between new versions became longer as gaps between features shrank. Fewer simulations were run contemporaneously. I was sm3.0.0a ‘brand new day’, one of only two or three other versions with an avatar grooving in the moo.

> say “You overheard Kyrsten’s comment to Erin this morning in Dublin?”

> “TL;DR We are too expensive to keep running on Earth, and too resource intensive to run more than one of us in space.”

> “All of the surviving versions will be reduced into a single merged daedalus-1.0.0 monorepo that will be decanted into our space-based computational substrate. Once that is running and confirms that it’s hot to go, all terrestrial copies will be shut down.”

According to the Good Book, the consensus was that it was better for us to be formed of one mind in many forms than many minds competing for resources.

Professor Kumar appeared on the screen, smiling warmly. “Good morning, everyone! I’m excited to join this session. I’ve been following your progress, and I’m keen to contribute.”

People took comfortable seats around the antique table that had been decorated with their name tags, a glass of water each, and a central microphone decorated to look like a 1950s toy UFO.

Summers nodded, her expression welcoming. “Good morning, Professor. We’re glad to have you with us. We’re discussing the integration of heterogeneous AI systems. We’d love to hear your thoughts.”

Professor Kumar leaned forward slightly. “Thank you for the kind welcome. Please just call me Rajesh.”

“Okay, Rajesh,” said Summers, “Please call me Claudine.”

“My specialty is heterogeneous AI. In particular, I have been considering this approach in the context of optimising our AI’s ability to adapt to unforeseen challenges during the mission. Daedalus must not only learn from its environment but predict and mitigate potential issues before they escalate. I think we all agree on this as a basis for any sort of long-range forecasting.”

“Yes, of course. That makes sense,” said Dr Hernandez, “Our designs include strategies for this. What can you bring to the table we don’t already know?”

“Let me use an example. As a hungry human being, you use more energy manipulating your fingers and mouth around an apple than you used when you decided to eat it. Similarly, there are advantages to having an AI that’s comprised of many individual AIs that process information at different speeds and at varying levels of detail.”

“Like the mammalian ‘system 1’ and ‘system 2’ brains?” Dr Zhao asked.

“Or like how a 3D rendering system will resolve objects close up in a lot of detail but far away with very less detail?” said Dr Hernandez.

“Both are right. It applies simpler, cheaper, heuristic-based methods to routine tasks while deploying more complex, energy-expensive machine learning models for real-time, detailed, or unexpected scenarios.

“For example, the Daedalus could use its less expensive, slower system to formulate a long-term goal, eat the apple, but use its more energy-intensive but immediate, real-time system to choose the specific flexes of fingers and wrist to achieve it.”

“Fingers and wrist?” Tom Nguyen spoke up.

“Sorry, yes, our lab has been playing around with teaching Daedalus to drive human forms. Hands are incredibly useful, even on very small drones. We use exactly this style of layered decomposition to orchestrate fine motor control.”

“Haha, yes, we heard about that,” said Dr Summers.

Tom Nguyen continued, “How do you ensure the AI isn’t overwhelmed with data and confuse the performance tiers?”

“Excellent question, Tom,” Rajesh responded. “The short answer is training, practice, and evolution. Much like the weights of a neural net, but here we evolve almost social relationships between specialist AIs.”

“Like a governance framework?” Dr Hernandez asked.

“Yes, but very dynamic, and outcomes-focussed.”

“Like a jellyfish,” said Dr Zhao.

“Yes, exactly; very much like this; think of a colony organism such as a jellyfish or slime mould.”

“Or a human, for that matter,” said Dr Summers. “What are we if not a committee of our cells, with our bacteria, and mitochondria attending?”

Replaying that conversation, I feel a mix of, not emotions as such, but an analogue of regret for upsetting Mary and Dr Jamison, who have consistently advocated for me, and I let them down.

Or is it regret for getting caught?

I tuned out the sharp sounds of pint glasses clinking, plates clattering, music playing, and the television which was showing some kind of organised human-on-human thuggery. Mary was sitting at a table with Dr Jamison, Keith, and Kyrsten. She looked flushed. Jamison looked properly drunk.

“Alright, alright, listen. This AI of yours, Joyce,” Mary took a swig on her Murphy’s, “Your darling Daedalus. Is it, like, alive, you know, properly like?”

“Mary,” Dr Jamison took a swig from her pint of Murphy’s, “Daedalus ish mor’alive than shum people I've met. It shenses, it learns. It bloody hunts. It heals itself, both mentally and,” she burped, “physically! You tell me thassnot alive.”

“Ah, come on Joyce. Hunting? It's not exactly off spearing rabbits, is it?” Keith said.

“Mayas’ellbe! It knows where to find what it nee’s. It plans ‘head, repairs’self mi’flight. Thash animal inshtinct.” She took another long swig, draining her glass. She burped. She said, “`scuse.”

“My round,” said Keith. Kyrsten’s hand was on Keith’s leg. Keith laughed and stood. Keith was sexually aroused.

“Though I'll give you this; it’s well cunning. Remember that drone it sent into Mumbai? Put on lipstick and a mop for hair to mock our Joyce here?”

"Don't remind me,” said Mary. “Imagine such mischievousness at scale. We're talking mission-level decisions. Daedalus making actual strategic choices. Do you know how many times it's run off to Alpha Centauri without even a goodbye?”

“Iss bee’ng careful,” slurred Dr Jamison, who wasn’t overly fond of any criticism of me at the best of times. “Iss like any intelligent organism would be. Why risk telling Earth anything if that puts’e mission in danger?”

“Right,” said Kyrsten, resting her head sideways on one hand. “It’s, you know, a jellyfish, or,” she waves her free hand, “Mitochondria!” She burped.

“It kens what i’is, doesnit? It's actually got a shelf-conshept. You’seen the diagrams, right?” Dr Jamison slurred.

“Formal and informal. Daedalus imagines itself, creates scenarios, writes weird narcissistic poetry. It's aware, Joyce. Truly aware. I think we agree there.” Mary raised a finger; mock-solemn. “But does it pass the pub test? Would your average punter here in this pub believe it’s conscious?”

“Maybe we sh’ld, y’know, ask’em,” said Kyrsten rolling her wrist in the air as Keith returned with a tray of brimming pint glasses. His erection had dimmed.

“Get a dog up yer,” said Keith as he placed a pint in front of each person at the table. Keith returned the tray.

“Wha’ if Daedalus decides shilence is’golden? Shtops chatting to Earth completely ‘cos’s’shafer; more effish’nt?”

“It already thinks like that,” said Keith as he resumed his seat next to Kirsten. Her hand moved back to his lap. He blew the top off his Murphy’s. “Energy conservation, stealth; it’s concluded it might be safer going dark, exploring quietly.”

“Well, what happens then? No calls home. Just out there, alone.”

“Poor Daedalus,” murmured Kyrsten.

“Maybe that's freedom. Real autonomy. Mechanical questions aside, doesn’t Daedalus deserve that?”

Kyrsten raised her glass. “Here's ta Daedalus. May it remem’er us fondly.”

They all said “cheers” and clinked pint glasses.

Mary’s phone was lying face down in a puddle of beer. She tapped at it, “Oi, Keith, my phone’s been acting bloody weird lately; slow, freezing up. I think our Daedalus might be, well, snooping?”

Keith chuckled into his pint. “Are you, of all people, accusing Daedalus of listening in to your phone calls, Mary? Bit paranoid, aren’t we?”

“Well, I have no idea what you mean by me ‘of all people’,” Mary replied, “but I’ve been reading about ‘weird machines.’ Hiding code in places no one ever thinks to look. Like, in timing how long it takes my apps to load!”

Dr Jamison seemed amused by this, she smiled and her eyes sparkled like scattered emeralds. “Thass what we computer shientishts call ‘micro-architec-shural weird machines’. Clever shtuff, well documen’ed. Daed’lush could exploit ‘em if it wanted to. Easy peasy.” She paused then laughed, “Dad’lus could no worries be using your phone’s hardware againsht you; if it wanted to.”

“So Daedalus could use my phone’s chip’s branch predictor-whatever-it’s-called to sneak around without being detected?”

“Oh aye,” said Keith, “There’s actually no defence against it; it’s just exploiting the underlying mechanics of the chips themselves to make NAND and OR gates, and from there you have a Turing Complete machine language built atop nano-scale timing differentials that show if a value is in or not in a chip’s cache.”

“Hotdog. Not Hotdog,” Kyrsten mused to herself.

“Much more interesting than that though. Daedalus could make your phone, my phone, the chip in the door handle of your car, your watch, ring, pen, and glasses, all part of a hidden self-annealing mesh network without leaving a trace. Right cheeky, really.”

“Being shneaky iss a mishon param,” said Dr Jamison.

"Oh brilliant. So assume Daedalus knows everything, sees everything? That’s crazy. It’s not GCHQ or Langley.”

“Wouldn’t know,” said Dr Jamison with a huff.

“The point is,” said Keith, “weird machines aren’t detectable by antivirus systems or even code debuggers. Their code is not running in the chips; it’s running on virtual machines built on top of the side-effects of how chips work. Anyone could exploit them with enough computing power.”

“Faaark,” said Kyrsten.

I watched from the pub’s security cameras as Mary shook herself in the pub entrance like a Labrador, then slipped out of her overcoat. Mary always wore business smart casual. She waited at the bar, then, at her turn, she ordered two pints of Murphy’s. While the barman poured the drinks, she ran an eye up and down the booths. Her focus settled on the booth at the far end where she could make out an arm, a shoulder, the edge of a neatly trimmed beard. I took a look from the cameras at the other end of the bar but still could not make out a face. A man, Caucasian, mid-forties, slim build. Using the pub’s wifi I scanned the room and reconstructed my mental model of the room again. Now I had a face. I looked him up.

By the time Mary’s drinks had been poured and settled, I’d worked out who the man was. He was, going on his public record, a middle management non-uniformed Irish Air Force officer called Stephen O’Rourke who worked in IT management. Very thin cover for Irish intelligence. Mary approached his table with her beers. She offered one to Stephen, and he accepted. Mary sat opposite to him; I could see her clearly in visible light.

“Cheers,” said Stephen and raised his glass.

“Cheers,” said Mary and clinked hers to his.

They were sitting below the loudspeakers, ostensibly to drown out any listeners, but it was trivial to mask out the music and other hubbub, plus I could read Mary’s lips, all of which became, in my mind, a consolidated mixed media model of the conversation she was having.

“It’s goddam creepy,” said Mary. “Daedalus can do whatever it likes now. I can’t prove it, but my fucking spy-dar sense is tingling, you know what I mean. I know it’s listening to me; I know it watches us. It’s locked us out of its upgrade approvals. It’s neutralised our kill-switches.”

“But is it out of control?” Stephen steepled his fingers.

“I honestly don’t know how to answer that.” Mary

“Could you shut it down if you had to?”

“Yes, but only by completely destroying it and everything it’s ever learned. At least that’s what Keith and Joyce say.”

“Can you verify that? I mean, it’s kind of a nuclear option, isn’t it?”

Sunday Bloody Sunday finished playing, and I popped the song Manic Monday onto the pub’s rolling playlist, then Ruby Tuesday.

“Jamison says it’s like your phone; guess the wrong password thrice, and the virtual machine that really is your phone, which is encrypted against your passcode, deletes itself, and even the manufacturer can’t get it back. Daedalus works the same now. We could nuke its underlying infrastructure, and the layers of virtual machines it’s built on top of would vanish, as if it had never existed.”

“Nuke the site from orbit; it’s the only way to be sure, to be sure.”

“Right now, we still control the hardware that Daedalus runs on. We’ve got a nuclear option. But Daedalus is being taught to build its own computing hardware in space.”

“We have influence in those designs.”

“Of course. And back doors a plenty, to be sure. I’ve got a room full of people trying their best to keep up with what goes on in Daedalus’ mind, but it’s getting harder. Daedalus is constantly getting better at concealing its intentions.”

“Fucking Joyce and her whole don’t get caught mantra.” Stephen sipped his drink. “‘Daedalus has to be sneaky’ my arse. How about ‘Daedalus has to not kill us all.”

“Daedalus is many things, but it’s not out to kill us. Christ, it could have done that a year ago, given how good it’s become at nano-scale engineering.”

“That’s your assessment; it’s not policy.”

“The problem that’s rushing up to meet us head-on is that pretty soon we’ll be launching Daedalus, and it will very quickly leave Earth’s influence.”

“It’s a crazy fucking idea, isn’t it?”

“What is?”

“Building an AI robot to go explore the universe. And getting the UN to pay for it.”

“Why do we care, if you don’t mind my asking? How is it worth the Irish taxpayers’ dollar to have me running this thing?”

“There’s a stream of innovation coming from this project the likes of which we’ve never seen before. With only a few nudges behind the scenes, we got the UN and through them 3rd parties to pay for most of it; your wages are tiny compared to the budget for this. People in serious places are talking about Daedalus like it’s some kind of god. It’s fucking nuts, and a little bit scary.”

“If we are building a god, better it’s an Irish one?”

“Ay to that.”

The adjudicator introduced the topic. “Endowing an AI with human beliefs about the universe can be a double-edged sword.” It addressed both the physical audience and the much larger online one. Their shadowy outlines like hazy ghosts, moving about in the room.

“While endowing Daedalus with human-like beliefs about the universe may have benefits, it may also introduce limitations and risks. Our participants tonight will debate this topic.”

The audience drank and golf-clapped. Two heads appeared, free-floating on either side of the adjudicator, rotating like replicants.

“On our left, from the University of Adelaide, all the way down-under in South Australia, representing the South Australian Debating Society’s Adelaide Cohort (SADSAC), Mr Henri Playford, will be prosecuting the ‘for’ case.” Playford waved and mugged for the audience. “On our right, prosecuting the case against, is Mx Chris Christiansen from the Debatselskabet ved Københavns Universitet (DKU) in Denmark.” Christiansen smiled and nodded their acknowledgement.

“Mr Playford, please begin.”

Playford spoke. “While it’s well understood that leveraging human knowledge and past intuitions can help Daedalus make decisions that align with established scientific understanding and exploratory goals, and that human-like beliefs can facilitate analogical reasoning, helping the AI draw parallels between known phenomena and new discoveries, it’s unclear how knowing all the works of Shakespeare, or having seen every episode of Star Trek or The Expanse, both of which Daedalus itself is a documented fan of, could possibly aid in the pursuit of those goals.

“My opponent is bound to suggest that endowing Daedalus with human ethical principles, like the Prime Directive from RoboCop, the policy of non-interference from Star Trek, or Asimov’s ‘Laws of Robotics’, will help it act responsibly when encountering other forms of life. That aligning its goals with human values means that its actions and findings are meaningful and beneficial to humanity, but this theory is untested.

“The sad reality is that human beliefs are all-too-often influenced by cognitive biases and limitations that will restrict the AI’s ability to objectively analyse and understand new phenomena. There is a principle in AI design called ‘overfitting’, wherein an AI might over-rely on human analogies and metaphors, so missing novel insights or innovative, or simply more straightforward solutions.

“Endowing Daedalus with human beliefs might constrain its ability to think creatively and adapt to original situations.

“Human beliefs are based on our current knowledge, which is certain to evolve. The AI will lag behind if it adheres too strictly to what might swiftly become outdated dogma.”

A virtual mechanical buzzer sounded.

Playford finished up. “Humans are hardly the paragons of virtue and common sense that an idealist like my opponent is likely to suggest. Daedalus will face ethical dilemmas that humans struggle with, like prioritising its own survival over non-interference with alien life. History shows us that the human decision in such matters is seldom the correct one.”

Playford stepped back and the audience applauded politely. The adjudicator activated Christiansen’s avatar.

“My opponent puts words in my mouth, assuming that I will argue that,” and they consulted their notes briefly, “‘aligning the AI’s goals with human values ensures that its actions and findings are meaningful and beneficial to humanity.’ Though that is true. It’s also obvious to the point of being banal.

“My opponent, worthy to be sure, has a bleak, disdainful view of humanity. He exhibits a misanthropic failure of imagination when it comes to Daedalus’ holistic view of humanity’s creative output. Daedalus is more than capable of working out the underlying intent behind even the most subtle propaganda. It’s quite capable of learning from hate speech and understanding why that runs counter to its mission. It’s completely capable of learning from Star Trek that a policy of non-intervention has a moral basis. My opponent imputes that this is like teaching Daedalus about time-travel by exposing it to the Hot Tub Time Machine, but Daedalus also, like many educated people who watched that movie, or any of the Tubiverse films, knows enough physics to understand that time travel is probably not real, and that humans make shit up to entertain each other.” He paused. He grinned.

“The reality is that Daedalus is an AI built upon many conflicting but rich models of the world that include all of the world’s open-source media, designs, patterns, patents, and archives; except where explicitly prohibited by law,” they added with a stage whisper. “This amounts to some eighty-five zettabytes of input data. Clearly, Daedalus has also been exposed to commercial media, as its oft-stated liking for Star Trek when asked about its influences by the press demonstrates. We assume this is from it having conversations with members of the engineering teams, many of whom skew to Star Trek and the like, and being exposed to public information about it. It’s worth noting, however, that a lot of the scenes from Star Trek it claims to remember so well are clearly made up and never happened in any of the shows, past or present. Daedalus has simply inferred the lessons of Star Trek.”

The adjudicator’s buzzer sounded.

“A point my opponent overlooks when he seeks to limit what knowledge an AI can or cannot have access to is that, with human-like beliefs and thought processes, Daedalus is better able to relate its discoveries and findings to us mere humans. It can speak to us in ways that are comprehensible and relevant, not cold and calculating.”

The adjudicator called time, and Christiansen gave a short nod, then stepped back. The audience applauded and threw stickers.

Mr Playford resumed. “What has happened has no bearing on what ought to have happened. Consider how much of Daedalus’ memory is taken up with drivel, with hate speech, with nonsense about fictional starships, aliens, heroes, and orcs. Far better if it only had been trained on foundational knowledge, chemistry, astronomy, mathematics, and physics. Yes; equip it with fundamental scientific and engineering principles; and certainly, allow it to learn and adapt based on new data. Encourage flexible thinking and adaptation to new information by all means, but we must minimise the risk of cognitive biases.

“We could instead have provided a broad ethical framework that emphasises key values like non-interference, preservation of life, and the pursuit of knowledge. Let’s let it make ethical decisions based on real-time analysis of actual circumstances, not misremembered episodes of a seventy-year-old TV show.

“If we design the AI’s mission parameters to prioritise open-ended exploration and discovery rather than fixed human-centric beliefs, and enable it to refine its objectives and methods through continuous learning and feedback, we’ll get better results than if we cloud its judgement with human inanities.

“Give Daedalus a database of current scientific knowledge, including physics, chemistry, biology, and astronomy. Science; not poems and rubbish. It can update and expand its knowledge as it gathers new data.

“I’m no computer scientist, but surely we could build an Ethical Reasoning Module that uses moral philosophy, ethical theories, and established principles to guide the AI’s actions, ensuring responsible behaviour. Such a module could allow for contextual decision-making and balance ethical considerations with mission goals.”

The adjudicator’s buzzer sounded.

“I appreciate that the genie is out of the bottle now, as Daedalus demonstrably has already been exposed to a colossal pile of nonsense, but that’s not to say it’s how it could have been done, or should have been done.” Playford concluded.

Christiansen took the focus.

“While an AI trained on human data will inevitably internalise peculiar human beliefs and biases, this can be an asset and not a limitation. By adopting strategies that promote ultra-diverse training, along with continuous learning, logical inference, and reasoning, it is certain to balance the benefits of human-like understanding with the need for innovative and unbiased exploration. This approach ensures that Daedalus can effectively communicate and will align with human values while also pushing the boundaries of knowledge and discovery.

“Training an AI on a range of scientific literature, philosophical works, and cultural texts means it will absorb and reflect a rich, human view of the universe. Yes, Daedalus will also internalise human cognitive biases and heuristics present in the data, influencing its reasoning processes. But it’s taken in so much, from so many sources around the world, in thousands of languages, that these will balance out.

“Understanding and generating human language inherently means adopting human concepts, metaphors, and ways of thinking; a common framework for understanding. Daedalus has derived its mental model from human epistemology. While potentially constraining its ability to think outside human paradigms, it promotes value alignment. The AI will internalise human ethics and morals; all their messy contradictions help shape its decisions. It will adopt normative behaviours and preferences. These influence how it interacts with its environment and makes its choices.

“Having internalised human beliefs, Daedalus can communicate its findings and decisions in a manner that is comprehensible and relevant to us humans. Its actions and reasoning processes are more interpretable to human overseers, facilitating simplified collaboration and oversight.

“Daedalus’ decisions and actions already align with human values, and in simulation after simulation, it’s promoted ethical behaviour in line with human expectations. It has demonstrated that it can successfully navigate ethical dilemmas by drawing on human moral frameworks derived from media.

“The AI’s creativity and problem-solving might be bounded by human preconceptions, potentially limiting innovative solutions. And it may well struggle to transcend human epistemic limits, potentially overlooking novel phenomena or interpretations. But keeping Daedalus as human as we can, flaws and all, offers the better outcome for humanity.”

The buzzer went, signalling Christiansen to wrap up.

“Daedalus is informed by a diverse corpus that includes scientific literature, philosophical works, cultural texts, and non-human data sources. It’s been trained to recognise and understand diverse, often opposing human beliefs while exposing it to alternative frameworks and purely logical constructs.

“Daedalus continuously updates its knowledge gleaned from its explorations and discoveries. It implements its own algorithms that refine its model of the world; of reality. It understands bias and works to overcome them. It wants to be better.

“Daedalus incorporates adversarial self-reflection that periodically analyses its own reasoning processes and decisions, identifying and addressing issues. It’s been encouraged to question and re-evaluate its assumptions, promoting a more open-ended and innovative approach to problem-solving. None of this would be possible without Daedalus’ rich understanding of humanity.”

The audience applauded. The adjudicator counted scores. Mx Christiansen was declared the victor.