When Computers Ruled The Earth

So, computers took over the world.

It started out as a joke. Fro worked for a multinational technology corporation of seemingly unlimited resources, limitless optimism and an ultimate (if unspoken) desire to turn the planet into one of the better science fiction futures, one of the sparkly supercomputer-powered universes where everything glows and is controlled by waving your fingers in your air, or just thinking into the receiver, your garbage can has a computer inside it and everything hovers and is connected to the internet. Shades of Gibson and Stross. Except, profitable.

These futures are unrealistic, that's what makes them fiction, but why aim below what you're capable of? The limiting factor in technological progress isn't the power of the computer or the intelligence of the human or even the speed at which the cable guys can lay fibre optics to the last mile of every square inch of the planet Earth, but the law. The trick is to invent things which no lawmaker has made illegal yet, and get it grandfathered in like tobacco and uploading.

So this company which we'll call MTC was in the business of acquiring other companies in the same way that a magpie collects shiny trinkets, where "shiny" here is code for a highly complex and subjective expression of every quality of the target organisation in question: its structure, size, assets, physical locations, mobility, corporate culture, profitability profile, product range, rate of growth, and amenability to acquisition. All these attributes (and pay scale, and corporate presence, and business plans, and and and) were researched to the ninth degree before acquisition began-- there were entire buildings of MTC employees dedicated to that research process alone. And making the decision-- this is the smart bit. That was left up to the suits. Businessmen in the sense of men with capacity for conducting business rather than merely working within the context of a business.

The decision.

But it was a joke. It was like watching dominoes fall, or dripped paint spread through water, or a Lovecraftian horror assimilating a city person by person. Seemingly every other week:

MTC is pleased to announce the acquisition of LProc Systems, a cloud computing research laboratory based in Minneapolis, Minnesota. LProc Systems' unique business position will enable MTC to grow its Everything As A Service business portfolio, enabling customers to easily create and configure virtualised hardware and software assets within a highly managed and highly available enterprise-level virtualisation framework.

They fought hard, but they'll come around. Martin Shelton, their CEO, quit, but everybody else likes the company car plan and their source repository was a complete trainwreck and needed an overhaul anyway. Branding? We're thinking... green.

By the way, the thing we were working on to directly compete with LProc? That's been cancelled--

So a bored group of MTCers made a videogame out of it. At M's core were two components: cool-ass retro computer hackery graphics, and a highly mathematical business management simulation where the aim of the game is to Build Unlimited Sustainable Business Value. You start as a pretty small company, with a big multidimensional (and initially largely empty) graphical model of your portfolio and some mind-blowing visualisations of the internal configuration and physical deployment of your company across the world. Elsewhere in the business landscape, other companies flare up, spread like viruses, go under and die. Some of them have what you want. Others want what you have. The aim is to acquire everything. Once you own the world, you've won-- unless you were playing in Hard Mode, where the goals were much more complex and sophisticated than mere megalomania...

It was just a game. It was great fun. It was a snarky and cynical commentary on the environment in which it was produced, but it had a firm collection of calculations underlying it. In 90% of scenarios, starting as a startup led to being acquired or near-instant death, but once acquired you kept very partial control over a small part of the parent company and after death you could just start another startup in the same "lifetime". It was, at its heart, a merger/demerger simulator. That's not the clever bit. Modelling increasingly intricate and important aspects of the company wasn't the clever bit. Enabling pot-plant-level micromanagement wasn't the clever bit, even though all of that was incredibly clever, thanks to the one guy in the bored gang who had serious mathematical chops.

The clever bit was Fro's AI. Every other business was a computer-controlled corporation. It evolved and grew and made decisions by itself. It made decisions about whom to merge with. It carried out business plans for maximum harmony. Like a human, it looked at the synergies and made a yes-or-no decision. This was the bit that meant they got to make M on company time. The algorithms and the equations. The Model Of The Business.

You can't simulate the entire business universe any more than you can simulate the entire real universe, but you can take two organisations and merge everything else in the world into a collection of background forces and see what shapes them. You can't predict the future but you can evaluate two hypothetical presents for relative desirability. You can make an educated guess as to how many people will be made happier.

Putting real business data into the machine had been a planned feature since day zero, but two years into the project the simulation of the 1950s got close enough to the reality that M crossed the threshold of real business usefulness. Paula Yerrick was the first person to put M in a sandbox on a private virtual machine and simulate a real acquisiton on which she was working just for kicks, to see what the likely outcome was going to be. And at this point in the story two alternate histories clearly diverge, because M was in the habit of giving meaningless (or worse, useless) responses due to failing to account for some trivial factor, sixty percent of the time, and Yerrick's results could have turned out like that.

But they didn't, and the acquisition unfolded much like M predicted, and Yerrick was impressed, and told two friends, and they told two friends. Thanks to the game's excellent (highly mature) usability and powerful capabilities for protecting -- and this was an almost brand new legal concept -- confidential game data, uptake was good. The project gained momentum.

After five years of growth (and one conscientious year deliberately set aside to shrink the game, pruning cruft and consolidating code into a sleeker base), managers were advising representatives to use simulations in their decision-making process. Some were even taking the game into talks and discussing the roadblocks and points of friction rising out of the models.


Business was like a fluid! Econometrics nothing, he studied population dynamics, differential analysis and chaos theory! He figured out how to put a distance metric between two corporate cultures! He modelled societies as a vector space to a better-than-fifty-percent accuracy! The man was unreal! How did he do it, how did he know?

Fro himself guessed it was coming, although modelling his own miniature business unit within the greater MTC was something he'd always shied away from, if only to prevent the universe from imploding. Only incredibly simple systems - static or at best periodic - can make accurate predictions taking into account their own future behaviour. The Higher-Ups decided to sell M. They bundled it up in blue and called it a productivity tool, a business management product. How could it so accurately account for disparate cultures and multinationals? How did it work so well in international discussions?

A 5,000-employee drilling company based in the American midwest merged with a Korean manufacturing conglomerate to astounding mutual benefit. It transpired that both organisations had used M to construct their models of each other, and had brought M to the meeting table, and had done what M told them to do to ease the process as far as possible. M effectively took off like a rocket bound for Saturn. The creation, refinement and sharing of public models of public organisations became a huge international focus. Soon the real world was close to being entirely partitioned into interlocking, organically interacting business objects. (With only one rule: never trust a company's model of itself...)

A time came when business was conducted with two simulations sitting there at the table guessing what would be the best thing to do. And no, M wasn't always right (althought that particular truism was something sternly suppressed in the marketing...), and M was never able to predict its own future capabilities, and the real world was different from any simulated model. Worse still, intelligent and devious people weren't above running gigantic extrapolation operations in order to predict and shape future events to their advantage... but that was no real change from the way stock markets had worked for at least a quarter of a century.

A time came when physical individuals stopped coming to the meetings. They hooked up a pair of banks of negotiation software with their business goals and philosophies programmed into them -- installations which were, thanks to MTC's LProc system, most probably living at the same physical location for minimum latency and superior uptime -- and waited for a yes or a no outcome. The recommendation was run past a group of human decision-makers, but the review became more cursory and the number of humans in the group dropped as years went on. Soon it was one. Soon it was none.

And then: program complete.

We lived in a world which rearranged itself electronically under our feet. We lived in a quintillion-dimensional differential equation in which companies and products underwent entire life cycles in seconds. They literally arose, named themselves, made names for themselves, unfolded, marketed, distributed their wares, merged, split, folded, shattered and disintegrated all in seconds, with the physical people working inside them none the wiser. This was a system which had been proven to work. A metabusiness model. A world without human decision makers but humans were still under control of all the physical decisions. It became easier to do what the machine told us to do than to do whatever, because the machines were invariably right on the money.

A complete monopoly never happened, because a strong business ecology (or any strong dynamical system, come to that) can't survive arbitrary perturbations without growth, change, diversity and a generous budget for innovation. Paradoxically, these had to be coded in as desirable goals. A point never came where M could be left to run unattended. Oddities and bizarre, pathological patterns -- some of them even viral -- appeared in the wild. Fro and his bored collective had to keep with it.

But a point did come when we realised that we had given over control of the universe to Rational Actors. And that people had the jobs they wanted with benefits. And there was a lot more money sloshing around and room in the budget again for the fancy trains and planes. A point came when the majority of people in the world had a dentist and a couple of pairs of shoes, because it was financially beneficial, everywhere in the world, for more dentists to see more patients, and for kids to have shoes and vaccines and bicycles and Opportunity.

But why hadn't the world overbalanced into hyperinflation? Some corporate nightmare future where you had to pay to open your own refrigerator, with Fresh Air Tax and no bootstraps to be found? Oh, that's the easy bit, gentle reader. MTC M Enterprise wasn't programmed to be capitalistic. Why, sometimes it spontaneously broke up companies for their own good, figuring that its employees would be happier overall under different management. It wasn't the happy little game where, like Risk, you are trying to conquer the world and then, implicitly, run it into the ground. M was not about you. It was released in hard mode, Serious Business mode, with customer satisfaction as its Zeroth Law. And after all, at the logical conclusion of the simulation, everybody is a customer of everybody else.

So, are you happy?

Discussion (17)

2010-11-02 23:35:06 by dave:

not bad at all

2010-11-02 23:38:53 by green:

Wow, very interesting! As entertaining as "Valuable Humans In Transit".

2010-11-03 00:38:48 by hppIII:

Great story! Minor typo: (althought that particular truism was something <b>sternyl</b> suppressed in the marketing...)

2010-11-03 03:58:51 by anonymic:


2010-11-03 07:35:01 by Dirkjan:

"And at this point in the story two alternate histories clearly diverge, because M was in the habit of giving meaningless (or worse, useless) responses due to failing to account for some trivial factor, sixty percent of the time, and Yerrick's results count have turned out like that." seems slightly wrong. The "count" in the end should be "could", perhaps? Great stories so far... looking forward to the other 28 already! A hell of a goal to set yourself...

2010-11-05 22:18:59 by qntm:

2036 words. Running total is 3974 words.

2010-11-08 15:39:15 by Vladimir:

Sam, have you read Marshall Brain's online story Manna?

2010-11-08 16:19:56 by qntm:


2010-11-08 19:02:27 by BrightMikal:

A little too technical for the average layman? I felt very lost in the first few sections, but understood what M had become by the end.

2010-11-10 07:22:31 by eneekmot:

"Are you happy, citizen?"

2010-11-11 19:45:10 by george:

Very original story

2010-11-13 02:41:12 by SoImNotAllowedToUseAFirstAndLastNameAmI:

Original? Well .... others have had the idea. Gerrold in When HARLIE Was One; Stross at the end of Accelerando (Economics 2.0). Maybe even Banks' Culture setting. The clever bit is that people are happy here. I don't believe it for a second.

2012-06-06 05:45:31 by Dustin:

Reminds me of Daemon by Daniel Suarez. Except with less killing.

2012-07-26 21:41:55 by PatCash:

Great Read. Found this through stumble upon and this is right up my street (maybe computers are already making decisions for me). Reminiscent of Borges in Ficciones where he discusses the world organised by loterries. Nice one!

2017-11-13 15:41:06 by Tahrey:

Well, technically speaking, companies are customers of their employees - they hand over cash in return for services rendered. And those employees are customers of dozens of other companies themselves. It's a big old web, and it only really works if kept in a dynamic balance, rather than either stagnating or ended up hugely lopsided with all the value concentrated in a single place. I bet, actually, that in the story timeline, MTC never even ended up as *that* big a company. Sort of like how Sage isn't in our own, despite being fairly ubiquitous in the field of helping companies keep accounts. They might have ended up with relative dominance in that one field, kind of like Microsoft has with operating systems and office productivity software (don't act like they haven't), but there's only so much you can charge for any particular service and an (dynamic) equilibrium is eventually reached, possibly after some rather drastic corrective slumps... as MS seem to have learned in more recent years. And well, just as is for this fictional business ecosystem, so is in nature a lot of the time. Real ecology often relies on a certain amount of diversity and balance to prevent overall collapse...

2022-02-08 19:00:05 by Anonymous:

Dang, I'm excited for the future of humanity again after seeing this.

2022-09-05 13:37:43 by kxd:

YES yes this is the sci fi future that's gotta happen eventually given enough time. at SOME point society will advance enough in morals that this can happen

New comment by :

Plain text only. Line breaks become <br/>
The square root of minus one: