Everything Becomes Connected

“I trusted you, Tristan !”, bellowed the bedside speaker.

Tristan rolled back, groaning, yawning, having failed to decline the call with a flailing hand

“What is it, sis? And good very early morning to you too.”

“It is not a good morning. And you know what it is. My lab called to let me know that my brother, who I trusted, had stolen a medical sample from the lab.”

“What …?”

“They saw you on the security camera.”

Damn. Awkward.

“A new source of urine was one of my principal targets.”

Target no.1, in fact. Sources of non-human urine.

“You said you weren’t doing that any more.”

“That baby porpoise in your lab was dead. There was no harm in taking its bladder. Urine from premature porpoises could repair damaged kidneys, or help bind better paint, or make new perfume, or cure cancer, or or something.”

“Are you serious? What do you know about any of that? You’re no biologist. You could’ve been, if you’d completed uni.”

“Don’t start that again. There’s no future in …”

“… experts. Yes, I know. You keep telling me. Every time I ask what you are doing with your life.”

“I’m making good money.”

“By stealing from your sister and the lab. This is a new low. Do you steal money from home too?”

“It wasn’t stealing. Ah, ok, maybe you might think of it that way.”

“You know what you are?”

“Making money? Seeing the world? Learning new things?”

“No. A body snatcher, digging up corpses for illegal medical research. And a drug addict. You’ll do anything to feed your habit. You don’t care who you hurt along the way just to get your fix.”

He guiltily checked his screen again. There, near the top, the line said 120% of quota achieved this month.

“Look, you said it was just a formality, a run of the mill post mortem before incineration. No harm done?”

“That’s not the point. What about trust? What about I can’t let you near the lab ever again? Or my friends”

“Sis …”

“So, those experts you don’t need? If you’d asked one, they might have told you’d pulled out one of the testes.”

The call ended abruptly, the sibilance of “testes” hanging in the air.

Sleep did not return, so up and breakfast.

Awake and focused, Tristan returned to the fridge and retrieved the sample bag, struck out the word ‘bladder’ and added the word ‘testicle’. Initialed, photographed, replaced the bag in the freezer. A cup of coffee, then online to update the description.

The dashboard updated immediately. Delivery urgency dropped from “immediately, by cab «click here for cab token»” to “within 3 days”, and the commission points halved. Ah well, easy go.

Exercised, lunched, Tristan addressed the issue of his sister’s ire.

He messaged:

Because I’ve been doing so well at my waste of a life, I have levelled up to Senior Scout. And with that comes a higher commission, more autonomy, and an invite to tour one of the processing facilities. Do you want to come along?

[silence]

This is your chance to get a look on the inside and ask all your questions.

[silence]

Go on. I’ll even say I’m sorry.

Sorry.

There. I said it.

[silence]

Now you have to come with. It’s the rules.

Izzy?

[silence]

I did think that li’l porpoise was due for the bin minutes later. It’s not like it was a human corpoise.

Not funny.

The processing facility was loud, annoyingly rather than painfully, as the sound echoed in the high hangar. Lots of space to expand into.

Once they had their visitor passes from the welcome desk, the introductions took place in an office in the small, temporary-looking, upper level overlooking the floor.

“I am Pat, Patrick Sears, Customer Relations for Core. Thank you for all coming. I hope you find it educational and inspirational. This is a glimpse of our near future. One that we will all be part of in experiencing and, importantly, shaping.

“I’ll give you a summary of how this site operates. Questions any time. We’ll go walkabout downstairs and I’ll be your tour guide. Returning here for final questions.

“We’ll start with Core’s mission.

“You’ll have heard the phrase, ‘Correlation is not causation’? Well, we have a response to that: ‘When you have a trillion correlations, it is not far off’. And that, in a nutshell, is what we are about. We are aggregating all possible facts and snippets, and pouring it all into our Core engine, looking for correlations, all coincidences, all associations, all links between all things. These links could be hints of new laws of physics, new materials, new psychiatric behaviours, new drugs.

“This approach is how Google automated translations. Not by building a complex AI that understands what it is reading, but by accumulating millions, billions, trillions of examples of real translations, looking for correlations. After a certain threshold was passed, the correlations derived from those many many examples produced good, usable translations of new texts.

“Edison, Thomas Edison, is another inspiration, with his development of electric lighting. He made a great use of theory, where it existed, but where it was absent, he industrialised the search for useful combinations of raw materials.

“For us, facts are just the beginning, the data points, the raw materials. And correlations are the combinations that may, or more likely not, be useful. Perhaps linked in some meaningful, useful way.

“The Core project is the Human Genome Project splitting apart DNA for useful fragments. It is the Large Hadron Collider smashing together items to explore their make up. It is Cyc capturing all common sense knowledge. It is all of these things and more. We take in anything, be it substances, ideas, facts, smash them together and study the pieces for meaning.

His enthusiasm was infectious. Tristan was pumped. This was why he had chosen this career. He looked at Izzy, who glanced back, shrugged, and wobbled her hand, still so-so.

The tour continued downstairs.

The site operated much like Accident & Emergency triage in a hospital. Tristan had dropped off his sample in a freezer box while Izzy was looking elsewhere - now was not the time to discuss which sample it might have been. The box was scanned, assessed, placed on the floor, and glided away briskly to and through one of the nearby low hatches.

The tour congregated in a positive pressure room for ‘enzymatic assays’, the inflated roof bulging. Each room specialised in a different physical or chemical test.

Izzy piped up. “But we are in here with dirty outdoor clothes, breathing, coughing.”

“Yes. Rather than lock everything down all the time, we log the fact that there was this specific group here during these specific tests on these specific machines. The full list, the context, is quite extreme. And link it to the outputs. If it proves to be significant, the system will request the tests be redone under stricter hygiene conditions. This applies across the site. We record the context with each and every data point. It may be that a context itself provides a correlation that triggers more interest.”

Back in the central corridor again.

“Where are the computers?”

“In the cloud, out there, along with the data. We just rent them. It’s cheap. On the other hand, here is where the costs are. Collecting the data in facilities like this.”

Izzy again. “How do you prevent illegally sourced samples being fed into your system?”. She didn’t look at Tristan, and neither, very carefully, did he look at her.

“Our scouts, of which we have many, operate under a strict code of practice where they first establish clear, unambiguous, documented permission before they obtain any samples. This all goes into the context. Any data and its correlations can be embellished or contradicted or removed entirely, at any time, which is a fundamental aspect of how the Core algorithms work.”

The tour returned to the raised office area. The muffled silence a relief. There were a few more questions, and a wrap up from Pat. The group, including Tristan slowly peeled away one by one back to the entrance desk. As Izzy turned to go, mind still troubled, Pat approached, “I sense you may have some more questions.” He looked down at the badge, “Izzy”. “Ah, would you be Isolde, Tristan Graham’s sister?”

“You know me?”

“Not directly, but we know Tristan. One of our better scouts. Some people seem to have a knack. And when it turns out they are related to someone with your skill set, things start getting very interesting.”

“My skill set? Tristan told you?”

“No. I assume you have a public bio where you work?”

“Ok, yes. That is public, I guess.”

“You are experiencing one of the main perception problems the Core has to overcome. Just using publicly available information, we can infer a very great deal. Anyone can. Governments and corporations. The advertising industry.

“There is no need for us to steal data. Eventually everything becomes connected, visible. Time is on our side.

“We are not interested in your and everyone else’s online behaviour. Well, we are, but not much. We want to pin down how the real world works.”

“Why do you pay your scouts so poorly?”

“We certainly do when they start. But the moment they demonstrate a grasp of what is needed, their salaries and especially commissions start ratcheting up quickly. Quicker than with most jobs. Do you happen to know what your brother earns?”

“No. He never said, other than he was happy with it.”

“Well, I’ll leave it up to him to share the details, but let’s just say that the next level up from where he is can expect to clear over £6K per month, and there are many levels higher than that. It is an irony of the times that we cannot envisage how Scouts, such as your brother, will ever be outperformed by robots. Homo sapiens evolved to be hunter-gatherers. Scouts are that in its purest form. The Core algorithm merely nudges them as to what they should look for, and then takes full advantage of what they find.”

“Our meeting, this conversation, not a coincidence?”

“Do you know of the DeepMind program AlphaGo, which became the best Go player in the world?”

Izzy half nodded. “I know it happened”.

“With an early version, the team took on, and thrashed, the then European Champion, before making improvements to take on the World Champion and winning comfortably.

“What is less known is that the AlphaGo team recruited that defeated European Champion to help them tune the later version and prep for the World Champion. He jumped at the chance to learn more about AlphaGo. Played relentlessly, morning noon and night, mostly losing, but uncovered a major problem. In large areas of the search space, AlphaGo was indeed more or less invincible, but in lots of small pockets, sometimes quite large areas, AlphaGo was not only not invincible, it was embarrassingly bad, where it would make the kind of moves even a beginner would know not to make. What is worse, AlphaGo did not know it was bad in those areas. You’ve heard of the Dunning-Kruger hypothesis?”

Nod.

“We need human experts to probe Core for systemic weaknesses of that sort, to help us expose them and alert Core to its own inadequacies.

“Allow me to introduce myself in more detail. I am Patrick Sears, Customer Relations.”

“You said that already.”

“And Head of Correlated Recruitment. It was partly my idea that led to you being here today. We turned the Core’s correlation algorithms on our own scouts, and noticed that the best scouts were connected with exactly the kind of people we need to help shape the Core. With a bit of a nudge to their picking lists, we can bring about collisions between the scouts and, well, you.

“Would you like to shape our future?”