I wrote this up Friday afternoon as part of an attempt at a business school essay. I’m too anxious to actually apply, but I do think it will be useful to use this blog as a collection of all the thoughts I have over the next few months, even if some of those thoughts feel tangentially related to language models. These posts in particular will be far from polished or thought out (and, TBH, probably not always properly cited), but will hopefully be something I can look back on as evidence of what I worked on and thought about during this time. So, taking all the MBA things out, here is what I’m thinking about this week:
Update (3/30): My tied-for-first-favorite-former-housemate Lucia published lessons learned from the Maker Movement so far. She is a literal doctor in these things and many of my thoughts here are largely gleaned from her tweets so… read that instead!
Physical infrastructure also needs a virtual counterpart
We need to understand how services historically offered by physical infrastructure might translate to the virtual. If we can’t occupy parks and Main Streets over the next few weeks, how will existing virtual spaces pick up the slack (some pun intended)?
Ethan Zuckerman has been arguing recently for greater attention on what he calls digital public infrastructure. He argues that social media today wasn’t actually designed to be socially valuable, but to grab attention and absorb personal information for use in targeted advertising and notes that societies with robust public media (the BBC in the 1920s and NPR in the 1970s) have better-informed publics.
While some companies – e.g. NextDoor, or standalone hyper-local sites focus on providing local information – what is needed is things that are both bigger and smaller: bigger in the sense that they command real attention from large swaths of the population – and smaller in the sense that they serve very specific needs of communities.
If existing social media platforms represent a market failure, how do we solve this? And, in the meantime, what virtual tools will people gravitate toward to substitute for public places? How will the failures of private platforms here influence organizing mutual aid and medical supplies over the coming weeks?
Disinformation detection is hard and important, but there are other roadblocks
Robust virtual public spaces and public media are important because they help stop the spread of misinformation.
There are two modes through which to think about building detection mechanisms for machine-generated disinformation: in one, we need to build a machine smarter than a model, in the other, it is sufficient to just train a new model on the output of the other. Either way, most people agree, that it becomes harder and harder to be right 100% of the time. Detection models can help narrow down the feature space, but they can’t make every decision for us. Moreover, it is difficult to get even the most sophisticated of detection models into the hands of those who need them.
Since detection is so hard, what should we do? How do we build low-tech interventions to help people with conflict-specific (or, in this case, COVID-specfic) knowledge sift through reports that come in and sort out false or machine generated reports?
Connectivity is already an issue
All of the above assumes that if we had robust online public spaces people would be able to access them. This isn’t the case, huge swaths of the US remain disconnected from high-speed internet. If digital public infrastructure isn’t your beat, don’t worry, there’s still plenty of traditional infrastructure work to do.