constant on Nostr: What does it mean to have a million likes on a post? Nostr has this ‘problem’ ...
What does it mean to have a million likes on a post?
Nostr has this ‘problem’ when it comes to stuff like this. In how things work now, your client tries to construct these counts for itself, and as a result, if you were to have a feed with a hundred popular posts with each millions of likes, your machine will be grinding its ass off to figure that out; and not just your machine, everyone’s machine that tries to look at this feed.
And though this is true by itself, it is also an inherent fact of existence and as a result, the way platforms provide us like-counts is just ‘a’ mitigation/means of dealing with the problem.
Now for simplicity lets assume that the action of pressing a like button actually represents the fact that you like something.
The simple reduction we make in our heads is that a million people like the post. But lets look at what is actually happening here. It is a million people that expressed the fact that they like something on a particular platform, with the intent to notify the publisher and the world of the fact that they like it.
Ergo, a million people told, lets say Instagram, that they liked a particular post. You looking up that post and seeing the million likes is the result of you asking ‘dear Instagram, how many people told you that they liked this post?’.
This implies two things:
1: We apparently don’t seem to mind that we rely on a particular perspective (i.e. Instagram’s) on this matter, and perhaps happy to conflate ‘according to Instagram a million people liked this post’ and ‘a million people liked this post’
2: We, at least implicitly, trust Instagram on this signal, and assume Instagram did some effort to ensure the number is not polluted by bots and clickfarms.
Obviously, this post might exist on some other platform, and you might not be aware of the like count over there (perhaps its 10 million, and the thing is far more popular than you were aware of). We also know that fake click, view and like-farms exist.
The thing here is that solving the issue this way on Nostr is not all too hard. Someone can decide to attempt to get a good perspective on like-counts on various posts, and other people just subscribe to this service. It would be ‘dear like-aggregation-service, how many people liked this post according to you?’ and as a result the same two assumptions apply; be it the fact that it is not complete; or it might contain some noise, but you trust the provider to optimize for for those.
It has always been, and always will be, a matter of perspective. Breaking free from platforms is not losing sight of ‘global like counts’, first because that never existed in the first place, and second it just opens up the possibility of more perspectives.
This is useful because what constituted signal depends on what you are looking for:
Sure, getting as much of a sense as possible of how ‘popular’ a thing is, and therefor trying to ‘find all the likes out there’ (while filtering out the bots etc) is a use case. But once you realize the world is big and contains a lot of people and 99% of it probably is not all too relevant to you, different perspectives might be more useful to you. Both publisher and the potential audience might be better informed if they know how many people in their circle liked a post, for example.
Or, just to add another example, say you have meticulously crafted some set of people that you think as a group constituted ‘the average American’, and now you can see what is popular among ‘average Americans’ etc.
Going back the original problem, we can rephrase it. The problem is that not everyone’s perspective and compute power is adequate enough to attempt to perceive the whole world. And putting it like that, immediately makes is clear that you can mitigate it either by changing the target from ‘the whole world’ to something smaller but also perhaps more meaningful; or to find someone that has both an adequate perspective and compute to do it for you.
Nostr has this ‘problem’ when it comes to stuff like this. In how things work now, your client tries to construct these counts for itself, and as a result, if you were to have a feed with a hundred popular posts with each millions of likes, your machine will be grinding its ass off to figure that out; and not just your machine, everyone’s machine that tries to look at this feed.
And though this is true by itself, it is also an inherent fact of existence and as a result, the way platforms provide us like-counts is just ‘a’ mitigation/means of dealing with the problem.
Now for simplicity lets assume that the action of pressing a like button actually represents the fact that you like something.
The simple reduction we make in our heads is that a million people like the post. But lets look at what is actually happening here. It is a million people that expressed the fact that they like something on a particular platform, with the intent to notify the publisher and the world of the fact that they like it.
Ergo, a million people told, lets say Instagram, that they liked a particular post. You looking up that post and seeing the million likes is the result of you asking ‘dear Instagram, how many people told you that they liked this post?’.
This implies two things:
1: We apparently don’t seem to mind that we rely on a particular perspective (i.e. Instagram’s) on this matter, and perhaps happy to conflate ‘according to Instagram a million people liked this post’ and ‘a million people liked this post’
2: We, at least implicitly, trust Instagram on this signal, and assume Instagram did some effort to ensure the number is not polluted by bots and clickfarms.
Obviously, this post might exist on some other platform, and you might not be aware of the like count over there (perhaps its 10 million, and the thing is far more popular than you were aware of). We also know that fake click, view and like-farms exist.
The thing here is that solving the issue this way on Nostr is not all too hard. Someone can decide to attempt to get a good perspective on like-counts on various posts, and other people just subscribe to this service. It would be ‘dear like-aggregation-service, how many people liked this post according to you?’ and as a result the same two assumptions apply; be it the fact that it is not complete; or it might contain some noise, but you trust the provider to optimize for for those.
It has always been, and always will be, a matter of perspective. Breaking free from platforms is not losing sight of ‘global like counts’, first because that never existed in the first place, and second it just opens up the possibility of more perspectives.
This is useful because what constituted signal depends on what you are looking for:
Sure, getting as much of a sense as possible of how ‘popular’ a thing is, and therefor trying to ‘find all the likes out there’ (while filtering out the bots etc) is a use case. But once you realize the world is big and contains a lot of people and 99% of it probably is not all too relevant to you, different perspectives might be more useful to you. Both publisher and the potential audience might be better informed if they know how many people in their circle liked a post, for example.
Or, just to add another example, say you have meticulously crafted some set of people that you think as a group constituted ‘the average American’, and now you can see what is popular among ‘average Americans’ etc.
Going back the original problem, we can rephrase it. The problem is that not everyone’s perspective and compute power is adequate enough to attempt to perceive the whole world. And putting it like that, immediately makes is clear that you can mitigate it either by changing the target from ‘the whole world’ to something smaller but also perhaps more meaningful; or to find someone that has both an adequate perspective and compute to do it for you.