The only reason we use Facebook is because everyone else is using Facebook, so it’s convenient to post events there because you know there’s a good chance that everyone will be able to see the event once it’s posted.
Exactly, it’s good enough.
I take your point that any feature could be better in any number of ways, but as a package Facebook is empirically good enough or else we would use something else. Put another way, desire all the ways it could be better, people keep making the choice to use Facebook when there are hundreds of other tools available. People already being on Facebook is just one of the assessment criteria for picking a tool to use, albeit a large one.
My argument is not so much that we can’t in theory do better than Facebook, but that we don’t because it’s a large bundle of product and service along a lot of different dimensions that results in people using it because nothing does better along enough dimensions, including the switching cost dimension, such that the final analysis of combining the vectors produces a score better than Facebook’s.
No, I don’t think that follows. One of the lessons of Inadequate Equilibria is that both individuals and groups get stuck in situations that are not “good enough” by any measure, simply because no individual has enough “free energy” to force a change to something better.
To me, Facebook is yet another Inadequate Equilibrium. We don’t use it because it’s good, we use it because that’s where we started out, and no one has enough pull in our community to force a global change. If someone manages to change it at a global level in the rationalist community, we would look back upon our usage of Facebook and wonder why we ever bothered to use such a terrible tool.
The relevant question here is “is there _currently_ an alternative, such that if everyone switched from FB to another thing at the same time, they’d all be better off?”. The context of the OP was “we built a tool by which people could all agree to switch from FB if and only if a bunch of other people switched from FB at once – why didn’t that work?”
And to that, I think there’s actually a pretty good case to be made that “no, there is not a good enough alternative to FB.” FB does a lot of things well – it can afford to because it’s a huge company and spends lots on R&D and fine-tuning itself.
The parts of FB that are bad are due to how it’s monetized and what incentives it has. Any alternative to FB would have to be monetized differently, or have to be (to some degree) altruistic project, or you’d have to have people making peace with a less polished project that didn’t have as much developer resources.
This doesn’t mean it can’t be done. (I think discord is actually monetized in a pretty reasonable way, where you get access to cooler emoji by paying $5/month while otherwise being free). But someone has to actually build it, and it requires a lot of effort.
I’ve looked at the attempts to build “open source aesthetic facebook” (that’s built from the ground up to leave the power with users rather than the monetization system). And… it was quite bad. Google+ might have been better, but I have no reason to assume Google+ would be any less evil than FB over the long term. (I think they couldn’t quite get away with being as evil initially since they had to tempt people away from FB).
To be honest, I think that if we start talking about “an alternative to Facebook” or “switching from Facebook to some other thing”, or “building a Facebook replacement”, we’re off to a bad start.
An analogy: suppose you have a Swiss army knife, which you use to chop vegetables, saw through wooden beams, trim your nails, etc. I observe that said tool is really quite sub-par. You ask me what I think is an alternative tool you should use. Should you switch from the Swiss army knife to some other thing? Do I have a Swiss army knife replacement to offer you?
No, of course not. It’s just that you should chop vegetables with a chef’s knife, saw through wooden beams with a saw, trim your nails with a nail trimmer, etc.
Now suppose I said this to you, and you replied: “A chef’s knife isn’t a Swiss army knife replacement! It doesn’t saw wood and it doesn’t trim my nails. A saw isn’t a Swiss army knife replacement! It doesn’t chop vegetables and it doesn’t trim my nails. A nail trimmer isn’t …”
You’d be missing the point, right?
FB does a lot of things well
Well… not really! That’s the thing; Facebook doesn’t really do a lot of things well. It does a lot of things, and it does some things well (probably; I have my doubts, but let’s grant it).
The better question, IMO (instead of those I list above) is: for each X, where X is something you use Facebook for, what should you use?
For some X, the answer will be “Facebook”.
For some other X, the answer will be “not Facebook; instead, [thing]”. The value of [thing] will not be the same for all X.
I do think the basic analogy you’re pointing at here is important to consider. Maybe FB should be split into it’s component parts. (And to be clear, even if we need a swiss army knife I think FB is quite subpar as a swiss army knife and someone should be making a better one, with the caveat that they need to actually address the incentives problem in a way that makes sense and I expect it to be a lot of work)
But much of FB’s value prop lies in having access to all of it’s userbase, to do various things with everyone I know, and have those things work in tandem. I can communicate with large or small groups that I want to communicate with, and invite them to events, and seamlessly blend between the two. Creating a userbase is hard, so there’s a strong benefit to swiss-army-ness.
Epistemic Status: Increasingly strained metaphor
I think FB is less like a crappy swiss army knife that doesn’t work properly, and more like a kitchen that doesn’t work. (Or rather, everything works, but the fridge and cabinets automatically restock themselves, and charges me when I take things out of it.… but they periodically move things around in the fridge to force me to bump into new things to take out and get charged for so I can impulse buy different things. (Supermarkets actually essentially do do this and it’s kinda annoying but less intrusive because it’s not my home that they’re re-arranging)).
This feature is undesirable to me and I want to fix it… and I could buy a new fridge, and a new cabinet, and all my own vegetables, etc. But buying any individual one of these means it has to live in a different kitchen that doesn’t come with all the rest of my stuff. I have to buy all the things in order for any of them to be nearly as useful. Which is quite expensive. [Also, I’d have to pay for them all (whereas the Auto-Wonder-Kitchen was free because it was running on weird product placement).]
Google has a similar set of products as FB, and does keep them a bit more distinct from each other. (i.e. you can use email to communicate with people, and calendar to invite them to events, and you can just use email and calendar in a way where you can’t just use any given segment of FB except for Messenger). And you get to re-use the same set of contacts in between google apps. But that segregation already has some issues (it’s not quit
But what Google doesn’t do (as well as FB) is basically provide ways to spam all your friends with stuff they might or might not be interested in, and make it easy-ish for them to opt into which collection of spam they want. (Google attempts to do this but has never really succeeded IMO). When I say I want a FB replacement, “spam almost exactly the right people” is basically the feature I want.
So, there’s definitely a lot to engage with, here, and I don’t think I could do justice to all of what you say in a single comment (in particular, I have a lot to say about the “spam all your friends with stuff they might or might not be interested in” thing[1]). Some scattered thoughts follow, with possibly more to come later…
… I could buy a new fridge, and a new cabinet, and all my own vegetables, etc. But buying any individual one of these means it has to live in a different kitchen that doesn’t come with all the rest of my stuff. I have to buy all the things in order for any of them to be nearly as useful. Which is quite expensive.
So, in the world of your kitchen analogy, that does seem true. But in the real world, and talking about Facebook, it’s actually not true!
Here’s an example. Eliezer has, for some years now, been “blogging”, so to speak, on Facebook. This has always perplexed me. Oh, I can understand the sentiment of “access to all my friends” that you mention, sure. (I may disagree, but I understand the argument, even if I find it ultimately unconvincing.)
But that is an argument for posting things on Facebook, it is not an argument for hosting things on Facebook!
You see? Eliezer could have a blog! He could have a Wordpress blog (wordpress.com or self-hosted), or use any number of other blogging engines; and then, he could link or cross-post or “share” those blog posts on Facebook. (I trust I do not need to go over the technical, accessibility, openness, archival, control, and many other advantages of doing so.)
So here is a potential “move away from Facebook” that does not require a “different kitchen”, etc. It’s fully interoperable with the existing stuff; it offers great advantages; and it facilitates a future “total abandonment” of Facebook (for this purpose), if such should be desired—and just as easy a “return” to Facebook (for this purpose), if that should be desired.
So why not do this? What is the reason? It can’t be technical, or a matter of competence, or effort; it is utterly inconceivable that—even if Eliezer himself can’t find the time, or doesn’t know how, to set up a blog—he cannot find someone competent and willing to do this for him. (I’d do it for free, and so would any number of other people!)
What’s the obstacle?
And to be clear, even if we need a swiss army knife I think FB is quite subpar as a swiss army knife and someone should be making a better one
It’s worth thinking about this analogy a bit more.
Consider this, for example: when, and why, do people need Swiss army knives? Actual ones, I mean. Certainly not for everyday uses, like “chop your vegetables when making dinner every night”, etc. Victorinox clearly does not intend their products to replace chef’s knives, yes?
But—speaking, still, of actual Swiss army knives—people do buy them, and use them. They are genuinely useful tools. Just… not as replacements for the proper tools for any given job, when those are available, and can be used.
So perhaps it’s worth having, say, an ability to use Facebook to invite people to an event, or the ability to post “blog posts” on Facebook, if we expect to encounter situations where the proper tools for those jobs are unavailable, or we find ourselves unable to use them. On the other hand, it seems like a clearly bad idea to plan on using Facebook as the primary tool for very job, forever, even when far superior tools exist, and we are perfectly capable of using those proper tools—just as it would be foolish, when stocking your home kitchen, to skip buying a chef’s knife, on account of you already having your Swiss army knife.
[1] And not all of it is “that goal is actually bad”! (Only some of it.)
As I said in my earlier comment, even if you want the sort of thing Eliezer says he wants, that’s entirely insufficient reason to make the choice he made. One would also have to not want all of the many advantages of choosing otherwise (i.e., posting on not-Facebook).
Let me emphasize again: for Eliezer’s choice to make sense, it is not enough to want the things he said. Neither does it suffice to want them more strongly than you want the other stuff I talked about, because it is not a tradeoff—at least, not in the sense that we have to trade off desirable features or advantages of one system against those of another system. The only costs are the up-front cognitive costs of considering the alternatives.
In order for Eliezer’s choice to make sense, one of the following seems to need to be true:
He has to not care at all about any of the advantages of not using Facebook as a primary host for one’s content.
He has to not have given any thought to the matter.
I do not have sufficient information to discriminate between these two possibilities. But I will say that #1 reflects much, much more poorly on Eliezer’s character than #2 does—which is why I assumed the latter to be the true answer.
There’s one major reason a blog will never be strictly superior to facebook: the dropoff rate for people clicking through to read your blog is large. (For example, I generally do not click through to read blogposts while scanning facebook – I only read whatever quotes people include on FB itself, and this is quite common).
People are not enacting the algorithm “scan FB, then click through to things and read them.” People are enacting the algorithm “scan and comment on FB.” The trivial inconvenience is huge.
If I want to comment on Eliezer’s blog, it’s worse: I have to sign up on that blog. And I also don’t trust most people people to have clicked through to read the comments on that blog because I know most people are only reading comments on Facebook. This happens quite frequently with rationalist blogs: all the substantive discussion ends up happening on the corresponding FB post, not on the blog itself.
(It’ll be necessary to sign up on the blog, since Eliezer specifically wants to be able to block people easily)
It’s noteworthy that Eliezer has multiple other blogs (OvercomingBias, LessWrong, and his own yudwosky.net, and a tumblr). He uses FB. I think it’s quite unlikely this is because he accidentally ended up on FB.
There’s one additional important feature FB has that most other blogs do not: casual feel. the textbox for entering things is small, the fontsize is small, the font is sans-serif and simple, the site is cluttered so your post doesn’t look like the most important thing on the page. All of these translate into a strong signal of “this is low effort and is not to be judged as a high effort thing.”
This is extremely important – it means I can actually get things written that I wouldn’t otherwise write. It’s hard to do with a dedicated blogsite (there may be wordpress themes that produce this effect)
There’s one major reason a blog will never be strictly superior to facebook: the dropoff rate for people clicking through to read your blog is large.
Cross-posting full text eliminates this disadvantage.
If I want to comment on Eliezer’s blog, it’s worse: I have to sign up on that blog. … (It’ll be necessary to sign up on the blog, since Eliezer specifically wants to be able to block people easily)
This is not true. Simply set up Facebook login for the blog; set that as the only available login method; and then block whoever you want.
Even if it were true, I did not say it was necessary to enable commenting on the blog; let people comment on Facebook only, if you like.
This happens quite frequently with rationalist blogs: all the substantive discussion ends up happening on the corresponding FB post, not on the blog itself.
While I do think this is bad, it does nothing whatever to reduce the advantages of posting the content itself on the blog.
It is also possible (and not very difficult at all, though not turnkey) to enable Facebook comments associated with a Facebook-shared blog post to also appear under your blog—thus letting people who comment on your blog, also see comments that other people have posted on Facebook.
[Eliezer] uses FB. I think it’s quite unlikely this is because he accidentally ended up on FB.
I agree that it’s unlikely that Eliezer accidentally ended up on Facebook. But I think it’s quite a bit more likely that Eliezer thoughtlessly ended up on Facebook—because he did not think about the tradeoffs, because he did not care to consider the pros and cons, because he made no attempt to search for better solutions. (And because no one else knew or cared enough to point them out to him.)
There’s one additional important feature FB has that most other blogs do not: casual feel. The textbox for entering things is small, the fontsize is small, the font is sans-serif and simple, the site is cluttered so your post doesn’t look like the most important thing on the page. All of these translate into a strong signal of “this is low effort and is not to be judged as a high effort thing.”
I am very skeptical of this sort of claim, truth be told. It feels very much like a post-hoc, just-so-story, sort of explanation. I do not expect the majority of such claims to survive rigorous testing.
But even if we accept what you say here, it is obviously easy enough to duplicate that effect by selecting an appropriate theme/layout for a blog. Once again, Eliezer Yudkowsky, of all people, is uniquely positioned to call upon a great quantity of technical and design talent to create bespoke solutions for him, for problems of this nature.
… Facebook is empirically good enough or else we would use something else.
This reasoning is not sound.
… people keep making the choice to use Facebook when there are hundreds of other tools available.
No, they don’t.
The second of the two quotes above is wrong in at least two ways—a specific one and a general one. The general one first:
Whatever may be claimed by adherents of certain strains of philosophy, there is a profound difference between commission and omission. Specifically, there is a great difference between these two situations:
You are presented with a choice: A or B? You must select one.
You will be given A. If you wish, you may request B instead.
Suppose we look at a population of people who have encountered this situation, and observe that most of them have A. Would it be reasonable for us to say, “These people have made the choice to have A”? No. Absolutely not.
And, separately and additionally, there is a great difference between these two situations:
You have A. You are presented with an explicit choice: keep A, or get B instead? (There is no default action; or, perhaps, if you make no selection, A is taken from you and you are left with nothing.)
You have A. If you take no action, you will continue to have A. At any time of your choosing, you may elect to switch to B.
Once again, suppose we look at a population of people who have A. We know that, at a time of their choosing, any of them could switch to B. Would it be reasonable for us to say, “These people have made the choice to have A”? Once again: no, absolutely not.
Note that neither of these distinctions is identical with the familiarity effect[1], nor do either of them have anything to do with external (as distinct from cognitive a.k.a. “internal”) costs of switching! Both of those are separate points.
So that is the general reason why saying “people keep making the choice to use Facebook” is dangerously misleading: it ignores how human cognition works—the reality of how people think, choose, and act.
I will describe the specific reason in a sibling comment…
[1] Consider a scenario where you always use some particular brand of toothpaste. Every time you run out, you go to the store, select a brand of toothpaste from the many brands the store carries, and purchase it. There is no default action (you can’t just tell the clerk “give me toothpaste”, you have to personally make the choice and personally take the item, etc.), and you are required to actively choose—if you take no action, you will have no toothpaste. Thus neither of the two considerations described by the distinctions I list apply. Nevertheless, we often buy the same brand of toothpaste, over and over. This is the familiarity effect. Once again, this is entirely different from the “cognitive costs of choice” effects which I describe above.
Once again, suppose we look at a population of people who have A. We know that, at a time of their choosing, any of them could switch to B. Would it be reasonable for us to say, “These people have made the choice to have A”? Once again: no, absolutely not.
By why is A in this case “use Facebook”? No one is making anyone use Facebook, people choose to use it, so it’s not really a default except that in some particular framing you may think of it as a default, which may be how some people think of the choice but is not how I think the choice would look to someone who was not already a committed Facebook user (the outside view is different from the inside view, and I’m taking the outside view). I feel like your entire argument is hinged on a hidden assumption about a framing that places Facebook as a default choice. That’s fine as far as it goes but I also feel it fails to address the point I’m making by picking a different frame of reference, namely one that is aiming to be more outside.
Also, as a meta note I’m establishing a new Said-specific policy for myself of only responding once to Said threads for the foreseeable future: I dislike the confrontational tone (that I perceive that) you take so I’ll only respond at most once to you in a thread to avoid getting into a back and forth I dislike. Anyone who notices me doing otherwise please feel free to remind me I said this.
By why is A in this case “use Facebook”? No one is making anyone use Facebook, people choose to use it,
Because the question is “why are people continuing to use Facebook”, not “why did they start using Facebook”. It doesn’t matter why they started. They did; this is now a brute fact. But why do they continue? And so “use Facebook” is of course the default choice, the non-action.
it’s not really a default except that in some particular framing you may think of it as a default
No, it’s a default in the obvious and straightforward way of “a person who is already using Facebook, by default, continues to use Facebook”. Switching takes effort (cognitive effort, time, etc.). Heck, it takes even knowing that you can switch, not to mention knowing what to switch to, and how. Continuing to use Facebook takes no effort and no thought.
I feel like your entire argument is hinged on a hidden assumption about a framing that places Facebook as a default choice.
It’s not hidden at all, it’s not an “assumption”, and it’s not a “framing”; it’s just how things are.
That’s fine as far as it goes but I also feel it fails to address the point I’m making by picking a different frame of reference, namely one that is aiming to be more outside.
No amount of framing changes the facts of this matter.
… Facebook is empirically good enough or else we would use something else.
I have heard this sort of thing many times, about a number of things (Facebook is certainly one of the usual suspects, though far from the only one). It’s worth examining this view closely.
Let me say, first, that quanticle’s point about Facebook potentially being an inadequate equilibrium is certainly a good one. But I want to take a somewhat different approach.
What do we mean when we say that Facebook is “good enough”?
Well, perhaps what we mean is: Facebook is good enough for us to continue using it—or, equivalently (?), not bad enough for us to stop using it. What we’re saying, then, is: if Facebook were bad enough for us to stop using it, then we would stop using it. Since we’re still using it, it must be good enough for us to keep using it.
This is clearly tautological. Under this interpretation, “Facebook is empirically good enough or else we would use something else” simply reduces to the observation that we are, in fact, still using Facebook. Which we already knew. So we’ve concluded a grand total of nothing.
The assumptions behind this perspective seem to be as follows. There is some dimension of “goodness”, along which a product or service may be “good” or “bad” (rather, of course, we should say “better” or “worse”). There is some point along this dimension such that at or below that point, the product/service is so bad that its users abandon it. If it is less bad than that, then its users continue using it.
Well, fair enough. We might quibble that there is more than one dimension of “goodness” (that “goodness” is multidimensional); we might quibble that “goodness” for one sort of person is not the same as “goodness” for another sort of person; etc. But let us take as an axiom, for now, that there is something sufficiently close to a single dimension of goodness (perhaps a vector through multidimensional goodness-space; perhaps the average of different people’s such vectors… or something; the details needn’t matter). Let us assume that we can, roughly, identify and agree upon this dimension.
And yet, having taken this axiom, we are actually still adrift, as far as having any kind of satisfying account of “goodness” or “good enough”, because we face two key questions:
What is the nature of this “goodness” dimension? Just what exactly makes a product or service in this category “good” or “bad”, “better” or “worse”? How, exactly—other than “do its users abandon it”—do “good” members of this category differ from “bad” ones?
Do we care about any other points or ranges along this dimension other than the “tipping point” of “bad enough to be abandoned by its users”? Is “is this product/service literally too terrible to keep using, yes/no?” the only question (as pertains to “position along the ‘goodness’ dimension”) that we have reason to ask of a product/service? Supposing we ascertain for ourselves that a product/service is just good enough for us to use—have we now no further reason at all to concern ourselves with its “goodness”?
The answer to both questions is actually the same—call it question A:
Question A: “Is this product or service accomplishing my goals / satisfying my needs / etc.?”
Now, might you argue that, actually, this is just the same thing as “is this good enough to not stop using it”?
You can argue that. But that argument would be absurd.
For one thing, you would have to argue that the answer to question A is binary: yes, this product or service is accomplishing my goals; or, no, this product or service is not accomplishing my goals. But of course it’s not binary; your goals might be accomplished only partly, or not as well, or only some of your goals might be accomplished, etc.
For another, you would have to argue that, empirically, if a product or service fails to accomplish a person’s goals, or fails to satisfy their needs, or fails to do these things well, then that person stops using said product or service. Of course no such thing is even remotely true.
So when you say “we are still using Facebook, therefore Facebook is good enough”—meaning, of course, nothing more than “Facebook is not quite so bad as to be too terrible to keep using”—then my question to you is this:
So what?
Who cares? Why should that be significant? OK, so Facebook is just “good enough” that we’re not motivated to abandon it—fine; we knew this already (it follows, tautologically, from us not, in fact, having abandoned it). So what? Does it accomplish our goals? Does it satisfy our needs? How well does it do these things? Is it good enough to make continuing to use it be the optimal strategy? By using it, do we win? Might we win more if we abandoned it, and switched to something else? These are the questions that matter!
P.S.: When I see people arguing against attempts to motivate abandonment of Facebook (or similar attempts to improve other inadequate equilibria like this) on the basis of “it’s good enough, or else we wouldn’t be using it”, I cannot help but think of this recent post.
Exactly, it’s good enough.
I take your point that any feature could be better in any number of ways, but as a package Facebook is empirically good enough or else we would use something else. Put another way, desire all the ways it could be better, people keep making the choice to use Facebook when there are hundreds of other tools available. People already being on Facebook is just one of the assessment criteria for picking a tool to use, albeit a large one.
My argument is not so much that we can’t in theory do better than Facebook, but that we don’t because it’s a large bundle of product and service along a lot of different dimensions that results in people using it because nothing does better along enough dimensions, including the switching cost dimension, such that the final analysis of combining the vectors produces a score better than Facebook’s.
No, I don’t think that follows. One of the lessons of Inadequate Equilibria is that both individuals and groups get stuck in situations that are not “good enough” by any measure, simply because no individual has enough “free energy” to force a change to something better.
By this logic, p-values are “good enough” statistics, because the scientific community persists in using them, even though they clearly lead to non-reproducible studies. Selling lifesaving formula in one state and making it illegal to sell across state lines is “good enough” because no one has bothered to change the situation.
To me, Facebook is yet another Inadequate Equilibrium. We don’t use it because it’s good, we use it because that’s where we started out, and no one has enough pull in our community to force a global change. If someone manages to change it at a global level in the rationalist community, we would look back upon our usage of Facebook and wonder why we ever bothered to use such a terrible tool.
The relevant question here is “is there _currently_ an alternative, such that if everyone switched from FB to another thing at the same time, they’d all be better off?”. The context of the OP was “we built a tool by which people could all agree to switch from FB if and only if a bunch of other people switched from FB at once – why didn’t that work?”
And to that, I think there’s actually a pretty good case to be made that “no, there is not a good enough alternative to FB.” FB does a lot of things well – it can afford to because it’s a huge company and spends lots on R&D and fine-tuning itself.
The parts of FB that are bad are due to how it’s monetized and what incentives it has. Any alternative to FB would have to be monetized differently, or have to be (to some degree) altruistic project, or you’d have to have people making peace with a less polished project that didn’t have as much developer resources.
This doesn’t mean it can’t be done. (I think discord is actually monetized in a pretty reasonable way, where you get access to cooler emoji by paying $5/month while otherwise being free). But someone has to actually build it, and it requires a lot of effort.
I’ve looked at the attempts to build “open source aesthetic facebook” (that’s built from the ground up to leave the power with users rather than the monetization system). And… it was quite bad. Google+ might have been better, but I have no reason to assume Google+ would be any less evil than FB over the long term. (I think they couldn’t quite get away with being as evil initially since they had to tempt people away from FB).
To be honest, I think that if we start talking about “an alternative to Facebook” or “switching from Facebook to some other thing”, or “building a Facebook replacement”, we’re off to a bad start.
An analogy: suppose you have a Swiss army knife, which you use to chop vegetables, saw through wooden beams, trim your nails, etc. I observe that said tool is really quite sub-par. You ask me what I think is an alternative tool you should use. Should you switch from the Swiss army knife to some other thing? Do I have a Swiss army knife replacement to offer you?
No, of course not. It’s just that you should chop vegetables with a chef’s knife, saw through wooden beams with a saw, trim your nails with a nail trimmer, etc.
Now suppose I said this to you, and you replied: “A chef’s knife isn’t a Swiss army knife replacement! It doesn’t saw wood and it doesn’t trim my nails. A saw isn’t a Swiss army knife replacement! It doesn’t chop vegetables and it doesn’t trim my nails. A nail trimmer isn’t …”
You’d be missing the point, right?
Well… not really! That’s the thing; Facebook doesn’t really do a lot of things well. It does a lot of things, and it does some things well (probably; I have my doubts, but let’s grant it).
The better question, IMO (instead of those I list above) is: for each X, where X is something you use Facebook for, what should you use?
For some X, the answer will be “Facebook”.
For some other X, the answer will be “not Facebook; instead, [thing]”. The value of [thing] will not be the same for all X.
I do think the basic analogy you’re pointing at here is important to consider. Maybe FB should be split into it’s component parts. (And to be clear, even if we need a swiss army knife I think FB is quite subpar as a swiss army knife and someone should be making a better one, with the caveat that they need to actually address the incentives problem in a way that makes sense and I expect it to be a lot of work)
But much of FB’s value prop lies in having access to all of it’s userbase, to do various things with everyone I know, and have those things work in tandem. I can communicate with large or small groups that I want to communicate with, and invite them to events, and seamlessly blend between the two. Creating a userbase is hard, so there’s a strong benefit to swiss-army-ness.
Epistemic Status: Increasingly strained metaphor
I think FB is less like a crappy swiss army knife that doesn’t work properly, and more like a kitchen that doesn’t work. (Or rather, everything works, but the fridge and cabinets automatically restock themselves, and charges me when I take things out of it.… but they periodically move things around in the fridge to force me to bump into new things to take out and get charged for so I can impulse buy different things. (Supermarkets actually essentially do do this and it’s kinda annoying but less intrusive because it’s not my home that they’re re-arranging)).
This feature is undesirable to me and I want to fix it… and I could buy a new fridge, and a new cabinet, and all my own vegetables, etc. But buying any individual one of these means it has to live in a different kitchen that doesn’t come with all the rest of my stuff. I have to buy all the things in order for any of them to be nearly as useful. Which is quite expensive. [Also, I’d have to pay for them all (whereas the Auto-Wonder-Kitchen was free because it was running on weird product placement).]
Google has a similar set of products as FB, and does keep them a bit more distinct from each other. (i.e. you can use email to communicate with people, and calendar to invite them to events, and you can just use email and calendar in a way where you can’t just use any given segment of FB except for Messenger). And you get to re-use the same set of contacts in between google apps. But that segregation already has some issues (it’s not quit
But what Google doesn’t do (as well as FB) is basically provide ways to spam all your friends with stuff they might or might not be interested in, and make it easy-ish for them to opt into which collection of spam they want. (Google attempts to do this but has never really succeeded IMO). When I say I want a FB replacement, “spam almost exactly the right people” is basically the feature I want.
So, there’s definitely a lot to engage with, here, and I don’t think I could do justice to all of what you say in a single comment (in particular, I have a lot to say about the “spam all your friends with stuff they might or might not be interested in” thing[1]). Some scattered thoughts follow, with possibly more to come later…
So, in the world of your kitchen analogy, that does seem true. But in the real world, and talking about Facebook, it’s actually not true!
Here’s an example. Eliezer has, for some years now, been “blogging”, so to speak, on Facebook. This has always perplexed me. Oh, I can understand the sentiment of “access to all my friends” that you mention, sure. (I may disagree, but I understand the argument, even if I find it ultimately unconvincing.)
But that is an argument for posting things on Facebook, it is not an argument for hosting things on Facebook!
You see? Eliezer could have a blog! He could have a Wordpress blog (wordpress.com or self-hosted), or use any number of other blogging engines; and then, he could link or cross-post or “share” those blog posts on Facebook. (I trust I do not need to go over the technical, accessibility, openness, archival, control, and many other advantages of doing so.)
So here is a potential “move away from Facebook” that does not require a “different kitchen”, etc. It’s fully interoperable with the existing stuff; it offers great advantages; and it facilitates a future “total abandonment” of Facebook (for this purpose), if such should be desired—and just as easy a “return” to Facebook (for this purpose), if that should be desired.
So why not do this? What is the reason? It can’t be technical, or a matter of competence, or effort; it is utterly inconceivable that—even if Eliezer himself can’t find the time, or doesn’t know how, to set up a blog—he cannot find someone competent and willing to do this for him. (I’d do it for free, and so would any number of other people!)
What’s the obstacle?
It’s worth thinking about this analogy a bit more.
Consider this, for example: when, and why, do people need Swiss army knives? Actual ones, I mean. Certainly not for everyday uses, like “chop your vegetables when making dinner every night”, etc. Victorinox clearly does not intend their products to replace chef’s knives, yes?
But—speaking, still, of actual Swiss army knives—people do buy them, and use them. They are genuinely useful tools. Just… not as replacements for the proper tools for any given job, when those are available, and can be used.
So perhaps it’s worth having, say, an ability to use Facebook to invite people to an event, or the ability to post “blog posts” on Facebook, if we expect to encounter situations where the proper tools for those jobs are unavailable, or we find ourselves unable to use them. On the other hand, it seems like a clearly bad idea to plan on using Facebook as the primary tool for very job, forever, even when far superior tools exist, and we are perfectly capable of using those proper tools—just as it would be foolish, when stocking your home kitchen, to skip buying a chef’s knife, on account of you already having your Swiss army knife.
[1] And not all of it is “that goal is actually bad”! (Only some of it.)
Eliezer has mentioned his reasons for moving to Facebook.
Having read said comment [working link], I stand behind my “thoughtlessly” hypothesis, which seems to have been entirely on-target.
Um, huh? That comments looks to me like the “Eliezer wants a very different thing than Said wants” hypothesis.
As I said in my earlier comment, even if you want the sort of thing Eliezer says he wants, that’s entirely insufficient reason to make the choice he made. One would also have to not want all of the many advantages of choosing otherwise (i.e., posting on not-Facebook).
Let me emphasize again: for Eliezer’s choice to make sense, it is not enough to want the things he said. Neither does it suffice to want them more strongly than you want the other stuff I talked about, because it is not a tradeoff—at least, not in the sense that we have to trade off desirable features or advantages of one system against those of another system. The only costs are the up-front cognitive costs of considering the alternatives.
In order for Eliezer’s choice to make sense, one of the following seems to need to be true:
He has to not care at all about any of the advantages of not using Facebook as a primary host for one’s content.
He has to not have given any thought to the matter.
I do not have sufficient information to discriminate between these two possibilities. But I will say that #1 reflects much, much more poorly on Eliezer’s character than #2 does—which is why I assumed the latter to be the true answer.
There’s one major reason a blog will never be strictly superior to facebook: the dropoff rate for people clicking through to read your blog is large. (For example, I generally do not click through to read blogposts while scanning facebook – I only read whatever quotes people include on FB itself, and this is quite common).
People are not enacting the algorithm “scan FB, then click through to things and read them.” People are enacting the algorithm “scan and comment on FB.” The trivial inconvenience is huge.
If I want to comment on Eliezer’s blog, it’s worse: I have to sign up on that blog. And I also don’t trust most people people to have clicked through to read the comments on that blog because I know most people are only reading comments on Facebook. This happens quite frequently with rationalist blogs: all the substantive discussion ends up happening on the corresponding FB post, not on the blog itself.
(It’ll be necessary to sign up on the blog, since Eliezer specifically wants to be able to block people easily)
It’s noteworthy that Eliezer has multiple other blogs (OvercomingBias, LessWrong, and his own yudwosky.net, and a tumblr). He uses FB. I think it’s quite unlikely this is because he accidentally ended up on FB.
There’s one additional important feature FB has that most other blogs do not: casual feel. the textbox for entering things is small, the fontsize is small, the font is sans-serif and simple, the site is cluttered so your post doesn’t look like the most important thing on the page. All of these translate into a strong signal of “this is low effort and is not to be judged as a high effort thing.”
This is extremely important – it means I can actually get things written that I wouldn’t otherwise write. It’s hard to do with a dedicated blogsite (there may be wordpress themes that produce this effect)
Cross-posting full text eliminates this disadvantage.
This is not true. Simply set up Facebook login for the blog; set that as the only available login method; and then block whoever you want.
Even if it were true, I did not say it was necessary to enable commenting on the blog; let people comment on Facebook only, if you like.
While I do think this is bad, it does nothing whatever to reduce the advantages of posting the content itself on the blog.
It is also possible (and not very difficult at all, though not turnkey) to enable Facebook comments associated with a Facebook-shared blog post to also appear under your blog—thus letting people who comment on your blog, also see comments that other people have posted on Facebook.
I agree that it’s unlikely that Eliezer accidentally ended up on Facebook. But I think it’s quite a bit more likely that Eliezer thoughtlessly ended up on Facebook—because he did not think about the tradeoffs, because he did not care to consider the pros and cons, because he made no attempt to search for better solutions. (And because no one else knew or cared enough to point them out to him.)
I am very skeptical of this sort of claim, truth be told. It feels very much like a post-hoc, just-so-story, sort of explanation. I do not expect the majority of such claims to survive rigorous testing.
But even if we accept what you say here, it is obviously easy enough to duplicate that effect by selecting an appropriate theme/layout for a blog. Once again, Eliezer Yudkowsky, of all people, is uniquely positioned to call upon a great quantity of technical and design talent to create bespoke solutions for him, for problems of this nature.
By the way:
Something seems to be missing here?
This reasoning is not sound.
No, they don’t.
The second of the two quotes above is wrong in at least two ways—a specific one and a general one. The general one first:
Whatever may be claimed by adherents of certain strains of philosophy, there is a profound difference between commission and omission. Specifically, there is a great difference between these two situations:
You are presented with a choice: A or B? You must select one.
You will be given A. If you wish, you may request B instead.
Suppose we look at a population of people who have encountered this situation, and observe that most of them have A. Would it be reasonable for us to say, “These people have made the choice to have A”? No. Absolutely not.
And, separately and additionally, there is a great difference between these two situations:
You have A. You are presented with an explicit choice: keep A, or get B instead? (There is no default action; or, perhaps, if you make no selection, A is taken from you and you are left with nothing.)
You have A. If you take no action, you will continue to have A. At any time of your choosing, you may elect to switch to B.
Once again, suppose we look at a population of people who have A. We know that, at a time of their choosing, any of them could switch to B. Would it be reasonable for us to say, “These people have made the choice to have A”? Once again: no, absolutely not.
Note that neither of these distinctions is identical with the familiarity effect[1], nor do either of them have anything to do with external (as distinct from cognitive a.k.a. “internal”) costs of switching! Both of those are separate points.
So that is the general reason why saying “people keep making the choice to use Facebook” is dangerously misleading: it ignores how human cognition works—the reality of how people think, choose, and act.
I will describe the specific reason in a sibling comment…
[1] Consider a scenario where you always use some particular brand of toothpaste. Every time you run out, you go to the store, select a brand of toothpaste from the many brands the store carries, and purchase it. There is no default action (you can’t just tell the clerk “give me toothpaste”, you have to personally make the choice and personally take the item, etc.), and you are required to actively choose—if you take no action, you will have no toothpaste. Thus neither of the two considerations described by the distinctions I list apply. Nevertheless, we often buy the same brand of toothpaste, over and over. This is the familiarity effect. Once again, this is entirely different from the “cognitive costs of choice” effects which I describe above.
By why is A in this case “use Facebook”? No one is making anyone use Facebook, people choose to use it, so it’s not really a default except that in some particular framing you may think of it as a default, which may be how some people think of the choice but is not how I think the choice would look to someone who was not already a committed Facebook user (the outside view is different from the inside view, and I’m taking the outside view). I feel like your entire argument is hinged on a hidden assumption about a framing that places Facebook as a default choice. That’s fine as far as it goes but I also feel it fails to address the point I’m making by picking a different frame of reference, namely one that is aiming to be more outside.
Also, as a meta note I’m establishing a new Said-specific policy for myself of only responding once to Said threads for the foreseeable future: I dislike the confrontational tone (that I perceive that) you take so I’ll only respond at most once to you in a thread to avoid getting into a back and forth I dislike. Anyone who notices me doing otherwise please feel free to remind me I said this.
Because the question is “why are people continuing to use Facebook”, not “why did they start using Facebook”. It doesn’t matter why they started. They did; this is now a brute fact. But why do they continue? And so “use Facebook” is of course the default choice, the non-action.
No, it’s a default in the obvious and straightforward way of “a person who is already using Facebook, by default, continues to use Facebook”. Switching takes effort (cognitive effort, time, etc.). Heck, it takes even knowing that you can switch, not to mention knowing what to switch to, and how. Continuing to use Facebook takes no effort and no thought.
It’s not hidden at all, it’s not an “assumption”, and it’s not a “framing”; it’s just how things are.
No amount of framing changes the facts of this matter.
(Continuation of my previous comment…)
I have heard this sort of thing many times, about a number of things (Facebook is certainly one of the usual suspects, though far from the only one). It’s worth examining this view closely.
Let me say, first, that quanticle’s point about Facebook potentially being an inadequate equilibrium is certainly a good one. But I want to take a somewhat different approach.
What do we mean when we say that Facebook is “good enough”?
Well, perhaps what we mean is: Facebook is good enough for us to continue using it—or, equivalently (?), not bad enough for us to stop using it. What we’re saying, then, is: if Facebook were bad enough for us to stop using it, then we would stop using it. Since we’re still using it, it must be good enough for us to keep using it.
This is clearly tautological. Under this interpretation, “Facebook is empirically good enough or else we would use something else” simply reduces to the observation that we are, in fact, still using Facebook. Which we already knew. So we’ve concluded a grand total of nothing.
The assumptions behind this perspective seem to be as follows. There is some dimension of “goodness”, along which a product or service may be “good” or “bad” (rather, of course, we should say “better” or “worse”). There is some point along this dimension such that at or below that point, the product/service is so bad that its users abandon it. If it is less bad than that, then its users continue using it.
Well, fair enough. We might quibble that there is more than one dimension of “goodness” (that “goodness” is multidimensional); we might quibble that “goodness” for one sort of person is not the same as “goodness” for another sort of person; etc. But let us take as an axiom, for now, that there is something sufficiently close to a single dimension of goodness (perhaps a vector through multidimensional goodness-space; perhaps the average of different people’s such vectors… or something; the details needn’t matter). Let us assume that we can, roughly, identify and agree upon this dimension.
And yet, having taken this axiom, we are actually still adrift, as far as having any kind of satisfying account of “goodness” or “good enough”, because we face two key questions:
What is the nature of this “goodness” dimension? Just what exactly makes a product or service in this category “good” or “bad”, “better” or “worse”? How, exactly—other than “do its users abandon it”—do “good” members of this category differ from “bad” ones?
Do we care about any other points or ranges along this dimension other than the “tipping point” of “bad enough to be abandoned by its users”? Is “is this product/service literally too terrible to keep using, yes/no?” the only question (as pertains to “position along the ‘goodness’ dimension”) that we have reason to ask of a product/service? Supposing we ascertain for ourselves that a product/service is just good enough for us to use—have we now no further reason at all to concern ourselves with its “goodness”?
The answer to both questions is actually the same—call it question A:
Question A: “Is this product or service accomplishing my goals / satisfying my needs / etc.?”
Now, might you argue that, actually, this is just the same thing as “is this good enough to not stop using it”?
You can argue that. But that argument would be absurd.
For one thing, you would have to argue that the answer to question A is binary: yes, this product or service is accomplishing my goals; or, no, this product or service is not accomplishing my goals. But of course it’s not binary; your goals might be accomplished only partly, or not as well, or only some of your goals might be accomplished, etc.
For another, you would have to argue that, empirically, if a product or service fails to accomplish a person’s goals, or fails to satisfy their needs, or fails to do these things well, then that person stops using said product or service. Of course no such thing is even remotely true.
So when you say “we are still using Facebook, therefore Facebook is good enough”—meaning, of course, nothing more than “Facebook is not quite so bad as to be too terrible to keep using”—then my question to you is this:
So what?
Who cares? Why should that be significant? OK, so Facebook is just “good enough” that we’re not motivated to abandon it—fine; we knew this already (it follows, tautologically, from us not, in fact, having abandoned it). So what? Does it accomplish our goals? Does it satisfy our needs? How well does it do these things? Is it good enough to make continuing to use it be the optimal strategy? By using it, do we win? Might we win more if we abandoned it, and switched to something else? These are the questions that matter!
P.S.: When I see people arguing against attempts to motivate abandonment of Facebook (or similar attempts to improve other inadequate equilibria like this) on the basis of “it’s good enough, or else we wouldn’t be using it”, I cannot help but think of this recent post.