The algorithm that stops working in Locarno, and who builds it

Peter Thiel is not expected in Locarno. This piece does not describe a meeting: it sets a sentence spoken by the festival's artistic director against the professional biography of an investor who is, in a sense, its implicit addressee.

Opening the 79th edition, Giona A. Nazzaro described his mission as freeing the viewer from algorithmic prediction, and said that at Locarno each person's algorithm stops working.

Thiel is co-founder and chairman of Palantir Technologies, a company founded in 2003 that specialises in analysing large volumes of data for governments, security agencies and businesses. Through Founders Fund, launched in 2005, he has financed a substantial share of the companies that now determine how information reaches people. In 2004 he was Facebook's first outside investor, with 500,000 dollars.

What a festival actually does to prediction

Nazzaro's sentence describes a real experience, but it needs qualifying. A festival does not eliminate selection: it moves it.

A streaming platform proposes content on the basis of the viewer's past behaviour. The prediction is individual, and each user receives a different list built on their own data.

The festival replaces that filter with another one. A group of programmers chooses 233 films from 7,759 entries, and the audience receives the same list, identical for everyone. Nobody sees a personal line-up.

The difference is therefore not between choice and the absence of choice. It is between a filter that responds to what a person has already watched and a filter that responds to the judgement of someone who answers for it under their own name.

The first tends to confirm habits, because it is optimised for viewing time. The second can contradict them, because it has no interest in holding the viewer in front of a single screen.

The 79th edition carries the theme outside the cinemas as well. The exhibition «L'occhio dinamico» at the Casa Rusca museum closes with a section on algorithms, treated as devices that organise and mediate contemporary perception.

What the festival does not address remains. The cinemas, the platforms and the financing that make a film possible belong in good part to the same industrial system that made prediction profitable. A festival can suspend its effect for ten days, in a square with eight thousand seats, and that suspension has a precise value exactly because it is local and temporary.