8th of March, 2025
The trouble began with a bookmark. A genuinely great web page, or a useful little gadget, which I saved to come back later. It would be a happy moment in my next day, I thought, when I come back to it. Surely enough, I did come back the next day, when I typed some letters into the browser trying to... well, what I was trying to type I cannot exactly recall. Probably something more important that starts with the same first few letters. Or was it muscle memory?
Anyhow, an autocomplete of the URL and a click later, I was on that blog from yesterday again, and stared at it, and did nothing for a few moments. Then I remembered; I was trying to look for the menu of a restaurant. Silly me, and what a coincidence that both addresses start the same.
In this example, who had bookmarked who? By giving up that screen real estate, I damaged my ability to find what I actually wanted. I'm stuck one inch deeper in Samsara, et cetera. It didn't even take a powerful algorithm to deplete my agency.
Such an extreme lack of focus probably only afflicts me personally. But what if instead, only as a thought experiment, it was a societal-scale problem that billions of people have to deal with? We imagine now that everyone had a kind of noise added to their intended action, that every time they wanted to click one of multiple buttons, there would be a good chance they would end up clicking the others. As if a chess engine were diluted by picking a random move, say, 20% of the time. And further, we suppose this random noise also translates to everything else, including the analogue world, somehow. What would be the implication?
The implication would be that the more options are recommended to the randomly-choosing people, the greater the chaos. Giving them recommendations would even be a bit unethical. You could steer people, simply by offering them things. The wider choice would dilute their intentions and lead to aimless behavior. Unlike in the real world, where recommendation can lead to discovery of previously unknown, better options, it would do the opposite — preventing discovery.
Let me try to be kind to the hypothetical diluted people by pointing out some design principles that would help them get through their day:
No, really, this is just a hypothetical. Actual people have much more agency than in the thought experiment. If, however, we dial the probability for a random action down, to 1% maybe, it might sound more realistic, and we have to ask which arguments still apply.
Recommendation is a genuine trade-off, good old. Getting to know more nice things versus getting to know just more things, without the nice in front. If we are lucky enough to have aligned incentives between builders and users, the core issue is often just a design oversight. So let's fix those.
Other times it's a tragedy of the commons, where e.g. an advertiser captures the value of a purchase by offloading the mental load of processing the ad (and other negative externalities) on everyone else. While that problem is waiting for a solution, we have to make do with browser plug-ins.
I'm reminded of a contemporary philosopher, Eva von Redecker, who offers up a concept called Bleibefreiheit — the freedom to stay put. It is meant as the opportunity to remain in an unshaken physical environment, and she showcases the lack of it in the face of industrial progress, political turmoil and environmental change. Let's at least attain a freedom to stay put in the digital world, where it is much easier to implement.