Faisal Zaidi

Why we cant stop scrolling. Two things people say about the feed in the same breath: it is a slot machine, and it knows you. Randomness and prediction are opposite engineering goals. Both claims cannot be true as stated. A slot machine holds no model of the person pulling the handle, and its payout ratio is fixed before anyone sits down. A ranking model does nothing but hold a model of you, and every hour of work makes the next item less surprising. The resolution is that the second manufactures the first. A recommender is not trying to be predictable to you. It is trying to be accurate to itself. Accurate means better than chance and nowhere near certainty. At your end that is a stream where most items are mildly worth continuing and an unpredictable minority are genuinely rewarding. The ratio is set by the accuracy of the model rather than by a randomiser. From the seat of the person holding the phone, it is a variable payout arriving after an unpredictable number of swipes. Which is what Ferster and Skinner described in 1957, from the other direction. The variable-ratio schedule produced the highest and steadiest response rates of anything they ran, and it was the hardest to extinguish. Stop paying on a fixed schedule and the subject notices at once. Stop paying on a variable one and it keeps going, because an absent reward is indistinguishable from an ordinary gap. A casino designs that schedule and tunes it by hand. A recommender falls into one by being imperfect, and it falls into a schedule tuned to the individual rather than to the average customer. No casino can do that. The strongest opposing case: intent is not what does the damage, and the structural argument works without it. Separating the two accounts would need someone to degrade a live recommender for randomised real users. The people who could, will not. And they would say the analogy was never about the dice. It was about the schedule. Where the schedule came from is a question for historians. It changes nothing at 1am. #psychology #attention #technology #habits #socialmedia thewhyweproject.com/articles…

Do you agree?

Virgil Abloh liked to argue with a cheap candle in a tin. In an all-white gallery it reads as art. In a garage it reads as rubbish. The wax never changes. His actual point was an allocation question. He could spend his time designing the candle, or spend it designing the room the candle sits in. Sounds like a designer being quotable. It is also testable. 2008. Hilke Plassmann, Antonio Rangel and colleagues scanned people while giving them wine and telling them the price. Some wines were poured twice, once labelled cheap, once labelled expensive. Believing it was the ninety-dollar bottle made people report more pleasure, and made the brain region that registers pleasantness respond more strongly. The price did not just change the review. It changed the experience being reviewed. 2001, Bordeaux. Gil Morrot and colleagues dyed a white wine red with an odourless colouring. Fifty-four oenology students then described it in red-wine vocabulary. One study, fifty-four students, so hold it loosely. Artists worked this out first. Duchamp signed a urinal in 1917, called it Fountain, put it on a plinth. Brian O'Doherty later called the gallery a machine for conferring status, one that hides the fact that it is a machine. The trap for anyone who makes things: effort goes into the object, a better formula, one more feature, a nicer finish, and the room gets whatever is left. Abloh came out of architecture and reversed that ratio deliberately. This is not an argument that context outranks quality. Someone still had to choose the plinth, the title, the timing, and that choosing was the craft. The room was never optional. Every object you have ever judged arrived inside one. thewhyweproject.com/look/con… #design #psychology