Humankind is facing a new era. For the first time, our intelligence, best reflected in our ability to produce relevant, intelligible, informed and to-the-point blurbs of consistent language, is no longer our prerogative. We created language-capable Artifical Intelligence models that leave us but two things to make us truly human: partying hard and the economic consumption of deluxe items. By playing, buying, and amassing board games, I actively engage into the second and feel more integrally humane as a result. Still, I believe we are not entirely outsmarted yet. At least, not if we put some effort into it.
The reason why ChatGPT and the like are working, is because we rely on prefab material – both to think, and to express ourselves. Thought is not created anew every single time. It’s an eclectic arrangement of scraps of ideas, notions, and reasonings that we had before and that we imperfectly stored in the activitation patterns of our brain. And so, when prompted to think, we rely on recurring schemes – same with language. Even though the power of combinatorics makes every sentence look rather unique, it is but a mish-mash of repeated syntactic and lexical patterns. These are the patterns that the algorithms learn, giving them the key to a human-like expression – especially given that learning algorithms have been created, raised, and trained, to pick up patterns. They even pick up patterns when there are none, just as we do with star constellations or cloud shapes.
What gives us an edge, at least for now, is that we are more creative, more subtle, more nuanced; well, with a caveat. Only if we strive to be so. And this is where BGG comes into play. Because on BGG, you often find threads from people asking for recommendations. They could have asked ChatGPT, but somehow they have resorted to an old tribal reflex and asked for the clan’s opinion. And it may look like a trivial demand, something that should not be worthy of any mental effort expenditure. But at stakes is our very capability to be more enlightened, more sharp, more accurate than a machine fed with a gazillion of patterns, designed to regurgitate the recurrences of our collective mind in a fancy dressing of human-like language.
So let’s not produce these same tired recurrences ourselves. Too often, I have seen the same stuff coming up and again. For instance, someone is asking for a proper alternative to Friday, with a slightly lighter theme. One of the answer is: Sprawlopolis. I mean, why? These games are nothing alike. It only happens that, both being light, card-driven, solo-focused games, they end up in the same bag of online recommendations. Another poster is asking for a good solo card-driven dungeon crawler game. Someone comes up with Palm Island. And gets massively approved. Here again, why? Because it’s a fairly popular game that fits in the loose category the request is a tiny part of. It’s no surprise if, when asked for a good solo game, ChatGPT can answer stuff like Catan or Carcassonne. Both are popular opinions that belong to the larger category solo games are a part of. The algorithm is not to blame: we are. Because we rely on ready-made patterns, we give answers that please the crowd, that are familiar to it, not answers that are relevant, and there is no social reward for the latter.
A rather puzzlingly erring answer to a 1PG recommendation request.
What would be a proper answer to these recommendations then? The truth is, a proper answer is not easy to come by – which is the whole crux of my argument. To be able to name a game that plays similarly enough to Friday, you first need to have played many games in the larger category Friday belongs to, in order to understand what distinguishes Friday as a game. It has cards. It has deck-building, but not like any other deck-building. It relies on card combos, but not in the way a dueling game would. So you need to know all this adjacency space to discriminate what makes Friday special. Then you need to know a large number of lesser known board games to assess whether one shares the features that you identified.
It takes mental work. More crucially, it requires you to possess a board game culture. And a culture is something difficult to build up, all the more so in our modern societies, where cultural contents are so abundant, so diverse, and where societal pressures demand us to be cultivated in a large number of domains. To face the demand, we replace cultures by patterns that give a semblance of it. And we end up with collectively approved facepalm island kind of answers. Leaving us no better than a goofy algorithm overfed with a mass of content that is worth no better than our own repeated failings in forming better judgments.
Note that a frequent alternative to the facepalm island type of recommendations is to spit out a few obscure games that almost no one has heard about. This, however, is about as equally misleading; the poster then verses into the pregnant belief in “hidden gems”, looks for far-fetched copies on the second hand market, plays a few times, realizes the game is probably poorly balanced, has messy rules, or just does not provide much interest past the initial discovery, and then is stuck with a copy of a game virtually no one wants or is looking for. The recommendation itself comes from someone who, likely, has played it once, thought it was nice to play a secret indie title, fancied the title or the art box, and never played it much more afterward, but retained a satisfied memory about it nonetheless, sufficiently fleeting and incomplete to maintain the illusion that the game is worth owning even though it will always be picked up last after a wealth of more successful, better known titles.
And if you ask for lesser known solo titles, ChatGPT is similarly apt to scrap up the dust in the end shelf and find a list of indie unknowns. The key thing is: when one recommends a game in response to a request, coming up with an explanation may be the proper way to do so – why this game fits the recommendation request, which specific features of the game matches the poster alleged expectations, and so on. Of course it takes more work. But otherwise, the reply may be no better than spitting out deluded advertisements.
Obviously, it's not like it's important. But again: the patterns we engage in to deal with things of no importance, are the same patterns that may dominate our thinking, our behavior, our decision, when it comes to stuff that really matters. Which is why, sometimes, it's worth fighting over the most insignificant oddities of the board gaming world.
All illustrations have been taken from the BGG database.