It depends on what your personal philosophy of blocklists is. If you believe that on a supermassive social network that no single person's content is worth anything and can safely discarded then what you're interested in is signal and a low false negative rate.
Who a person follows is a pretty strong signal of their values. Is it always accurate, no, you stand as an example of such a person but their feed is still full and the kind of people they have in mind they don't want to see are gone.
Reddit has a similar culture with extensions that let you hide users that subscribe to or have ever posted in certain subreddits. It works better than it has any right to.
> Who a person follows is a pretty strong signal of their values
I strongly disagree, and believing this to be true is a problem for public discourse IMO.
What a person posts and who they repost is a strong signal of their values. I dislike/disagree with the validity of a social construct where people who choose to stay informed are assumed to hold values they do not.
+ most of my inclusion in these blocklists is due to actions taken by someone i follow after i followed them (some AI person I followed who posted bluesky posts on huggingface). it really is ridiculous to read a follow as an endorsement
i think this is more of a cultural issue with bluesky than a technical issue/flaw
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. But the fun thing about statistics is I that I can assume a great many things about the set of all people who follow Evil McBaddington.
I see this "scale invariant fallacy" all the time with programmers. The forward direction being a thing and that thing in a for loop are equivalent in essence, and the reverse being that a property being true about a set means it's true for its members (can decompose aggregate statements into an equivalent for loop
acting on each member).
> 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.
> 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."
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.
Nobody cares about your public discourse. Your opinion is not that important to 99.9999% of the people on this planet. I use social media to follow people I like and see funny jokes and if a few people undeservedly catch a block oh well, that's the price of a curated feed.
Call it an echo chamber, but nobody has a right to have their posts be seen
I was reacting to the very direct and unqualified statement that who a person follows strongly indicates their values.
If you’re just using social media for entertainment, I don’t care who you block or how. But perpetuating the idea that “follow equals endorsement” is a problem regardless.
This has less to do with people having a right to be seen and more to do with the perpetuation of ideas that make society more polarized in spaces where such things matter.
To put this another way, I’d react less strongly to “blocking people based on who they follow weeds out almost everyone I don’t want to see and I accept the false positives”. I’d still express concerns about the resulting echo chamber but that’s a far more reasonable position.
Who a person follows is a pretty strong signal of their values. Is it always accurate, no, you stand as an example of such a person but their feed is still full and the kind of people they have in mind they don't want to see are gone.
Reddit has a similar culture with extensions that let you hide users that subscribe to or have ever posted in certain subreddits. It works better than it has any right to.