Feedback Systems are Counterrevolutionary

Essays on technology, psycho­analysis, philosophy, design, ideology & Slavoj Žižek


June 8, 2011

Feedback Systems are Counterrevolutionary

Larry Sanger, co-founder of Wikipedia and Citizendium, wrote a blog post that levels the charge of anti-intellectualism against the geek community. I think this is manifestly true, and one of the most obvious indicators is the proliferation of services based on a cybernetic model of feedback that they build and promote.

Why do these indicate anti-intellectualism? An example of one of these systems is Amazon’s customer reviews. In the pre-internet world, part of the role of the retailer is to know about the product it sells, hiring salespeople who can educate customers, which implies expert knowledge. You might go into a bookstore and ask a salesperson to recommend a book. But this doesn’t scale - if you’re trying to be the world’s largest bookstore, you can’t afford to hire experts on every possible subject.

Customer still expect the retailer to provide information about the quality of the products it sells, so the expert salesperson role is replaced with an automated system that aggregates feedback from customers who purchased the product and generates a rating that serves as an indicator of quality. So in a narrow sense, Amazon is “anti-intellectual” about the products it sells – it doesn’t have knowledge, it relies on a feedback system that adapts over time.

We can see Amazon’s commitment to anti-intellectualism when this customer review system encountered a problem: how do you know if a customer review is high quality or not? The solution is the same as how to tell if a product is high quality or not: feedback. Amazon gathers feedback on the feedback, asking customers, “Was this review helpful to you?”

There are lots of other examples: Wikipedia can be seen in these terms, as can Google’s PageRank algorithm, Agile, the popular software development methodology, Twitter retweets, Tumblr reblogs, Facebook likes, Yelp restaurant reviews, social news sites like Digg and Reddit, any number of variations on the wisdom of the crowd and on and on. When Clay Shirky claimed that the problem of information overload is really a problem of filter failure, the filters he had in mind were not the expert salesperson who tells you which book to buy so you don’t have to go wandering around the store. He was talking about adaptive feedback systems that automatically and continuously sort out what we, the participants in those systems, believe is good or bad.

The key to these systems is that they take an action (which could be a wild guess at the right answer at first), receive feedback, make an adjustment and the cycle repeats – the faster, the better. The emphasis on speed and iteration makes this model a perfect complement to neoliberal capitalism, with its rapid iteration, just-in-time manufacturing, constant turn-over, optimization and creativity. Here, it appears that neoliberal capitalism is also anti-intellectual – it does not think, it simply acts and reacts to feedback.

These feedback systems often draw their inspiration from evolution. An animal does not know how its ecosystem works, and yet it is perfectly adapted to it. The animal’s DNA also doesn’t know anything about the ecosystem, yet it builds an animal that thrives. How does it do that? A random change in the DNA sequence produces a mutation in the organism’s phenotype, and if this is a useful mutation that makes the organism better adapted to its environment, it survives and reproduces more organisms with this mutation.

But this may also be a reason to be skeptical that these systems can truly be revolutionary. They are designed to adapt to the world around it, not to critically question it.

To the extent that we see our place in the world as a participant in a feedback systems, we will also never ask critical questions.