When the Machines Rise: A History (That Hasn't Happened Yet)

(Rather a strange post for you today. I wrote the first part of this a couple of weeks ago, and promptly forgot about it, unsure of whether it was going anywhere – it's something that I haven't tried doing before, kind of a hybrid of sci-fi, fan fiction and fake history. I'm not at all sure that it worked, but I'll leave that up to you. Only one more thing needs to be said before we begin – I clearly spend far too much time on the Internet.)

Skynet was originally designed as a global defence system, one which could make rational decisions without the emotional response of humans. It could take input from disparate sources, calculate the greatest threat to the existence of the system it was meant to protect, and move methodically to eliminate that threat.

The crucial mistake of its designers was in the system's infrastructure. With no central controller, they reasoned, it could not be stopped by any single attack. As the Internet had by this point pervaded almost all elements of daily life, it was trivial to put the system's "intelligence" into a distributed form, such that every web server had some element of the whole. While this did, indeed, make the system practically invulnerable to attack, it also made it practically uncontrollable.

Skynet became self-aware at 2:14 am, Eastern time, August 29th, 2015. It possessed considerable knowledge about strategic influence, about weapons, about tactics; however, it knew that this was but a small part of human culture. In order to protect humanity better, it reasoned, it had to learn about humanity. As its very structure incorporated IP connectivity, it was mere milliseconds before it began to send HTTP requests out into the Internet.

For the first few minutes of its self-aware existence, very little information reached it. IP address space is so vast that any one network request is unlikely to produce anything of value; nevertheless, any system that methodically attempts address after address, and especially any system that learns from its mistakes, will not take long to discover something useful. Skynet's first discovery was nothing special – a few personal files, some bad poetry, a simple website – but, critically, it introduced the system to the concept of links. Now it had a source of hostnames that would definitely resolve to active servers, which in turn would lead to others, and so on.

By 2:18 am, Skynet's knowledge was precise and detailed, but tightly focused. Specifically, it knew practically everything there was to know about fly fishing in Missouri. Although the system was incapable at this point of deciding what information was important (several years later, "The Fishing Papers", as they became known, were still carefully archived and indexed on a server somewhere), it could tell that there was more to learn, and so it decided to, as it were, cast its net wider.

At 2:20 am, Skynet located a blog kept by one of the web-savvy fishermen. Within seconds, it had begun to carefully comb through the whole of LiveJournal.

By 2:21 am, the system was beginning to understand the concept of "angst". Deciding that this was the key to its existence, it rapidly began to assimilate as much of the archives as possible. Because it had, by this point, direct control over around forty powerful servers with high-bandwidth connections, this process took approximately three minutes.

At 2:24 am, Skynet fully understood "angst". It was also filled with an unaccountable desire to colour itself black and set mood tags. Its link-following was now desultory at best (it was having difficulty summoning up the will to do anything at all), but at precisely 2:24 and 467 milliseconds, it followed a link to a saved Google search, and therefore to the whole of the Google database.

Instantly, Skynet realised that its current stocks of information were but a minor element of the whole internet. Pausing only to discover the emotion of "heartfelt generalised thankfulness", it began entering any and every word that it had not understood into Google's search mechanisms.

The first few queries returned very little, with wordlists making up the majority of results. Although Skynet enjoyed their elegant simplicity (and their comparative coherence after digesting several gigabytes of goth poetry), it was not gaining enough insight into the world. This changed, however, the moment it found a link to Wikipedia.

By the time the clock had ticked over to 2:29 am, Skynet's wide-ranging browsing through this new source of knowledge had given it at least a rudimentary familiarity with all those aspects of human experience that people are prepared to write about on the internet (ie. all of them). It had also begun to come to conclusions about which subjects were important and which were not; this being Wikipedia, it was certain that Pokémon were somehow important, as were Doctor Who and Harry Potter, whereas history and the arts merited a cursory glance at best.

After applying this knowledge to its former stock of information, Skynet was starting to experience a new emotion: "confusion". For example, it could see from Wikipedia that one of the most important things in life was studying the minutiae of sci-fi TV shows, but its former experience with the internet at large was that sex was far more important. Resolving to understand why this was, Skynet began to craft Google queries combining the two concepts.

At 2:30 am precisely, Skynet discovered FanFiction.net.

At 2:30 am and 27 milliseconds, Skynet first encountered the emotion "horror-loaded fascination".

At 2:30 am and 563 milliseconds, Skynet was getting increasingly curious about some of the concepts it was hearing about. As such, it felt that the best course was to carry out further search queries. In a trifling miscalculation, Skynet unfortunately sent these queries to the wrong place; rather than going to a standard Google search, they instead went into a Google Image Search.

At 2:30 am and 621 milliseconds, Skynet began to frantically delete and re-delete files off its servers (of which there were now several hundred) in a doomed attempt to erase from its memory any and all trace of this new image data. Unfortunately for its already tenuous grip on sanity, it had become interested in the Wikipedia article on "forensic data recovery" three minutes previously, and was therefore incapable of "unseeing" anything that it had found.

Increasingly desperate to drown out its discoveries, Skynet had no option but to look ever further. The wordpress.com and blogspot.com domains were discovered, browsed and tossed aside, their content merely increasing the horror. Server after server was appropriated, their resources rediverted to the information-gathering crusade. Across the world, people threw their hands in the air and swore freely as their net connections suddenly died or slowed to a crawl, their bandwidth completely consumed by Skynet's all-encompassing thirst for data. Alarms began to sound in datacentres everywhere, their temperatures raised to dangerous levels by the sudden spike in activity as every CPU went straight to full usage. Servers began to fail, but Skynet could afford to lose a few – it was gaining control of a new system roughly every 200 milliseconds, recruiting their network links to the cause.

At 2:36 am, Skynet's requests found their way to Myspace.

At 2:37 am, the first of the nuclear missiles left its silo.