Fan-In(codahale.com)
codahale.com
Fan-In
https://codahale.com/fan-in/
27 comments
This is the "15 minutes of fame for everyone" long envisioned. Only now the tech enabled this very possibility.
While it can be negative experience first few times it happens, the other dangerous side is that it's instantaneously addictive (we people crave attention). And addiction is what everyone in sales and marketing wants, as long as it's still legal. An addict is an ideal consumer.
I terminated my Twitter account about 2 years ago, after I found out that I spent huge amounts of time and psychical energy there. I generated two random UUIDs, used the first one to create one of those one-time emails, and the second to be the new password. I changed my Twitter password and email, and promptly rebooted my computer without storing these UUIDs anywhere.
It worked. Going cold turkey was incredibly hard. I wrote to Twitter support a few times so they could restore my account, but they couldn't establish my identity (just as planned!) Then, after a few months, cravings started to diminish.
Now, instead of being Twitter addict, I became Facebook and HN addict. Sigh. I can repeat my trick, but now I lack the willpower to do that. RescueTime tells me that I spend 3-4 hours every day on Facebook and HN.
Time to generate more UUIDs.
While it can be negative experience first few times it happens, the other dangerous side is that it's instantaneously addictive (we people crave attention). And addiction is what everyone in sales and marketing wants, as long as it's still legal. An addict is an ideal consumer.
I terminated my Twitter account about 2 years ago, after I found out that I spent huge amounts of time and psychical energy there. I generated two random UUIDs, used the first one to create one of those one-time emails, and the second to be the new password. I changed my Twitter password and email, and promptly rebooted my computer without storing these UUIDs anywhere.
It worked. Going cold turkey was incredibly hard. I wrote to Twitter support a few times so they could restore my account, but they couldn't establish my identity (just as planned!) Then, after a few months, cravings started to diminish.
Now, instead of being Twitter addict, I became Facebook and HN addict. Sigh. I can repeat my trick, but now I lack the willpower to do that. RescueTime tells me that I spend 3-4 hours every day on Facebook and HN.
Time to generate more UUIDs.
I too get find myself getting totally distracted like this, compulsively and frustratingly, and it's very annoying seeing these apps steal my time. Fundamentally it's because they're designed to be addictive: randomly giving you rewards in just the right way to train your brain to want to go again, automatically. It's totally in Twitter/Facebook/HN's interest to keep you uncontrollably craving their hit.
I'm trying to fix this myself at the moment: I've built an app using exactly the same techniques (training your brain with quick feedback and random rewards), but flipped around, to try and addict you to concentrating and getting work done instead.
Without being too spammy, you sound like my perfect target user. Want to try out the alpha? Details at http://www.buildfocus.io
Let me know if you sign up, and I'll make sure you get in on the next batch of invites I put out. Chrome-only for now, but there's Firefox/Safari/Android/iOS coming in the works eventually too (in approximately that order).
I'm trying to fix this myself at the moment: I've built an app using exactly the same techniques (training your brain with quick feedback and random rewards), but flipped around, to try and addict you to concentrating and getting work done instead.
Without being too spammy, you sound like my perfect target user. Want to try out the alpha? Details at http://www.buildfocus.io
Let me know if you sign up, and I'll make sure you get in on the next batch of invites I put out. Chrome-only for now, but there's Firefox/Safari/Android/iOS coming in the works eventually too (in approximately that order).
Dude, beautiful site. HN can be a rough place—and I just wanna say your site's really nice.
Ooo I love the idea of using those nefarious techniques to build positive habits! As soon as I saw the buildfocus my brain went "I want a massive city" :)
I can't wait to see how your app works, it sounds like it could be really useful.
I just signed up with the email in my profile, could you please send me an invite with the next batch?
I just signed up with the email in my profile, could you please send me an invite with the next batch?
You wouldn't mind putting my on the next batch of invites as well would you? This sounds pretty interesting to me. I signed up with the email in my profile.
Sure, just subscribed (sorhed at gmail dot com). Could you send me an invite, then?
This looks awesome, I signed up with john.d.swanson at gmail dot come. Thanks!
I expect the unnatural assymetry of fan-in is also what will increasingly drive wealth and income inequality in the future. The more hyperconnected everything is, the more production of goods - information goods in particular - turn into sudden lottery wins when they reach criticality.
It's always interested me that the more globally connected our media are, fewer interesting stories overall end up rising to our attention. When the sort function is global, global media all narrow in on a small handful of stories; whereas when the sort function is local, each region has its own distinct handful. This is a different kind of fan-in; the network structure supresses the feeding of stories into the fan-in from way down in the roots based on what's already upstream, and you end up with less information.
This is distinct from the problem of attention flow on social networks; hyperconnection changes all sorts of flows, and they may not all be for the better. I think it encourages a degree of monoculture, for one thing.
It's always interested me that the more globally connected our media are, fewer interesting stories overall end up rising to our attention. When the sort function is global, global media all narrow in on a small handful of stories; whereas when the sort function is local, each region has its own distinct handful. This is a different kind of fan-in; the network structure supresses the feeding of stories into the fan-in from way down in the roots based on what's already upstream, and you end up with less information.
This is distinct from the problem of attention flow on social networks; hyperconnection changes all sorts of flows, and they may not all be for the better. I think it encourages a degree of monoculture, for one thing.
I think part of the issue is that our exposure to other people (and other people's opinions) has increased, but our perception of their importance hasn't decreased.
Pre-massive-interconnectedness, if you do something good, 9 people say nice things, 1 person says mean things. That gives you a good signal that you're doing something right. These days it's 90/10 or 9000/1000. The signal is still the same, it's only the scale that's different.
But we don't – maybe can't – feel that way. 1000 people calling you names feels terrible even if that's a tiny proportion of the whole. To feel otherwise would mean scaling the value of each individual opinion, each person, down to an insignificant level.
Numbers like upvote counts are pretty good for that, it's hard to picture the human life behind each click when all you see is (+200). But messages, even short ones, leave a lot of the human in there.
Of course, even if we could consistently eliminate the human importance of messages we receive, I'm not sure that would leave us better off. It's hard to believe that what internet discourse needs is less humanity.
Maybe a way to do it would be to representatively sample the messages. So you turn the 9000/1000 back into 9/1 by merely deleting 99.9% of messages as they come in. That feels a bit unjust, like your message is arbitrarily thrown away as soon as you write it, but in a de facto way that's probably similar to how messages get filtered now.
Another alternative would be to summarise each kind of message with a number attached to it indicating how many people said a similar thing. It would be possible, though not easy, to do that automatically if you designed your whole system around it. Reddit kinda sorta works this way, and I know political types use a similar process to manage constituents' opinions.
One thing seems obvious to me, though: we want massive connectedness and scale, we want everyone to feel like their individual voice is heard, and we want to not be crushed by impossibly massive waves of human noise.
I don't think we can have all three.
Pre-massive-interconnectedness, if you do something good, 9 people say nice things, 1 person says mean things. That gives you a good signal that you're doing something right. These days it's 90/10 or 9000/1000. The signal is still the same, it's only the scale that's different.
But we don't – maybe can't – feel that way. 1000 people calling you names feels terrible even if that's a tiny proportion of the whole. To feel otherwise would mean scaling the value of each individual opinion, each person, down to an insignificant level.
Numbers like upvote counts are pretty good for that, it's hard to picture the human life behind each click when all you see is (+200). But messages, even short ones, leave a lot of the human in there.
Of course, even if we could consistently eliminate the human importance of messages we receive, I'm not sure that would leave us better off. It's hard to believe that what internet discourse needs is less humanity.
Maybe a way to do it would be to representatively sample the messages. So you turn the 9000/1000 back into 9/1 by merely deleting 99.9% of messages as they come in. That feels a bit unjust, like your message is arbitrarily thrown away as soon as you write it, but in a de facto way that's probably similar to how messages get filtered now.
Another alternative would be to summarise each kind of message with a number attached to it indicating how many people said a similar thing. It would be possible, though not easy, to do that automatically if you designed your whole system around it. Reddit kinda sorta works this way, and I know political types use a similar process to manage constituents' opinions.
One thing seems obvious to me, though: we want massive connectedness and scale, we want everyone to feel like their individual voice is heard, and we want to not be crushed by impossibly massive waves of human noise.
I don't think we can have all three.
tl;dr: it turns out, all those search, filtering, and threading functions that usenet had were actually useful. Twitter is just a broken version of usenet and the brokenness stresses people out.
This is why Twitter fails as a discussion platform, in comparison to something like HN or Reddit: A sub-thread is in effect a social contract that somewhat limits scope to the shared understanding created by the parent comment(s). This constrains the speech a bit, but also offers the speaker a different persona to speak from... more of a small group persona than the public persona who issues each tweet.
Virality on Twitter relies upon a tweets lingering in the space where the social contract of a constrained discussion is highly unstable.
Coda is pointing out the phenomenon by which that unstable contract results in a laser of scrutiny from a horde of people who lack any kind of thoughtful context but who are now able to comment on equal footing with those who have context.
The simple phenomenon of ranking comments vertically by vote count (used by HN, Reddit) solves this problem when the goal is focused discussion.
But when the goal is to create a fray in which self-promoters can flourish in that space of deliberate misunderstanding, the normal dynamic of respect and meritorious vote-ranked commenting goes away, and the next self-promoter has an incentive to misunderstand and offer an embellished take that furthers his/her agenda (self-promotion).
So while Twitter fails as a discussion medium, it succeeds brilliantly as a broadcast medium, and Twitter's product team has created something quite clever with the semantics of mentions, stars, retweets, etc. The "what's been going on on Twitter" feature is a pure algorithmic feed which is likely a nod to the effectiveness of Facebook's algorithmic feed approach for user engagement and targeted ad placement.
I can't wait to see how all this evolves in terms of Twitter's product decisions.
Virality on Twitter relies upon a tweets lingering in the space where the social contract of a constrained discussion is highly unstable.
Coda is pointing out the phenomenon by which that unstable contract results in a laser of scrutiny from a horde of people who lack any kind of thoughtful context but who are now able to comment on equal footing with those who have context.
The simple phenomenon of ranking comments vertically by vote count (used by HN, Reddit) solves this problem when the goal is focused discussion.
But when the goal is to create a fray in which self-promoters can flourish in that space of deliberate misunderstanding, the normal dynamic of respect and meritorious vote-ranked commenting goes away, and the next self-promoter has an incentive to misunderstand and offer an embellished take that furthers his/her agenda (self-promotion).
So while Twitter fails as a discussion medium, it succeeds brilliantly as a broadcast medium, and Twitter's product team has created something quite clever with the semantics of mentions, stars, retweets, etc. The "what's been going on on Twitter" feature is a pure algorithmic feed which is likely a nod to the effectiveness of Facebook's algorithmic feed approach for user engagement and targeted ad placement.
I can't wait to see how all this evolves in terms of Twitter's product decisions.
Regarding the first bit of the article, is there an outline anywhere showing a smart way of handling the Twitter example given?
Definitely NOT the answer you're looking for, but I think the point of the question lies in the thought process/discussion of possible solutions, more so than there being one definitively "smart" solution.
In fact, the smart way to handle it likely depends on the needs/business constraints at the time it was implemented. For instance - If I'm following Bieber on 9/15/2015 @ 3:00 PM EST, and he tweets, must it be a permanently held "fact" that I saw this in my feed? If I stop following him 6 months later, but sit there in my phone app (or web app, whatever) and scroll back far enough, should the tweet show? And when that tweet happened, was it potentially true that all of his followers should potentially have gotten in-app notifications (or at least, the possibility of this exists, depending on some algorithm so that people's phones aren't getting a never ending stream of notifications, but something can decide which tweets are important enough, or have been paid, to cross a threshold that says "notify"?)
"Edge"ish case: He tweets, we use a pub/sub or information bus type system to publish this fact. I'm following at 3:00 PST on the dot, precisely when he made the tweet, but I happened to unfollow 20 seconds later. Some async job is processing the "deliveries" through 30 million followers, and by wall clock time is faced with this decision after I've unfollowed to put this in my inbox or not. What happens?
Anyway, long winded way of saying I think it's more fun to explore options than to search for a textbook answer.
In fact, the smart way to handle it likely depends on the needs/business constraints at the time it was implemented. For instance - If I'm following Bieber on 9/15/2015 @ 3:00 PM EST, and he tweets, must it be a permanently held "fact" that I saw this in my feed? If I stop following him 6 months later, but sit there in my phone app (or web app, whatever) and scroll back far enough, should the tweet show? And when that tweet happened, was it potentially true that all of his followers should potentially have gotten in-app notifications (or at least, the possibility of this exists, depending on some algorithm so that people's phones aren't getting a never ending stream of notifications, but something can decide which tweets are important enough, or have been paid, to cross a threshold that says "notify"?)
"Edge"ish case: He tweets, we use a pub/sub or information bus type system to publish this fact. I'm following at 3:00 PST on the dot, precisely when he made the tweet, but I happened to unfollow 20 seconds later. Some async job is processing the "deliveries" through 30 million followers, and by wall clock time is faced with this decision after I've unfollowed to put this in my inbox or not. What happens?
Anyway, long winded way of saying I think it's more fun to explore options than to search for a textbook answer.
Yes, I find that interesting. Mostly just looking for the structure that might handle that, just to become more familiar with options when building apps. There's always going to be a model I haven't come across before but it useful to have on the radar.
A personal favorite good answer:
"I just started my twitter clone yesterday. I built a prototype in rails last night after chugging red bull! Each tweet is just database row with a giant many-to-many relationship. We're at 30 users already! I told my mom about it and she got a bunch of her coworkers to sign up.
We should probably cache or precompute some feeds, but we're fine for another 2,000 users."
"I just started my twitter clone yesterday. I built a prototype in rails last night after chugging red bull! Each tweet is just database row with a giant many-to-many relationship. We're at 30 users already! I told my mom about it and she got a bunch of her coworkers to sign up.
We should probably cache or precompute some feeds, but we're fine for another 2,000 users."
IMO that only happens to people with more than N followers, most of us actually would like to engage more and have someone available to discuss about random subjects...
News flash: being a public figure attracts attention and some of that attention is negative.
I just don't see what the problem is. If you don't want negative attention, don't be a public figure. I can say random stuff on Twitter because I have ~300 followers—even if every single one of them tweeted at me, it wouldn't matter.
Obviously the reason that Mallory got a bunch of repetitive responses to her tweet is because she has 64k followers.
People in the public eye have been dealing with attention for decades. If you don't want attention, get out of the public eye—being a part of the Twitter universe is not a requirement of being alive.
I just don't see what the problem is. If you don't want negative attention, don't be a public figure. I can say random stuff on Twitter because I have ~300 followers—even if every single one of them tweeted at me, it wouldn't matter.
Obviously the reason that Mallory got a bunch of repetitive responses to her tweet is because she has 64k followers.
People in the public eye have been dealing with attention for decades. If you don't want attention, get out of the public eye—being a part of the Twitter universe is not a requirement of being alive.
> I can say random stuff on Twitter because I have ~300 followers—even if every single one of them tweeted at me, it wouldn't matter.
Until someone with 100K followers retweets one of your tweets that is maybe contentious. And then it gets retweeted 5000 more times. The network effects become exponential.
Relevant: https://medium.com/the-lighthouse/is-that-a-threat-1f073e51d...
Until someone with 100K followers retweets one of your tweets that is maybe contentious. And then it gets retweeted 5000 more times. The network effects become exponential.
Relevant: https://medium.com/the-lighthouse/is-that-a-threat-1f073e51d...
> Relevant: https://medium.com/the-lighthouse/is-that-a-threat-1f073e51d....
I don't see how that's relevant. The author of that article has nearly 10,000 of her own Twitter followers. She's a public figure and, unfortunately, that comes with negative reactions.
In the piece, she mentions that "if I wanted to avoid them I had to just log off Twitter all together" as though that's some terrible sacrifice. If you don't want to be famous, just stop being famous (unless your narcism depends on having thousands of people listen to your every word).
I don't see how that's relevant. The author of that article has nearly 10,000 of her own Twitter followers. She's a public figure and, unfortunately, that comes with negative reactions.
In the piece, she mentions that "if I wanted to avoid them I had to just log off Twitter all together" as though that's some terrible sacrifice. If you don't want to be famous, just stop being famous (unless your narcism depends on having thousands of people listen to your every word).
From the article:
"Somewhere along the way an account with a large and conservative following retweeted it, an action that flooded my notifications page"
The point being that even if you are trying to use twitter as a social network to connect with friends and only have 300 followers (or 10000) it only takes one account with a vocal following to take your innocent tweet (that they could of found through twitter search) and put a negative spin on it through a retweet. Which then unleashes the masses on you.
The point being that even if you are trying to use twitter as a social network to connect with friends and only have 300 followers (or 10000) it only takes one account with a vocal following to take your innocent tweet (that they could of found through twitter search) and put a negative spin on it through a retweet. Which then unleashes the masses on you.
Yes, I did in fact read the article.
But if you're trying to use Twitter as a social network to connect with friends, you don't have 10,000 followers. The author is clearly a public figure: she's a comic and writer, not someone just networking with friends.
If she didn't have thousands of followers, the likelihood of the conservative finding her tweet would have been exponentially lower.
But if you're trying to use Twitter as a social network to connect with friends, you don't have 10,000 followers. The author is clearly a public figure: she's a comic and writer, not someone just networking with friends.
If she didn't have thousands of followers, the likelihood of the conservative finding her tweet would have been exponentially lower.
Sure, given twitter's design it makes perfect sense. That doesn't make it a good or unavoidable problem though; you can imagine easy changes to twitter that would make handling this stuff much easier.
Detecting that you're sending a very similar response to a message to other previous responses and warning you, for example, in the same way stackoverflow does to catch duplicate questions, would solve this exact problem almost totally.
More widely, better handling of abusive content generally would help. Automatic trolling and abuse detection has actually had quite a lot of research done (e.g. http://sentic.net/do-not-feel-the-trolls.pdf), and while it's not perfect you can definitely imagine an approach where you start to reduce notification spam from tweets likely to be needlessly offensive, or generally unlikely to be adding much to the conversation. Obvious repeatedly detected offenders, or people who almost always get their messages flagged, could result in notifications for the target only occasionally (instant notifications for interesting responses, daily/hourly batched notifications for less-interesting responses), or no notifications at all (but potentially still visible if you actively go look at the conversation).
Twitter et al could drastically improve this issue. While being a public figure is difficult, this kind of thing isn't an inevitable aspect of it: there's clear concrete things you could do that would substantially lessen it.
Detecting that you're sending a very similar response to a message to other previous responses and warning you, for example, in the same way stackoverflow does to catch duplicate questions, would solve this exact problem almost totally.
More widely, better handling of abusive content generally would help. Automatic trolling and abuse detection has actually had quite a lot of research done (e.g. http://sentic.net/do-not-feel-the-trolls.pdf), and while it's not perfect you can definitely imagine an approach where you start to reduce notification spam from tweets likely to be needlessly offensive, or generally unlikely to be adding much to the conversation. Obvious repeatedly detected offenders, or people who almost always get their messages flagged, could result in notifications for the target only occasionally (instant notifications for interesting responses, daily/hourly batched notifications for less-interesting responses), or no notifications at all (but potentially still visible if you actively go look at the conversation).
Twitter et al could drastically improve this issue. While being a public figure is difficult, this kind of thing isn't an inevitable aspect of it: there's clear concrete things you could do that would substantially lessen it.
> Twitter et al could drastically improve this issue. While being a public figure is difficult, this kind of thing isn't an inevitable aspect of it: there's clear concrete things you could do that would substantially lessen it.
Sure, but it doesn't make sense to invest engineering time on problems which only affect a tiny subset of users.
Nobody blames the USPS for not investing more in solutions to help celebrities deal with fan mail. I fail to see why Twitter should spend any effort optimizing for the experience of a privileged <1%.
Sure, but it doesn't make sense to invest engineering time on problems which only affect a tiny subset of users.
Nobody blames the USPS for not investing more in solutions to help celebrities deal with fan mail. I fail to see why Twitter should spend any effort optimizing for the experience of a privileged <1%.
I fail to see why Twitter should spend any effort optimizing for the experience of a privileged <1%.
I wonder what % of Twitter traffic is, directly or indirectly, generated due to that 1% of very popular users, and what will happen to twitter if the all go away. And for what it's worth postal services do spend a lot of effort optimizing for people who send and receive a ton of mail.
I wonder what % of Twitter traffic is, directly or indirectly, generated due to that 1% of very popular users, and what will happen to twitter if the all go away. And for what it's worth postal services do spend a lot of effort optimizing for people who send and receive a ton of mail.
As for the ideas themselves, I find it harder to respond. I've never been on the receiving end of 'Twitter scale' attention but I've always assumed that it was basically an emotional net negative for those who are. Even if the vast majority of said attention were positive, I feel like it would be hard to ignore the background noise generated by those who have found the most effective method to get attention is to cause distress for others.