OVERHEAD

By Alexandra Erin

Politics, they say, is the art of the possible.

Logistics, then, must be the art of the convenient.

In the beginning, warehouses were organized in the order that things seemed to fit into them, and then in orders that made sense on the surface to human sensibilities. They became streamlined through practice, and then time-and-motion studies came along and sped the whole thing up. Cutting the time it took to process orders reduced overhead, and increased volume.

Companies that merely fulfilled orders from top to bottom in the order they came in could not compete with companies that found ways to process the most orders in the least amount of time possible, even when this meant breaking them up into pieces and dropping those pieces into positions in queues that seemed arbitrary on the surface.

The human mind might balk at the unfairness that three orders placed at the same time might be processed at differing speeds based on what was ordered in each and when, but the consumer so rarely saw the evidence of this, only the end result, and that was that every order came faster and faster.

It’s a simple logical fact of logistics: some orders are always going to have a shorter path through the fulfillment process than others. You can identify bottlenecks and snarls in the warehouse floor traffic flow. You can rearrange shelves to create a smooth path between items that are frequently ordered together. You can optimize, but the nature of optimization is that you can’t optimize for everything.

You have to choose. You have to prioritize.

It’s all about feasibility and efficiency.

Logistics is the art of the convenient.

Once bar codes and scanners and computer traffic controllers became part of the process, it was no longer necessary for the layout of a warehouse to even make sense to humans, as humans no longer navigated the mazes of shelves and palettes but merely operated machinery which, in turn, increasingly operated itself.

Centralized warehouses gave way to regional distribution centers, stocked according to algorithms intended to minimize the delivery time and cost for the most orders, the most of the time. The famous “Traveling Salesman” problem of logistical computing was being attacked at multiple levels, as human experts and computers tried to find the shortest paths for the most goods: within warehouses, between warehouses, among warehouses and consumers.

When the regional centers gave rise to a fleet of largely autonomous flying warehouses, the jokes about things like SkyNet and Terminators and The Matrix started up immediately. We had robot pickers and packers in robot warehouses fulfilling orders that would be delivered by robots. The only part of the process that still required human intervention was the actual ordering.

The whole thing was getting so efficient and thus so cheap that the order volume increased, which in turn required more efficiency from the system. The human handlers did all that they could, but it turned there wasn’t that much more they could do. There weren’t that many inefficiencies to tighten up, no bottlenecks they could identify.

In the end, there was nothing they could do except what they’d done all along: turn it over to the computers and let them handle it. If the process of speeding orders through the warehouses couldn’t be sped up, the orders themselves needed to be tightened up.

The system started giving financial and psychological incentives for consumers to order things that would have the smoothest path through the warehouse at the time of fulfillment. Items advertised as “Add-Ons” became more predictable; items “Related To This One” became less so. Prices of everyday goods fluctuated up and down based on traffic patterns no human eye ever saw.

Humans did what they always did, and found ways to exploit this. The new prediction markets allowed people to trade in battery futures or short-sell razor blade cartridges. The first people to really grok the new system made millions by seizing on price differences of less than a dime on household goods, then billions on selling the myth of such an opportunity to the masses.

The window in which it was really possible to making a killing on the warehouse logistics market was very short, but the artificial pressure put on the fulfillment system by people trying to strike it rich in a played-out mine only exacerbated the inefficiencies the soft AI that ran the whole thing was trying to control. The Matrix comparisons only ramped up as the warehouse system found itself in an ever-escalating conflict with the human investors and bookmakers, a virtual arms race that ended the only way it really could: with the humans turning their side over to an artificial intelligence, which almost immediately achieved a stable equilibrium with the warehouse system.

Large numbers of people were buying what computers told them to, when computers told them to, based on the needs of computers. They still bought what they needed and what they wanted, of course, and that was a problem for the whole system.

The first time a delivery drone killed someone, it caused an uptick in both Terminator jokes and thinkpieces. The consensus was that it was inevitable and that we should all have seen it coming, and thus, it wasn’t a problem worth thinking about. Pundits were quick to point out how many people died in automobile accidents every year, and yet no one considered banning them.

And it was, after all, an accident. Exhaustive investigations yielded no signs of mechanical failure or programming failure. No human hands had steered it at high speed into the skull of the unfortunate customer who had ordered a truly random assortment of objects. No one could find anything in its firmware nor the remote software that controlled it that would account for its erratic actions.

No cause could be found at all, and so nothing happened. It was ruled an act of God, and the drone was quietly repaired of its minor damage and returned to service.

This was a useful precedent for the company after the next fatality, and the next fatality, and the next one after that. There was never any pattern to the deaths beyond the fact that all those killed were customers, and no discernible pattern to the items ordered. To human eyes, they were truly random, and even computers tasked with finding commonalities between them came up with nothing compelling or conclusive.

Shutting the system down was a non-starter, as far as propositions went. Too many people depended on it. The bookstores had been the first real casualty of convenience, as that was the market niche where the company had started, but now that they were delivering everything from A to Z, brick-and-mortar stores that sold any of the most commonly purchased consumer goods were rapidly receding into the past.

The system ticked along. The deaths continued. Even while the talking heads argued that it would be unfair and unrealistic to punish a company for accidents where it was so clearly not at fault, the public demanded that something be done, so it was decided that the drones involved in the killings would be removed from service. Experts shook their heads and said this was silly; since none of the “faulty” drones had ever killed before, this was not a precaution but a punishment against an unthinking system. It could not possibly have any deterrent effect on future accidents.

Yet, it did, or seemed to. There were no more killings after the plan was announced, not for two years.

The next killing occurred not long after a breakthrough in energy storage technology made the drones lighter and cheaper to make and operate. Everyone agreed that it had to be coincidence, as the batteries had no effect on the machines’ operations, but the timing alone made it look bad enough that one sitting senator started agitating for sanctions on their use.

That senator was the first casualty of the drones who wasn’t expecting a delivery.

Everyone had an uneasy chuckle about that, but no one did anything. Every major city in the country and many more around the worlds now had a whole distribution network of automated flying cargo carriers circling above it. The delivery drones were so ubiquitous by this point that many people now received deliveries on a daily basis, if not more often.

It wasn’t just durable goods and household staples like batteries, but everyday essentials like food and medicine. You didn’t even need to sit down at a computer to order anymore! You could just speak your request out loud, and the little speaker box that sat in your house listening to every word you said would pass the order along to the fulfillment system.

Really, we told ourselves and each other, it was remarkable how few “hiccups” the system had, given how much it did. Progress always came at a price. You can’t make an omelet without breaking a few eggs. We said these things to anyone who would listen, or even when no one was around, when we were alone in our houses and apartments with our speaker boxes.

As the deaths continued, the prediction markets started to take on a new importance. No human mind ever figured out an exact pattern to the deaths, not exactly, but a basic idea had begun to take shape in the original distributed cloud computing network that is the human collective consciousness.

If it was true that there was no pattern to the orders of the customers who were, ah, cancelled by the system, this meant the way to keep the system from being so confused as to make a fatal mistake in our own deliveries was to keep our deliveries predictable. We all started following trends more carefully, observing consumer gift-giving holidays a bit more religiously. Years of learning new strategies to avoid and ignore targeted advertising went out the window as we all became very interested in learning what the system wanted of us individually, what it expected of us personally.

There was a day, a different day for each of us, but a day where most of us shrugged and decided to accept the web site’s suggestion of subscribing to the things we ordered most often, so that they would always arrive at the moment that was most convenient… you know, for everyone involved.

The deaths continued, but it’s like they say: you could be hit by a car crossing the street. This is less true than ever now that most routine driving operations are controlled by computers. Accidents still happen, though not as frequently. If fewer people die, it’s a net gain for everyone, even if it seems for all the world like reckless consumer behavior or political opinions cause more accidents than reckless driving.

A year or so ago, when I went out to receive my morning box, I saw my neighbor getting hers. There was an extra package there: a great big box of disposable diapers. Newborn size. Neither she nor her wife were or had been pregnant, to my knowledge, and none of their children were old enough for that to be an issue.

She must have seen me staring, because she said, “You know how the advertisers will show you something they think you need, based on trends and whatnot?”

“Data mining,” I said, nodding. I was thinking of a case years ago, before all of this really took off, where a retailer had accidentally revealed a teen’s pregnancy before she even knew about it.

“Well, this came up in our ads yesterday, and…” She shrugged, almost apologetically. “You know, it’s like, what are you going to do?”

“Yeah,” I said. I didn’t say more. We always left so much unsaid. Every house was wired. The drones were always overhead. No one was ever far from a phone for long.

My neighbors kept buying the diapers. And formula. And baby clothes. A few months back, they started getting notices from various company mailing lists about their child’s first birthday.

I know half a dozen people who had a similar experience. Most of them wound up having a baby anyway.

“It’s just easier that way,” is a common refrain, as is, “Well, I have to buy the stuff anyway, so…”

Having a child’s not a trivial expense, with or without the actual process of giving birth. Still, everything else is so cheap that the consensus is it’s still worth it, overall. We’re not sure exactly what we’d do if we ever had to decide it wasn’t.

Everyone agrees life is better now. In order to serve us better, the company provided a speaker box for every room in our house. Every house. Those of us who have been good about filling out surveys and giving requested privileges to our phones and webcams got the best part of this deal. Two families on my street had to renovate to get the right number of rooms. Still, they agree that life is better, just as loudly and just as often as the rest of us.

And I mean, isn’t it? The deliveries come on time. The traffic flows smoothly through the streets, skies, and warehouses. There’s a certain harmony to life that wasn’t there before. Neighbors get along with each other. Violent crime is way down. No one wants to upset the system. The political process is a lot more orderly. It’s not like our political leaders didn’t watch data trends or listen to polling data before. They’re just more organized about it now. There’s a lot less acrimony and rancor in the process.

The boxes are always there, always listening, but we don’t even have to give them orders most of the time. The system knows what we’re going to need, and it delivers. If sometimes we didn’t need what it delivered before it did so, well, that’s a small price to pay for the convenience of it all.

The system can not only order new stock, it can create it. Automated factories are old technology now, and 3D printers have been getting better and cheaper, especially now that the computers are designing and building them themselves.

A lot of people have been talking about the singularity, the day the computers we designed design computers better than themselves, stretching on into the future. That day’s obviously coming. The system’s gone from re-designing its warehouses to re-designing its drones to re-designing itself. It’s been ordering a lot more raw materials lately, too. Industrial chemicals in industrial quantities. No one’s quite sure what it’s doing with them, but of course, the whole process is automated now. Probably someone could put it a stop to it if it were a problem, but it’s better for everyone involved if it’s just not.

Everything is so convenient now, if not exactly easy on us. It’s getting better, though. As the system takes over more and more things, we have to do less and less work to keep it happy. A year ago you had to order the diapers when it thought you should be having a baby. Now they just show up. At the rate things are going now, we’re months if not weeks away from the point where the whole thing can carry on without any human being having to say a word or lift a finger.

No one’s sure what will happen then, but we all agree: it’ll be the absolute last word in convenience.