Issue 1/2018 - Asoziale Medien?

The Match

Lost and Found at the Computer Trade Fair

Thomas Raab

“There is only one kind of intelligence but there are 7.5 billion kinds of stupidity – true artificial intelligence would have to simulate both, otherwise it would be stupid.”
(Tatjana Raskolnikova)

After treating ourselves to two cans of beer at the petrol station and laughing at the term “pre-loading” with two after-work drinkers, we finally set off towards the computer fair. The brightly lit trade fair building drew us towards it as if by magic; we could hardly take our eyes off it. Jenny was of course already more of a habituée and was not so fascinated by the matt silver Fringebooks, the glistening towers of servers, cooling equipment and xPhone posters that we could gradually see more clearly through the squeaky clean panes along the trade fair building. Jenny worked for sie. She tweaked algorithms to improve a partner-matching app's machine-learning paradigm. Her firm was called Colorit.
Even beforehand, she was bored by the thought of the obligatory trip to the annual trade fair, which is why she had hooked herself an artist friend a week beforehand at a Devendra concert, so that this time someone would keep her company in that hell. It was indeed a preview of hell that awaited all of us, artists included. A perfect, cosy hell, with long drinks, beer, cigarettes and drugs on tap. With fun, cybersex and an after-show disco. And with excessively bright lighting, over-polished displays and too much hetero-SM. I was her artist friend.
The quality of Jenny’s new algorithm was measured on people. The benchmark was a match-rate of over 90 per cent measured in terms of real life (which however, in my view, had gone downhill anyway). Jenny had been one of the leading engineers who switched the process from searching for correlations between data-sets (people) and multi-variable methods like factor analysis to an experimental machine-learning system. She had been brought from California to Berlin specially for that.
Anyway, Jenny was good. Somewhat tipsy, we swayed between the brochure stands, glossy display cases and sales desks laden with gadgets, and headed for the buffet, spotting newspaper stands that had been set up in front of it. There was not a single staff member to be seen far and wide, nor anyone handing out the papers. We stopped to take a look. Interestingly, among all the Network Raids and Computer Grafix I also spotted a pile of the art magazine springerin, and grabbed one before we moved on to the high bar tables around the buffet.
Jenny had ordered a Dos Equis, and I indulged in a Cuba Libre as she frantically scanned the undulating horizon of crème-de-la-crème heads, trying to spot any of her bosses, who were all from the financial sector, of course. I actually wasn’t a big fan of long drinks, they were too expensive and I couldn’t stand Coke even without rum. But here, among all those programmers, middle managers, tech-groupies, computer scientists and USB-3.0 cable scroungers, I had apparently been overcome by the need to look like someone special – although, no, because I was pretty short of cash right then. But maybe Jenny would ...
Typical artist, she perhaps thought, as she spotted my ridiculous long drink. The barkeeper in his silly catering uniform was definitely also an artist, for now a song by Adam Green was blaring out across the trade fair buffet, which had been made to look as “cosy” as possible despite the glaring lights (garlands, vintage armchairs, ashtray smoking stands). I stumbled in all wide-eyed to the computer show, and followed you into every store, please tell me what I’m for. Listening to Green's New York sounds, it occurred to me that Jenny had had a doctoral grant at MIT a few years earlier. She must be a genius! Certainly as far as computers and bosses went. But here and now, she needed me, she needed creativity, which is why she had picked me up.
When the usual muzak started playing again, she explained to me that the experimental aspect of her algorithm was that the matching wasn’t based on any positive correlations. She took a swig of Dos Equis from the bottle. Because it was all about machine learning, she continued, it was also impossible to offset negative correlations and positive matches. To be honest, even as an artist, that wouldn’t have seemed particularly experimental to me.
Jenny suddenly gave a start. She stared at a silverback just about to head for the buffet, with a middle-aged woman in a grey suit glued to his side, a telltale smile on her face. “The big investor”, Jenny whispered, turned away from the two of them and faking an interest in the art magazine she had picked up and begun leafing through.
Out of the corner of my eye I could see that Jenny had stumbled on an article entitled “What the Heart Desires”. As the tagline revealed, it was about the last refuges of the human realm that still withstood calculations by AI and Big Data. That naturally interested Jenny, because her experimental algorithm was intended to compute falling in love, still considered the headliner of romantic unpredictability.
En route to the buffet, the silverback was quickly intercepted by a friend and, with his besuited companion, drawn into one of the “lanes” that held a series of smaller company stands. I guessed that the highly risky innovation was being sold there.
Jenny breathed a sigh of relief – but whichever way you looked at it, she was soon going to have to make her product presentation – with me discreetly in the background.
What was the innovation she had come up with?
She had been one of the first to recognize that the old matching programmes had found similarities, and had therefore linked people narcissistically. With all the experience of a 35-year-old, she had realised that this was not the way to calculate love. As she had explained to me right after the Devendra concert, it was also not about matching opposites, in other words, contradictory characteristics and external traits like education, wealth, appearance and style. A mixture of positive and negative correlations was also not an option.
“So what do people fall in love with then?”, I had asked after Devendra.
Jenny put the copy of springerin down on the table and downed her Dos Equis in one.
“What is love?”, I had insisted, my ears still ringing. Even just talking about that subject had made me start to like her. A fascinating person, I had thought to myself, and suddenly her eyes, which I had found pretty boring before, also seemed to shine.
“I don’t know that, do I?”, Jenny had cried enthusiastically.
I’d been astonished. What a person!
Meanwhile Jenny had grabbed my hand. Hers felt warm, fresh and good. Fascinated, I continued listening to her as my tinnitus gradually faded – standing in the foyer after Devendra’s concert, which we hadn’t really liked. And for a few seconds I had forgotten my (Performance) art.
“What do you mean?” I had asked, confused. “You don’t know what your algorithm calculates?”.
“That’s right”, she answered gently, as if she were talking to a little deer. Why was she being so gentle? Was she calculating me? Suddenly the secret of Jenny, with everything that entailed, loomed large before me. All of a sudden I wanted to know who her parents were, what she liked to eat, why she had become a computer scientist, what her favourite YouTube channels were, if she liked Santana, even her shoe size ...
“We use experience-based values! We take successful matches from the past and use them to train the neural network with deep learning and support vector machines.”
Left! Left!
I suddenly felt hot and had pulled my hand away. What did she mean? The unpredictability of love as a romantic postulate in a neural network?
Deep in thought, I just had time to shove my empty glass onto the bar table next to her bottle, and a moment later I was scuttling along after her through the crowd of geeks, gamers, tech-junkies and investors. What a woman! Of course, I was also fascinated by this alien environment, for up until that point I had only ever been with artists, and for that reason alone - at least that was my helpless guess after the Devendra concert - Jenny glittered for me like a rare blue gem.
The Colorit stand stood slapbang in the middle of the central trade fair plaza, brightly top-lit by huge halogen lamps. Squinting upwards, I could see the stars coming out in the dark-blue sky beyond them. A large crowd had gathered. Colorit, with its hip bright-blue CI, had the reputation of being a cool employer. Even so, I hadn’t expected such a crowd. Just walking towards it made me feel a little nauseous. Was our coup going to work?
When Jenny reached the throng just before me, she stopped so abruptly that I bumped into her back. She turned around briefly, smiled and winked at me. My heart leapt. Was I in love? Nonetheless I thought her trusting look had more to do with our plan than with any affectionate feelings.
During the day we had entered our profiles in the Colorit form, which had taken a few hours (which I could however write off with an entirely clear conscience as Performance art preparation), and then tweaked them until Jenny was certain there would be a 99 per cent chance of a match between us coming up live as well. I had my doubts, but after all, she knew her algorithm best ...
She had already climbed up onto the small stage. She looked stunning in her bright-blue catsuit – objectively, I mean, for all eyes, not just mine, were glued to her as she took the microphone and began to speak. My heart was beating like mad, the Coke, complete with the rum aroma, was repeating on me, and I closed my eyes for a moment. When I had pulled myself together again, I saw Jenny talking into a headset while swiping like mad at the Colorit app on her xPhone, which was transmitted onto the large screen behind her.
Non-matches to the left, matches to the right. The bad ones go into your crop, the good ones go into the pot. The algorithm would then filter out from the latter category the right partner “for life” – me! All I would have to do at that point would be to yell in triumph, go up on the stage and fling my arms around Jenny. That was exactly what I wanted to do anyway! Ever since our evening at Devendra, after which I had lain sleepless with longing night after night.
Left! Left! Left! Right! Left! Left! ...
Aha! Jenny swiped away all the photos with dogs, cats, horse or flowers, apparently without paying the slightest attention to the face depicted.
Left! Left! Left! Left! Left! Left! Left! Left! Left! Left! Left! Right! Left! Left! Left! Left! Left! ...
How convincing she seemed! The bad ones into the pot! The bad ones into the pot!
As a study has shown that as almost all men on matching apps tend to swipe right, and into the pot, for just about everything, as a woman you really have to watch out who you swipe right, because it means that every man you match with comes to the first date with such enormous expectations that it would be almost impossible to refuse sex without making them really angry, which would filter through into dislikes on Colorit and then, via its parent firms, Facebook and Google, would affect your credit rating and damage your career chances. The women knew that the men knew that, and they in turn knew that the women knew it etc. I hated matching apps, and I hated the hetero cliché. But it wasn’t any different on the LGBTQ platforms. You could spot, although you weren’t allowed to say so, who was attractive to most of the target group and who wasn’t, and therefore had the same problems as everyone else. We were all only human, after all.
Jenny swiped so elegantly and nonchalantly that some of the nerds just stood there gaping. I was hot all over at the thought that she could be my new girlfriend.
Meanwhile she confidently explained a few details of her programme, and the nerds nodded enthusiastically. Despite my romantic daze, intensified by the enthusiasm of all those “dudes”, terms like support vector machine, kernel function, separable data, tensor flow and Bayes network seeped through to me, but I couldn’t get a clear sense of what it was all about, the picture of the algorithm fused with Jenny’s picture, I couldn’t hear a word any longer, my tinnitus suddenly flared up, crescendo, Jenny’s face grew even brighter in the already glaring ceiling lights of the entirely halogen-lit trade fair hall ...
There! Jenny’s algorithm had calculated something, which, as is standard, was announced by a flashing green display, broadcast on a huge screen above our heads. Green! Everyone held their breath. Green! In a few moments the light would stop flashing and my face would appear!
There was the photo!
But it was Jenny’s photo!
The crowd cheered, because they naturally thought that Colorit, reflecting its company profile and Corporate Identity, was scoring points by being ironic. Gratification.
Jenny was distraught, as I could see from her twitching right eyelid. She had failed. What would the bosses in the back row think, what would the investors think?
It only dawned on me gradually: We had failed! I couldn’t think clearly any longer and fished my xPhone out of my pocket to repress the embarrassment and welling pain by fiddling with any old app. Performatively, so to speak.
I kept swiping until I matched with myself too ...


Translated by Helen Ferguson