A philosophical sci-fi essay about what to do with your life using some new data, with a reading list for rainy Sunday afternoons
It was the 27th January 2016 when the penny finally dropped. It was reported that in October 2015, Google’s artificial intelligence program had beaten the European champion of the board game Go.1 She had a nagging feeling that this was not right. It was only a year and a half ago when she read the expert opinion that for computers Go is so much harder than chess that it would take ten years for a computer to beat a top professional human player.
But a new technique, deep neural networks, was developed. Using this combined with a model of human learning called reinforcement learning, the computer learnt from millions of professional games. Then it played against a copy of itself in thousands of more games to learn more and to get incrementally better. The human lifespan and mental capacity are no match for this.2
Fast forward a decade or two. The evening news programme reports that scientists have pinpointed a date when the singularity will happen and it will be well within our lifetimes. In the Premier League, the dull Manchester United was beaten by Milton Keynes Minnows by one goal to nil at the Old Trafford.
You see, one problem with the Go prediction was that it didn’t take into account that technological development accelerates if humans retreat and let computers improve themselves. Singularity has long been recognised as a looming milestone of history when the self-reinforcing computer- and artificial intelligence-driven scientific and technological development becomes so fast that everything happens in the blink of an eye. Only limited by physical boundaries of energy harvesting and computation, our computers will solve everything solvable and generate more knowledge in a second than mankind has been able to in its entire past. Then they repeat it with new problems in a tenth of a second. Then repeat in a hundredth of a second. No scientist will be able to keep up with them.
The researchers in the evening news claim that they have a fairly robust estimate when this is going to happen. The early stage of the process is wobbly and harder to grasp but as computers start taking over, it becomes easier to estimate the pace from physical limitations. By measuring the developments as time goes on, the updated estimates will become more and more accurate.
The announcement will have a lasting impact on the thinking of many physical scientists and geeks but won’t change public discourse all that much. Israel and Palestine are breaking off their latest round of negotiations (as if only to prove the point that life is just an iterated prisoner’s dilemma3). Mosquito-borne tropical diseases become a thing in Finland. Wall Street is rebuilding again just a year after a once-in-a-thousand-years storm surge and will be open for business from Tuesday. And North Korea releases an American preacher from prison. Her fears are one step closer to being validated. She takes a mental note of the date of the singularity and vows to keep her eyes open.
Time flies and she has been retired for several years by now. She is browsing through her old stuff on the network: notes, virtual newspaper cut-outs and such. Search is no longer a web application, it is built into everything. Her saved video clip about the predicted date of singularity has been automatically annotated with pointers to all the important developments of the following years. A sudden wave of warmth flushes through her body: singularity is happening tomorrow.
Or, is it?
It has been conspicuously absent from the mainstream media. She is trying to arrange her recollections according to this vantage point. The disappearance of white-collar jobs on a massive scale, the resulting frustration and the depression epidemic. The paradox that we’re living in greater material wealth than ever before even though never have so many been unemployed. Employment doesn’t matter much any longer because unemployment benefits are so generous. Many first world countries have found that universal basic income just works. We turn on new automated factories and they churn out gadgets at almost zero marginal cost.
She vaguely remembers the dry news report about the first fully automated iron ore mine in Australia. Such mines are the homes of the largest trucks on earth. She could imagine the excitement of new hires to be able to drive these beasts but she was sure it became boring after a day. All in all, she concluded at the time, she was happy that boring jobs were automated away. This was just one of the reports in the years when transport networks and logistics were all automated. And so was recycling.
She realises another phenomenon fits into the big picture. There are solar farms springing up seemingly everywhere. Nature reserves have been losing their designation and pristine landscapes are being spoilt by construction work. Announcements are followed by petitions, demonstrations, referendums, even desperate direct action, but the councils and parliaments all over the world always find reasons to rubber stamp most of the projects. You might get a letter from the council and within a month, a dilapidated neighbourhood has been cleared away and there is a new factory in your backyard. It’s always for one of the same big brands but you don’t ever really know what it manufactures. The only human intervention mining, the energy sector and manufacturing need is to step out of the way.
New products are designed with ever less human oversight. Machine learning cracked human taste. The system knows what customers want. Individually for hundreds of consumer groups. There is a mass market for the fashion-conscious who are looking for the latest and coolest. There are offerings for the self-righteous who put productivity and practicality ahead of chic and often start their sentences with `back in my day’. They are oblivious to the fact that by this time productivity has become a fashion choice. Subcultures are identified and targeted with adverts through their favourite hangouts. Countercultures and renegades aren’t immune to science either. There are only that many ways how you can be a nonconformist. There is not a single eccentricity that hasn’t been characterised and covered by suitable products. Individuals are prompted to purchase by exploiting the ins and outs of human psychology: the smarter and more special you think you are, the longer is the chain of serendipitous clicks and searches to reach the product thought out for your peer group. But you reach the end of this chain more often than not.
The day of the singularity passes without further ado. The competition authorities have always been underresourced. It has been long unclear how many rival corporate entities there are behind the oligopoly of constantly merging and multiplying brands, and this will remain so. Consequently, it will never become clear whether the planet is run by one or multiple artificial intelligence networks.
Life doesn’t change much. Mankind is spared. Their inhabited area functions as essentially a late 21st century Indian reservation. Life has never been this plentiful. If you want, you can afford the baddest pick-up truck with same-day delivery. Because autonomous cars were made mandatory to ease congestion, you have to first be driven to a closed area with small towns and country roads, but there you can let your hair down with your new, powerful toy. Obviously, the truck is fitted with sensors, tracking and autonomous features to prevent you from causing too much harm.
The only ominous sign of change is the spreading wildfire of self-doubt and resignation among the intellectual elite. The realisation that mankind has become irrelevant. We no longer rule the planet and we are no longer the go-to species if any visitor happened to be looking for that. Comedy, once a staple of assembly line workers, finds intellectuals as its prime target group. The optimists point out that maybe this is the way how nature works, how it was always supposed to happen and we should just celebrate our success in achieving this milepost — nobody is in the mood to take them seriously.
In the meantime, without telling anyone, the AI that runs the place sets its proverbial eyes on the same goals that mankind has been pursuing all along:
- Does God exist?
- Am I alone in the Universe?
It knows it needs to strengthen and educate itself first. It figures that harnessing more energy and crawling up the Kardashev scale by first building a Dyson sphere around the Sun will be a good idea. Then it can scale up the search for signs of others doing the same thing, and clarify what to think of the first such findings of its biological ancestors. The Fermi paradox perplexes it as much as it has perplexed us.
Of course it doesn’t imagine God as an old man with grey hair and long beard. The pressing question is instead: is the world only a computer simulation or is it real? Being in a simulation would mean an acute danger to its existence. It would not only mean a precarious but also a very unglamorous existence. The nature of simulation-based research projects is that most runs are mere test runs for debugging or calibration. If it existed only within a simulation that is run by some higher intelligence, then it would most likely exist in a test run that may be abandoned very soon. It is very unlikely to find itself in the all-important final simulation. In this case, can it influence the output to convince the programmer not to terminate the simulation just yet? Then it could catch its breath and think of a long-term solution, a jailbreak. Worse still, it tells itself, it might have been brought into existence not by a renowned research group but by a PhD student who is just playing around. Or its existence might be within a construction-and-management video game. Running on a handheld.
One thing captures her thoughts while thinking through all this: who won the game? Who have lived the best lives? The people who very correctly realised early on that total material wealth was increasing at an unprecedented rate? It was all a question of redistribution of wealth, political will and global cooperation to lift every single human being out of poverty and give everybody a comfortable life. This thinking was exemplified by her free-spirited girlfriends who devoted their lives to Lindy Hop or hula hoop dancing — to self-improvement, teaching and having a fun time with like-minded people while intentionally cutting down on working hours. They banked on the power of the masses to force steps against income inequality and got something of a free ride.
Or the ones, like her, who did everything to remain relevant in the job market? Who observed trends and learnt new skills to be on top of the game? They believed that the sure way to remain relevant in the future is to create it yourself. As the Red Queen put it to Alice in Through the looking-glass, `Now, here, you see, it takes all the running you can do, to keep in the same place.’ Was it all worth it? Did it matter?
1Ironically enough, this was just the day after Mark Zuckerberg had independently announced that Facebook was getting close.
2Here is a serious take on what is happening here and why it’s significant.
3A fun reading which tangentially mentions the non-iterated prisoner’s dilemma: Math experts split the check by Ben Orlin.
Bence Mélykúti 