Monday, April 24, 2017

On Social Spaces

This post is basically a response to Against Facebook: Comparison to Alternatives and Call to Action. There are some important needs that I believe are not addressed by Zvi’s proposal that we all switch to blogs and Twitter instead of Facebook.

First off, the things we agree on:

  • Facebook is evil and Molochian. Its incentives are not your incentives. 
  • Specifically, Facebook is conspiring to eat more and more of your time with its news feed. This is a serious problem. 
  • Facebook isn’t good for building deeper relationships. It’s not a good place to say things of real substance that take time to think about. 
  • More people should start blogs and put their extended thoughts there, instead of on Facebook. (see: You’re Good Enough, You’re Smart Enough, and People Would Like You
  • Conditioned on there being a good Facebook alternative without evil incentives, we should definitely all switch to it. 
I should also note here that, while I think Zvi is wrong*, I don’t think his views are going to cause any long-term harm. I predict that the community, if it tries to take his advice, will soon naturally snap back to its usual mode of online communication in a mixture of Facebook, Tumblr, other social media, and blog posts. More on why I believe this later.

Things I disagree with Zvi about:
  • A blog and a public Twitter account are a good Facebook alternative. 
Or more specifically:
  • A public Twitter account is a good place to put brief thoughts that don’t seem appropriate for a blog. It has all the important features of Facebook, without the evil downsides. 
To start on a point of agreement: Starting a blog feels scary. It seems like you have to say things that are Important or Well Thought Out to have one. Now, one might say that “You’re Good Enough, You’re Smart Enough, and You Should Start A Blog Anyway.” That’s true. I appreciate that, and that’s part of what encouraged me to expand this into a blog post from a Facebook comment. But the fact is, I don’t think I would have had the idea to write it if it hadn’t started as a Facebook comment.

So: Not all thoughts that are worth sharing can overcome the barrier to entry of writing a blog post. And many of those thoughts would eventually overcome that barrier, given a safe place to be nurtured into growth. But Zvi’s proposed alternative for briefer thoughts, public Twitter, is not a safe place.

Here’s the thing: sometimes people want things to be ephemeral and undiscoverable after the fact. In other words, for some kinds of communication, not having a search function is a feature.

Twitter is public by default. Sometimes people want to write things that only their friends or people they trust or people they know can see. It seems that Zvi doesn’t need this type of private communication, and that’s fine for him, but for me and many other people, it’s really not fine to have all online spaces be findable by doxxers 10 years from now. Or by my weird conservative uncle 10 minutes from now.

(You can make your Twitter “protected,” but as far as I know, it doesn’t give you any granular control over which of your followers get which content. You can’t segment your identity, the way you can on Facebook or Tumblr.)

Twitter is discoverable. It’s easy to search someone’s history. It’s usually indexed by search engines.

There is an expectation on Twitter that you will be able to read all of someone’s tweets. Thus, Twitter is more of a stage. Many people find it difficult to speak when they feel that too many others are watching them.

Twitter isn’t great for conversations, either. Zvi mentions that Facebook has poor comment threading, but Twitter’s is even worse, by any reasonable standard.

Facebook Is Better Than You Think It Is

Facebook makes it feel easy to comment on things, because you know you won’t be forced to face the consequences of your words being repeated for all of time. The UI makes it also feel simple and low-barrier-to-entry - “type in the small box, press Enter,” instead of “type in the large box, enter your personal details, publish.” Publish is a scary word.

I’ve had several interesting conversations in Facebook comments that made me feel closer to the people involved; I would credit Facebook with several important friendships of mine being nontrivially improved. I could not have had those conversations on public Twitter or a blog, where I felt that people were watching me. I’m not sure I could have had them on private Twitter, either. Sometimes I want my words to stay in small text, rather than large.

In other words, Facebook creates an online space for conversations that feel very much like in-person conversations at a party, with a sense of privacy and social safety that I don’t get in almost any other online space. And when I don’t feel safe enough outside of that space, it’s the only place where I can speak at all.

Tumblr is rather good at this too, despite being public, by being pseudonymous and poorly discoverable. This makes it harder to actually find your friends on the network, though, so it often feels lonely. Facebook allows you to find your real friends and also segment your identity (using Friends Lists and Groups and so on), so that you can interact safely with less need for pseudonymity.

What I’d like to say here overall is that the structure and affordances of social spaces are important, and also hard to predict from first principles. You can’t just say “well that doesn’t matter” about certain features or changes (or things that look to you like bugs!) that nonetheless affect people’s willingness to engage in particular ways. And there are frequently reasons that people are on one website rather than another, that aren’t immediately obvious and in fact take significant thought and research to find out.

Case Studies of Switching Social Networks

Network effects aren’t nearly as hard as you’d expect to overcome, provided that the alternative is actually compelling. I think people (who are likely to read this blog) systematically overestimate the power of switching costs and underestimate the importance of how good the existing network is in keeping people in the same online spaces.

An unsuccessful example I want to use here is the movement from rationalist tumblr to Diaspora. I tried to start a movement to get everyone on rat-tumb to switch to Diaspora instead, and it sort of worked. The thing is, Diaspora isn’t that good. It has several missing features compared to Tumblr, like asks. It’s buggy. It’s hard to find people. It’s hard to sign up. (I tried to find the Diaspora account I started to flesh out the details here, but the server I was using has shut down.) I got a bunch of people to sign up for Diaspora, and even post some content, but they switched back to Tumblr in short order.

By contrast, a successful example: my friend group has entirely switched over to Telegram now from other messaging services. Telegram is awesome. It is solidly better than any other messaging app I’ve used, and it mostly does this by not sucking in various small ways. However, the thing that got us to switch over initially was the stickers. Telegram allows custom, user-created stickers, and that makes it fun and exciting to use at first. This convinced my partner to try it, who convinced me, and I convinced several other of our friends; now most of my close friends are on it. And though it was the stickers that made us try it, it’s the consistent quality that makes us stay. If we moved back to Hangouts, I would miss the speed of the app, its ability to handle poor connections gracefully, its superior desktop clients, and yes, the stickers.

Another unsuccessful example: the recent rationality community attempt to move everyone over to Signal messenger. This can be succinctly described as an amusing minor disaster. Someone started a party chat that numerous (20-30?) people joined. It was extremely annoying to use, because whenever someone switched devices, every single other person in the chat would get a notification telling them to verify that person’s security number. There was a period when every message sent in the chat was repeated four or more times. The last nail in the coffin was this: it was impossible to leave the chat. People on iDevices who tried to leave the group chat continued getting push notifications on their phone, until they uninstalled the app entirely in disgust. We made a group pact to never enter the chat again.

My conclusions from these experiences: People do have a limited tolerance for this stuff. If you’re asking them to try a new messaging app every week, they probably won’t bother. In other words, their tolerance is a common good that is easily consumed.

But if you can pitch a decent alternative, and the commons is still in good shape, it’s actually pretty easy to get a bunch of smart people to try a new social network. It’s much harder to get them to stay on the new thing if it’s worse than the old one, or missing something that’s important to them. And you may have a good pitch for your alternative, but making sure it has those other, subtler things is another game entirely.


What I want to leave you with is this:
  1. Social safety and private spaces are important. We can’t make all online conversations public; that will just stifle private conversations. 
  2. Writing good social software, with the right affordances, is hard. It’s harder than you think, and most people get it wrong. You might think, “Because I’m smart and rational, I can write a better app / switch my friends to a better app that I’ve found!” But beware: the crowd may have more wisdom than you give it credit for. And if your alternative isn’t truly compelling, they’ll drift back to the last local maximum pretty quickly.
  3. But getting people to switch to a better social network is easier than you think - as long as it’s really better.
Trying to kick the community out of a local maximum and into a better social space is a great idea. I’m glad people are doing it - once in a while, so we don’t use up that precious commons.

I’m pretty sure this particular attempt won’t succeed, because of (1). And a lot of them won’t succeed in general, because of (2). But I do think it’s possible to do better, because of (3). It just might take some time.

*PS: Lastly, a Possibly-Unnecessary Note on Zvi As A Person

Zvi is a nice person who took me to a pastrami place once, and it was tasty. I also appreciate his encouragement for more people to blog. I hope he does not take this disagreement as an indictment of him or his needs.


  1. The note is entirely unnecessary; this is exactly the kind of thoughtful argument I hope to see to things I post online, and I greatly appreciate your past attempts described here to make things suck less in various ways.

    I'm sad to hear about the Signal issues; I joined but don't message much so I haven't had any issues. My guess is that it is worth having around because when you need it, you really need it, but that using it for regular day-to-day is way more trouble than it could possibly be worth.

    I see two core arguments here that run counter to my post: One is that Facebook is actually pretty good at facilitating low-cost, low-investment, ephemeral and private communication. Two is that keeping such communication private is important, and that there is no known superior way to fulfill this need.

    We agree that if there was a (let's call it) Goodbook that could do these things without the evil parts of Facebook, we would all be best served to switch there, but that such a site doesn't really exist. Twitter is very good at some parts of it (threading has gotten worse, and is non-intuitive right now, but it actually works once you understand it, although in turn I agree that in general 'once you get used to it' is mostly an argument you make when you're losing, interface-wise), but the privacy thing is a dealbreaker for some tasks.

    Thus, for some things, Facebook is still the only game in town, despite being evil, and maybe you need to compromise with sin. I respect the decision to use FB purely for casual, ephemeral, private things, given the circumstances. Others have more need of this than I do.

    I think a lot of this disagreement is that people seem to think that Facebook's privacy settings will 'stick' in an important sense - that what you post there is actually private. I don't buy it. You are posting to hundreds or thousands of people. In my mind, that's basically public. I do understand that there's a norm that it's NOT public, and there are trivial barriers not to be underestimated, but I would not post anything to Facebook (other than in small secret groups) I would not be OK with being in a public forum. Small secret groups I would think of as similar to email, which is 'this isn't public exactly but the government is reading this if it wants to, and there's a good chance it will come out eventually in exactly the cases you really don't want it to, so watch what you say in writing even to people you completely trust because nothing is ever reliably deleted.'

    1. Two things make Facebook a little more private for me. One is that I only comment on posts of my friends' that are listed as "Friends Only," which mostly prevents people I don't want to see from seeing my comments. I don't actually care much about people I don't know seeing them, I care much more about people I *do* know but who *aren't* in the intersection of my friends list with others' friends lists (e.g. family members, former teachers, people I went to school with).

      The second thing is that I post to Facebook mostly on a curated list of friends, not everyone I'm 'friends' with. This has improved my ability to use it by quite a bit.

      But yeah, Facebook is not as private as it wants you to think it is. This is a problem that ideally "Goodbook" would solve.