4 ms·

> I don't think you can really decide whether something is a useful signal or not. I don't and can't assume anything about an individual because they follow Evi

> I don't think you can really decide whether something is a useful signal or not. I don't and can't assume anything about an individual because they follow Evil McBaddington. Maybe I misunderstood your comment, but isn’t that exactly the opposite of what you’re saying here? > Who a person follows is a pretty strong signal of their values If your position is closer to “on average, people who follow other people tend to be more aligned with the people they follow”, I’d probably agree. But this is a vastly different statement than “following someone strongly indicates endorsement”. > I see this "scale invariant fallacy" all the time with programmers. I think it’s problematic to bring too much of an engineering mindset to a social issue like this. Algorithms can be cleanly and rigidly defined. Social issues tend to be messy and complicated. In terms of pure utility, it may be true that blocking people based on who they follow results in a clean feed at a low (effort) cost. But perpetuating the idea that following someone equals endorsement has longer term social costs and gets into dystopian territory pretty quickly, and depending on why someone is using social media, the impact of entirely excluding the group of people that “follow but disagree” is likely not small. If you’re on there for cat memes, probably not a big deal. If you’re trying to stay informed or be politically active, probably a bigger deal. And regardless of use case, believing things about other people reflexively based on something like follow status is not a healthy place for society to be. Most people do not appreciate the nuance of what you’re describing as “scale invariant fallacy”. My biggest issue here is the framing and the questionable/harmful conclusions that can be derived from it.

Spivak 2y ago
> If your position is closer to “on average, people who follow other people tend to be more aligned with the people they follow” Maybe it's a semantic thing but is that not what is meant by this usage of the word signal? "A child growing up in an affluent neighborhoods is a strong signal of their future success." implies => "The proportion of the set of all kids who live in affluent neighborhoods who achieve success as adults is high."
- haswell 2y ago
  That’s not how I read your use of signal in the broader context. Putting people on lists for who they follow is a growing trend and it’s pretty clear in Bluesky culture that many people actually believe following someone implies some kind of support for that person and “taints” the follower somehow. This goes alongside a host of problematic behaviors ranging from all-or-nothing thinking “if you’re not with us you’re against us” to simple bad faith. My goal isn’t to argue here, but to understand and clarify what you meant. It sounds like you didn’t literally mean that on an individual basis, following someone strongly indicates their values, but that in the aggregate (due to the current social climate), it’s more likely they won’t follow people they disagree with. I don’t think this is just a matter of semantics given the context, but a necessary clarification. In the affluent neighborhoods example, there’s no call to action (block people), the situation isn’t a highly charged issue in the cultural zeitgeist, and there aren’t a lot of reasons someone would interpret that example in some other way. The same is not true for the current social media climate.

Spivak 2y ago

> If your position is closer to “on average, people who follow other people tend to be more aligned with the people they follow” Maybe it's a semantic thing but is that not what is meant by this usage of the word signal? "A child growing up in an affluent neighborhoods is a strong signal of their future success." implies => "The proportion of the set of all kids who live in affluent neighborhoods who achieve success as adults is high."

haswell 2y ago
That’s not how I read your use of signal in the broader context. Putting people on lists for who they follow is a growing trend and it’s pretty clear in Bluesky culture that many people actually believe following someone implies some kind of support for that person and “taints” the follower somehow. This goes alongside a host of problematic behaviors ranging from all-or-nothing thinking “if you’re not with us you’re against us” to simple bad faith. My goal isn’t to argue here, but to understand and clarify what you meant. It sounds like you didn’t literally mean that on an individual basis, following someone strongly indicates their values, but that in the aggregate (due to the current social climate), it’s more likely they won’t follow people they disagree with. I don’t think this is just a matter of semantics given the context, but a necessary clarification. In the affluent neighborhoods example, there’s no call to action (block people), the situation isn’t a highly charged issue in the cultural zeitgeist, and there aren’t a lot of reasons someone would interpret that example in some other way. The same is not true for the current social media climate.