Authors at Google: David Weinberger

Uploaded by AtGoogleTalks on 08.02.2012

>>Presenter: Okay. Welcome, everyone. It's my pleasure to welcome you here today to our
Authors at Google series. Welcome to everyone here in person, and everyone on webcast as
well. I'm Bradley Horowitz, head of Product for Google+. It's my pleasure to introduce
David Weinberger. David has one of those resumes and CVs that just goes on and on. I'm going
to pluck out a few highlights so you know-- to remind you why you're here.
He's the co-author of the bestseller the Cluetrain Manifesto, which Information Week called the
most important business book since Tom Peters' In Search of Excellence. He's also the author
of Small Pieces Loosely Joined: A Unified Theory of the Web, and the author of Everything
is Miscellaneous: The Power of the New Digital Disorder. You have a knack, sir, for titling
your books. I think these are wonderful and intriguing.
He's also a senior researcher at Harvard's Berkman Center for Law. He's the co-director
of the Harvard Library Innovation Lab at Harvard Law School and is a Franklin fellow at the
United States State Department. He's a frequent advisor, commentator on NPR and other news
organizations. He wrote gags for Woody Allen's comic strip for seven years. He notes that
this raises expectations he will not meet in the talk you're about to hear. [laughs]
He has a Ph. D. in philosophy from the University of Toronto and taught philosophy for six years.
And on and on. The way that I think about David's work is
to think about the parallel to athletics. There are some people who are just blessed.
They're born six foot one and can throw a 100 mile an hour fastball. When you pair these
people with someone who's an expert in biomechanics and can actually teach them about fast twitch
muscle fibers and air resistance and things like that, it will take the already good and
make them great. All of us here at Google, you're exceptional. You're here because you're
the best at what you do. We train each other and become steeped in our Googley principles.
Then great things happen. The way I think of David's work: he's part
sociologist, part anthropologist, part social psychologist, economics-- what else?-- journalist.
All of this is mixed in. What he does is he holds up a mirror to the work that we're doing.
He helps us better understand ourselves. He frames it in ways that are valuable in understanding
our own work. That's my experience, always, when I read his books. I end up with a better
appreciation for my own work. Things that I did and took for granted are not reframed
in ways that I can better understand myself. I really encourage you all to read his works,
to pay attention today. Without further ado, I want to welcome to the stage David Weinberger.
>>David Weinberger: What a great introduction, which I can't live up to. Thank you very much.
It's especially delightful to me to have this coming from you, Bradley, where we haven't
seen each other in a few years. In the five minutes we had coming up here, we already
had a conversation that changed my mind about a couple of things. [laughs] Typical of Google
and of Bradley. I do want to talk about the topics in this
book. In particular, a question that I think has become pretty pressing, which is this
bizarre phenomenon, unexpected phenomenon, that the most important pillars of our culture,
especially when it comes to knowledge, have started to just fall over, fall apart. Whether
it's encyclopedias--. Encyclopedias, besides being an actual collection of knowledge, were,
for many years, the sign of your home's commitment to knowledge, if you had the encyclopedia
in your house. And suddenly, it's just gone! They're not gone, but you know what I mean.
As a cultural reference. Newspapers: obviously also getting aggregated and getting reaggregated.
Libraries, which are very much--. Every librarian is wondering, "What's going to happen to libraries?"
And nobody knows. We just don't know. This core institution in our culture, of knowledge.
We don't know what's going to happen. It's pretty amazing to be living in these times.
These changes have occurred because of what, in some sense, is a really small technology:
because of hyperlinks. Little hyperlinks. This little bit of code, --granted, embedded
into much a bigger system, but nevertheless-- this little bit of technology just pings against
these mighty institutions of knowledge, and the institutions fall over. This is unexpected,
and it's certainly worth asking, "Why?" I am not going to be able to answer that question,
but I want to talk about it in terms of--. In particular, knowledge as an institution,
which is a peculiarly important, it is a key institution, knowledge is, in our culture.
I'm only talking about the West and western knowledge, by the way. In large part because--.
If you look at this quote from Senator Daniel Moynihan, it's a quote that's coming up more
and more these days. It makes a certain promise. This is the promise of knowledge in the West
for the past 2500 years: "Everyone is entitled to his own opinion, not his own facts." That's
a very cool expression, really nicely put. I think we respond to it, these days, because
it offers some comfort, some very traditional comfort, that "Yes, we all disagree and there
are fights all over the place, but if we could just sit down and focus on the facts, then
we can come to agreement. We can live together in agreement." That has been the promise of
knowledge from the beginning, from the Greek beginnings 2500 years ago. That we will be
able, that knowledge will enable us, to live together across our differences.
Really quick trip through what knowledge was. You know all of this. We thought about knowledge
as a picture of the world that we build up point by point, brick by brick. We do this
across generations. If you work hard, you might be able to contribute a brick or two
to this picture of the world that we've been creating. This is a picture in which each
piece is nailed down. We're nailing down stuff. I know I'm mixing metaphors like crazy, but--.
Finally, we're able to do this because in the great rush of sensation or the great mix
of mere opinion, which is where the concept of knowledge originally began with the Greeks:
so many opinions you have to figure out which ones are worth holding on to, which ones are
true and worthy of belief. We're able to do this because we go through these streams and
we pan for gold. Yes, four metaphors in under a minute. That knowledge is relatively scarce,
that most of what we see is not knowledge. It's wrong. It's illusion, it's opinion, it's
mere sensation. Knowledge is the real stuff that we are able to filter out. That's been
the project of knowledge in our culture. From the beginning, it has recognized and tried
to deal with the fact that the world is obviously way, way, way, way bigger than what can fit
into anybody's skulls. The Greeks knew this. We know this all the way through our history.
The world is much bigger. Our knowledge gets much bigger. Our skulls don't get much bigger.
That's a real problem. How do we small creatures know the world?
We adopted, in the West, a particular strategy, which is, to put this crudely, to reduce what
there is to know so that our brains can manage it. Our brains collectively can manage it.
One of the more prominent techniques has been to break off brain-sized chunks of the world
and to allow somebody to master that chunk, to dominate that chunk, and become an expert.
Then we can go to the expert when we have a question, or go to the expert's book, and
we can ask the expert a question. We can get an answer and we can move on because it's
been answered. This is a hugely--. And if we were not sure if we trust the expert, then
we can look up her or his credentials and move on. "Oh, I didn't know you have a degree
from wherever." This works. This system absolutely works. It's made us the dominant species on
the planet. It's incredibly efficient. But it is primarily constructed as a system of
stopping points. You've asked the expert, and the efficiency of the system is, assuming
the credential is acceptable, you don't have to ask any more. You got your answer. You
can move on. You don't have to repeat the experiment. You don't have to go look at the
primary research. The expert told you. You are right, generally, of course, to rely upon
genuine experts. But this is a system about creating a set of stopping points so that
we can manage in a world that is just way, way too big otherwise.
So it works. It's a fantastic system. It's an amazing system. But it's not what knowledge
is. It's what knowledge is when its medium is paper. This is--. [pause] Paper is itself
a stopping point. It's just the way that it is. You can't click on the page and get to
the next source. It has footnotes, you're in a book and you're reading and there's a
footnote. So there are obviously connections out to the world. We know books are written
by people who are deeply involved in social networks. All that's true. But the book itself,
as the medium of knowledge, is a hugely disconnected and disconnective medium. There's the footnote.
You don't pursue footnotes. It's very rare that anybody actually uses the footnote as
a link, because it means getting on a bus and going to the local library and hoping
that they'll have the work. And they probably won't. If they do, it's probably out. Meanwhile,
it's very expensive to get past the disconnected nature of books.
So our idea of knowledge is actually, unsurprisingly, it's completely a McLuhan thing. It's unsurprisingly
a reflection of the nature of its medium. Disconnected medium. You have a book and you're
an author. You've got to try to put everything into the book that the reader needs, because
you know the reader's not easily going to be able to hop out of the book and continue.
We have a sense of knowledge that comes directly from books, but it's not knowledge in itself.
It's not knowledge's own nature. It's what knowledge looks like in books.
But now, of course, we have this different medium. It's hugely, hugely connected. You
can think about links as being a new type of punctuation. The old type tells you where
to stop, links tell you how to continue, and provide you the means. The smallest possible
click of the human finger, and you've now done the thing that otherwise requires books
and old card catalogues in the old days, and crawling through shelves.
We have a hugely connected medium, which is fantastic, obviously. But there are also issues.
So my hypothesis is that knowledge is taking, just as it took on the properties of its old
medium, it's now taking on properties of the new medium. I want to look at four of those
properties. The first is, obviously, it's too much. The
old paper-based medium: very restricted in how much could be output or stored. Now: no.
It's way, way too much stuff. Clay Shirky said recently, and put it brilliantly, as
he tends to do-- oh, Clay-- that "There's no such thing as information overload. There's
only filter failure." Brilliant way of putting it. He's trying to provide us with some comfort,
historical comfort, some historical continuity. I think he's basically saying, "Don't freak
out. We've always, in our history, have at times felt overloaded by information, and
we know what to do about it. We fix our filters." I think he's absolutely right to give us this
sense of continuity, but I want to point to two discontinuities, which I think Clay will
probably be okay with, but I'll let him speak for himself.
I also want to-- Just a shout out to Ann Blair, who makes the same point in a book that came
out six months before my book, which is Too Big to Know. Hers is called Too Much to Know.
Just one of those things. Hers is a very scholarly and interesting look at what happened to information
in the age of Gutenberg. It's actually a little broader than that, but it's looking very carefully
at what information overload actually was, and what the filters were.
The term "information overload" itself came to prominence through Alvin Toffler's book
Future Shock, 1970. Very good book in which he models the term on the prior term, which
was "sensory overload." He did not come up with the term. Sensory overload is 1950s,
the idea which was that your senses get overwhelmed at a Grateful Dead concert. There's too many
lights and sounds and music and people pushing in on you. That can cause your sensory circuits
to blow and you will fall down as a quivering mass with your eyes rolled back in your head.
It was an individual psychological syndrome. Same thing once we decided for whatever reason,
which I don't understand but actually am quite interested in, that brains are information
processors, which let's not get started on. But once you decided that, then of course
there has to be the equivalent thing, which is information overload, with the same sort
of effect: too much information and you will fall down and be a quivering mass, irrational
mass. In fact, Toffler says that "Sanity itself hangs upon avoiding information overload."
It's very much an individual symptom. A problem that happens that strikes you as an individual.
What did information overload look like in 1974? Some marketers did some research, investigating,
in which they gave 192 housewives 16 brands, each of which had 16 different categories
of information. They discovered, these marketing researchers, that with that much information,
[pause] the quality of the housewives' decisions went down. This was information overload:
16 brands, 16 categories. The poor housewives had to be protected from this much information,
which was the conclusion. "Therefore, we're doing the housewives a favor by not taxing
them with that much information." But we look at this now, and this is like a joke. This
is laughable. That's what information overload is? 16 brands, 16 categories? I mean-- Yes,
that is what information overload looked like. It has gone, in the past, since 1970 when
the term came to prominence, from an problem for an individual, that you will fall down
quivering, to simply being not only scaled up, but just a feature of the environment.
It's not a complaint about an individual, it's about our culture. There's information
overload. When people complain about information overload, they don't say, "Oh, I feel a seizure
coming on. Somebody put a piece of wood between my teeth so I don't swallow--." They are,
instead, really saying, "I want more information. I want the right information. Get me more
information. Hurry, stat. More information." Our idea of information overload has shifted
very much. It is a different concept than we have had in the past. [pause]
Second discontinuity in information overload is that the nature of filters themselves have
changed. You know, sure, there's filter failure. But filters have changed. In the physical
world that we all are very familiar with, when you filter something, you filter the
other stuff out. If you are on the acquisitions committee of your local library, you choose
the new books. They get put up on the nice shelf in the library. New books. Everybody's
very happy when they go in and they see that. What people don't see is that a million books
are published every year. They don't see the million minus a thousand, whatever it is,
stacked in the sad and lonely trucks that are backing away, their hearts broken, from
the library loading bin. You do not see that. You only see what has been brought forward,
because that's how physical filters work. You get rid of, you hide, you dispose of,
or never make available, the stuff that doesn't make it through the filter. The articles that
get sent into the-- the manuscripts that get sent to book publishers that get rejected,
you never see those. They're filtered out. On the web, obviously, we filter differently.
When I make my list of top ten science articles of the week or whatever on my blog, all that
I'm doing is choosing ten and reducing the number of clicks that it takes to get to them.
I'm not throwing anything out. I'm not making anything invisible or hiding anything or rejecting
anything. I'm just simply shortening the number of clicks. All that material that didn't make
my top ten list is still there and will show up in the next search that somebody does on
some search engine, for example. Or it will get passed around in email. Or will be in
somebody's social network software. Or whatever. All that stuff is still there. It's still
available. It may show up. On the net, we don't filter out, we only filter forward.
And furthermore, the search engines are very, very eager to show us exactly how much is
there that they aren't showing us on the first page. We're now constantly aware of just how
much there is, something that was hidden from us before. We filter forward, we're now much
more aware of the overload. And we seem absolutely fine with it. In part because we don't have
to reduce the world by filtering out, we can enhance it by filtering forward.
This leads to a new strategy that one sees adopted all over the place, which is in general,
include everything. If you're going to try--. Rather than try to curate a collection, --and
there's tremendous value there, too, of course-- but rather than seeing this as the first alternative,
instead include everything. The economics of deletion obviously have shifted so that
it's often more, it's cheaper to include than to exclude. The problem with building curated
collections-- and as I work in a library, I do appreciate the value of them-- but the
problem with them, with deciding what's going to be in, what's going to be out, is that
you can never do that adequately. You cannot anticipate what people are going to be interested
in because we're a crazy, squirrelly species and there's no telling what we're going to
be interested in. If you are curating a news collection and
you decided to keep out of it all the gossipy, trashy stuff about celebrities, Lindsay Lohan
and all the rest of it, probably a good decision in one sense. A reasonable decision, in any
case. But you will have just deprived the set of scholars of the primary materials,
the source materials they need in order to investigate their topic, which, for example,
is the effect of media on women celebrities. You will have completely killed this academic
conference about the same topic. You don't know what's going to be interesting or important
to people. Furthermore, you can't predict history. Nobody
could have predicted that the 1996 records of the library committee in Wasilla was actually
worth including in some curated collection, because you could not predict what was going
to happen in 2008. Nobody did. So there are good reasons to try to include
everything where everything is an overstatement, but you know what I mean. If you do that,
you have to give users ways of filtering on the way out, according to what they're interested
in and how they think about how the world is organized. So much activity over the past
15 years. You can track the history of the web by watching the history of how we are
learning to filter large collections for ourselves and for one another. I'm at Google, I'm not
going to-- You know that. Okay, next characteristic of the net that
I think knowledge is picking up is messiness. We're really good at organizing things. We've
gotten incredibly good at doing it. And not simply because we want to be able to find
things, though that would be enough. But for most of our history in the past 2500 years,
we believe that we are organizing things not simply for finding but in order to discover
what they are. To know what something was, in the West, has been to know where it fits
in the order. The order of the universe. If you're Aristotle, you're arranging things
by, --and in the West, we're basically all Aristotle-- you're trying to figure out how
this is different from the other objects in its class and how it's the same as--. And
you do this nested hierarchy. To know what something is is to discover its position in
its hierarchy. The order that God established. That is our spiritual mission as well, at
least post-Greek. Or it's to see the one logical order, which is also beautiful, by the way.
The order, truth, and beauty, same thing. The order is a thing of magnificent beauty,
and it was our task to discover it. It wasn't simply a matter of finding.
Well, there's a reason why we thought this way. I believe it has a lot to do with the
fact that the physical world does work in that way. That is, when you and your spouse
are trying to figure out how to organize the CDs in your collection, and you want to do
it alphabetically and your spouse wants to do it by genre or whatever, only one of you
can win. Because in the physical world, everything does have to be in one and only one spot.
You can't have two things in the same spot. Atoms. We've dealt with it. We've taken that
limitation and we've imported it into the world of ideas so that we have the idea that
the order of the universe--. Also everything has a single spot and there's only one right
order. In arranging our physical stuff, somebody had to decide whether Philly Wacko here, who's
playing cracked avant folk while ranting against the-- can't read it-- oh, the Broads, who
wronged him, whether that's actually under British psych or whether that should be under
acid, visionaries, and weirdos. I'm sure there's a very intense argument about this. But somebody
had to win. Whereas obviously, digitally, it's a digital collection: "Well, I'll just
make playlists. We'll have as many as we want, overlaying them. We don't have to argue about
it anymore." Each of these actually exposes some meaning.
Now, to think that there is a single order that says what a bird really is and that it's
worth our time to argue over that, to do that in a hyperlinked world, a tabbed world, the
multiclassification and playlisted world? It now seems ridiculous! It's not an argument
we're interested in having anymore. It just seems [pause] pointless to think there's got
to be one order of the universe. We have this idea that dominated our culture for millennia,
has dropped away. If I say to you, "And there is a single order of the universe, of course",
this thing that was not only taken for granted but was taken as human destiny departs. It
now seems to you like I'm a crazy guy. There is a very large change. It actually,
I think, overall is a wonderful change, because it turns out that messiness is how you scale
meaning. Multiple playlists, multiple tabs, all piled up on top one another. Each one
of them has some meaning. It has some information to contribute. This is how you scale meaning.
That's what we're doing now. Okay, unsettled. [pause] Back to Moynihan.
Remember this idea that we'll all just focus on the facts and agree and sit down, preferably
over coffee--. The coffee house ideal from the Enlightenment, very powerful image. Sit
down, we have a rational conversation. We can figure this out and we can agree because
there are facts. Of course there are facts. But then you go out on the web and every link
expresses some difference between the page that you're on and the one that you're being
sent to. Here's a page that expands on what I said or disagrees or it's funny. You say,
"Why are you sending somebody to that place?" That's where the value of the web is. Take
absurd cases. If every page was exactly the same, the web would have negative value. It's
only because it's this web of difference that the web has value. It's not just difference,
it's linked difference, which is an amazing development that we not only have a moment
in our history where we're able to see all the ways we disagree and enhance one another,
the ways that our points of view are different, those points of view are now linked. Those
of you who are fans of Hegel, this is a wonderful Hegelian moment where [pause] it's-- yeah.
Here we have these differences that are in conversation. Sometimes quite literally in
conversation. This is very different than the picture of the world that Moynihan paints
for us and that we've had as part of our Enlightenment ideal. It really bothers us. I think there's
truth in each of these worries, each of these fears. We should be taking them really seriously.
There is truth in them. You go out on the web, and what you see is all of this disagreement.
Well, I think what's actually happening is that the net is exposing a truth that we've
always known in our hearts, but we couldn't quite acknowledge. We didn't have the environment
which made it possible to acknowledge it, which is, you know what, we actually don't
agree about anything. We don't agree about anything! There's nothing on the web that
is--. There's no fact uncontested on the web. There are facts. I like science, I like facts.
So I want to be really clear: there are facts. Some statements are false, and there are people
who lie. I totally believe that. My point is is we're not going to agree about who those
people are. We're just not. It doesn't matter how many facts you bring to bear. Nope! There's
going to be continuous disagreement, and there always will be. My evidence for saying that
is all of human history, we've never agreed about anything. Now, we can see that we don't.
It is evident before our faces. One of the more-- . This is a problem. One
of the more hopeful things, though, is that we have also been evolving rapidly ways of
dealing with difference and disagreement that enable us to move forward and in fact benefit
from that disagreement. This is actually a terrible example, but it's a real example
from a couple of days ago. There's this YouTube site that famously has long conversations.
They're not particularly well threaded conversations. It doesn't seem to be what whoever-- Oh, I
see, sir, you seem have a YouTube sweatshirt on. You must have been to their offices at
some time. Those conversational threads at YouTube serve a different purpose. They're
not for long threaded conversations. Here's the new Batman trailer. 4.5 million views.
From about three days ago, there's a thread of about 30 back and forths about circumcision.
It's pretty learned. [laughter] It gets a little nasty, but it's pretty learned. They're
flinging facts back and forth and so on. It's a little hard to understand exactly how this
got started, and it doesn't matter. The point is that in a different conversational medium--
At YouTube, it doesn't matter because it will just be pushed down, there's such volume.
But in a different conversational medium, there would be a way of forking conversations.
Forking is a really powerful tool that's hard to do in the real world. It's rude and disruptive
to do it in the real world. But it's usually helpful in the online world where people are
going down, in a very heated fashion, down some thread that has nothing to do with the
original thread. A fork takes us somewhere--. Continue it. Anybody else who would like to
continue arguing about circumcision, please feel free to do so. We're not ending the conversation
by any means, just maybe not on the Batman site. Because we want to talk about something
else. Forking is a really powerful tool for enabling difference to continue without telling
people, "Shut up!" [laughter] which is what we would do in the real world if that conversation
broke out. The second technique that I think is amazingly
powerful for dealing, not only dealing with difference but benefiting from it is namespaces.
In the 19th century, we spent an awful lot of time trying to figure out whether that
thing is--. First of all, is it really a mammal or what is it? And second of all, could it
possibly even exist? Because it violated the classification schemes that we had. It didn't
fit into our taxonomy. There was very real thought that it must be a hoax. Even when
one was delivered with eggs in it, it was assumed it must be a hoax because such a thing
could not exist. We spent a lot of time arguing about how to classify this. Now, we don't
so much because, and other species and genes as well because, for example, because we can
just set up a namespace. Say, "Well, okay, I call it a platypus and I think about it
this way. You call it a water mole. That's fine, so long as I know that what I call platypus,
you call water mole and you call this thing." Namespaces are a wonderful way of dealing
with differences. At Encyclopedia of Life, they in fact have
two servers. One--. If you're a scientist who wants some information, you can use whatever
nomenclature you want, because they track them all, that they can find. It doesn't matter
what you call it. Call if what you want. That's the least interesting thing about it, what
name you use for it. You can put it within whatever taxonomy you prefer. That's not--.
And if two scientist want to talk about that thing, even though they call them something
different and classify it differently, great. You can do that because it's namespaces. We
can coordinate namespaces. Another way that we are evolving very rapidly
to deal with difference and to benefit from it is the linked data, which I'm not going
to explain because I'm at Google. But it is, in fact, a way of letting people release data
very easily without having to worry about getting it classified right in the uniform
way. No, you do it in some reasonable way. You point to disambiguating pages. If somebody
does the mapping right, you're pointing to Encyclopedia of Life, I'm pointing to the
Tasmanian Field Guide, which I've made up. We want to know--. If we want to enable computers
to know that we're all talking about the same thing, we can do a mapping and say, "That
page in Encyclopedia of Life, that's really referring to the same thing as that page in
the Tasmanian Field Guide." It's a really good way of getting past differences that
don't matter so we can collaborate. "Yay for difference! Difference is wonderful, difference
is great." It is, but. [pause] This is what the net looks like a little bit.
As it is, this is what the net looks like in terms of how we experience it. We do seem
to tend to want to hang out with people who are like us. There's reason to think that
this is a human thing, maybe even a genetic thing, maybe a brain thing. But we do it.
We like to be with people who are like us. This is the echo chamber argument that you're
all familiar with. I just want to spend one slide on the echo chamber, because I think
it's a real issue. I am not nearly as convinced, as many others are, that it is the killer
issue that some have claimed. But I do think it's a real issue. The problems with understanding
it are--. So I say one slide, but I've got lots of words on this slide. Sorry. The problems
are difficult and deep and, I think, emblematic of a difference about how we think about knowledge
and conversation, how knowledge is developed. First of all, we don't have a lot of data
that tells us exactly how much, what the diversity of sources are that we engage with. That's
more data. Part of the problem is, compared to what? We want to know if the net is making
us more open or not. The echo chamber argument says, "It's easier to hang out with people
that you like." Hanging out with--. Excuse me. "People who are like you", which turns
out to be often the same thing. If you hang out with people who are like you in their
beliefs, then you end up reconfirming your own beliefs and, in fact, becoming more extreme
in your beliefs. If that's the case, the internet is making us less open. But the question is:
compared to what? Compared to pre-net, when we had, when I grew up, three television channels
and you had a couple of newspapers. The newspapers were this big and reported a range of opinion
that was actually even smaller than that? Compared to what? We don't have a baseline.
And we have a romanticized view of what those days were like. "Back in the old days, we
all read newspapers. Newspapers were full of news." Well, first of all, A: we didn't,
B: we're reading tabloids, which are full of crap, and third of all, we don't know what
people were actually reading in a newspaper. The New York Times is full of wonderful stories.
Most of us read it like this, looking for a headline that catches our eye. It's not
like 15 years ago, we're all studiously going through every article. So it's really not
clear what we're comparing it to. Clay Shirky points out that we may even have
the causality wrong. The echo chamber theorists. [indistinct] in particular. That they may
have the causality wrong. That maybe, in fact, we are becoming more polarized because we
are in more contact with people with whom we disagree. That causes us to harden our
opinions. [pause] There are times in the day when I'm totally an echo chamber guy. If I'm
going onto a site of a politician, a candidate who I support, I want an echo chamber. I need
an echo chamber, because that's how political movements are formed. If I'm going to a sports
site. I don't, but I can imagine. If I went to a sports site, I don't want to particularly
maybe want to hear from everybody who hates my team. I'm from Boston, so I'll say the
Yankees. I don't necessarily want to go to a site where the Red Sox fans and the Yankees
fans are always mixing it up. I want to hang out with Red Sox fans. It's baseball, right?
[laughter] It's not at all clear how this varies by education level, if it's there at
all. By education level, by socioeconomic class, by topic, by time of day, and the rest
of it. But for me, the two more important, when it
comes to knowledge, what's happening in knowledge, the two more important points are the next.
The first is that the echo chamber argument, in at least some of its forms, assumes that
the role, the only good conversation is one where two people who really disagree can sit
down in the Moynihan sense, over some coffee and a big bowl of facts. They each say, "I
disagree with you, my friend. I'm a Jew, you're a Nazi. I disagree with you. But hey, I respect
you. I respect your opinions. And I'm open. Maybe you're right. I'm open to being converted.
Let's sit down. Maybe I'll become a Nazi and maybe you'll become a Jew." That's a good
conversation. That's a conversation that never happens. Never ever. Never happens. I'm going
to tell you I'm never going to sit down with a Nazi with an open mind and think, "Yeah.
That's a good point. You make a good point there, Adolf." Not going to happen. It's not
how conversation works, nor should it be how conversation works, except under very special
circumstances. And so to say, "Well, the web is not full of these Jew meets Nazi conversations.
That's a failure of the web." No, it's not. It's a failure of expectation about humans
and conversation. The same thing for--. It seems to me that
the echo chamber argument, the people who are most worried about it have a different
idea about what conversation is for. Excuse me, how it works. In order for two people
to talk, diversity is wonderful. Keep in mind, the whole point of this little subsection
I'm trying to make is that knowledge, now, can include diversity and difference. Disagreement
within itself. This is a big change from knowledge, and it's wonderful. It's really important.
At the same time, in order for two people to talk, there has to be so much agreement
and so much sameness. Same language, interested in at least one topic they can talk about,
enough common assumptions about how that domain works and what the facts in that domain are,
some sense of the norms of conversation: who gets to talk, how much, when you interrupt,
all that sort of very subtle thing. Conversation, to be successful, needs to be 99% sameness.
I made the number up, but it's overwhelming sameness, and then very small iterations on
top of that of difference. That's a requirement for conversation. It's also a requirement
for the conversation that you have with the people with whom you're working and you want
to move the ball forward. You need an enormous amount of sameness. The fact that there isn't
an enormous amount of difference in these conversations on the web is again, not a failure
of the conversations on the web, it's a failure to recognize how conversation works.
Now, having said all that, I don't know how real the echo chamber effect is. I suspect
it is in at least some ways very real. But it doesn't actually matter too much whether
we have an exact measure of how damaging it is. We, I think, should have some confidence
that human beings do prefer to hang out with people whose beliefs are the same as theirs,
that there is damage that can be done by doing that, and that there is benefit to opening
ourselves up to more and further types of belief. No matter how damaging the echo chamber
effect is, doesn't matter. We still need to be doing everything we can to counteract it.
Both personally and institutionally and as a business.
I will just take the, I'm very happy to be here. Nevertheless, the including of--. Running
Google's filters through my social network pretty well ensures I'm going to get more
of the same, when in fact, I would rather--. Speaking as a member of a culture I'm trying
to live in, I would rather have you disturb me more. I do appreciate that there's a button
that allows you to turn off the personalized, so I can see. I would also want you to--.
This is very similar to the long tail argument, by the way. You want to pull stuff out of
the long tail, and you want to disturb people like me so I don't get too comfortable in
my beliefs that need challenging. Okay. And clearly, it's not simply a Google issue. Bing
does it, and the rest of it. It's an industry issue.
So. [sighs] Here's Moynihan, here's the promise, the comfort that we will all come together
and we will come to agreement and that's what knowledge is for. No. Knowledge is better
off because, now, we recognize that we don't agree and because we are linked in those disagreements.
Knowledge is not the thing that is settled anymore. It is where this disagreement happens.
I should point out that this quotation, it's unattributed. We don't know if Moynihan actually
said it, who said it, or what the exact quote is, because irony remains the fundamental
law of the universe. The fourth and last point is the web's too,
the internet is so scattered, so knowledge is getting to scattered. Well, no. Books,
the opposite of scattered. They're literally bound together in sequence. You can not get
less scattered than a book. This has been the medium of knowledge. And from this has
come a way of thinking about how ideas go together, how knowledge, and thus how the
world itself, because knowledge is a picture of the world. Knowledge as it comes through
books is a picture of the world taken in short steps, each one of which has been tied down
with deductive logic or with strong evidence. If you're writing a book, you're going to
start your readers at A, you're going to try to get them to Z by taking these short, measured
steps. That is the ideal in our culture. That has been the ideal of knowledge. It's the
person who's sitting there writing the long-form work that proves her or his point.
But there are problems with this that go unnoticed when all we have are books. For example, it
means that we write in private, which, if you're designing a knowledge ecology, seems
entirely backwards. You want to be working on this in public, so that when, for example,
Darwin, who obviously wrote one of the great long-form works. In The Origin of Species,
the first five chapters lay out the theory. The next six chapters is Darwin sitting alone,
by himself, thinking about all the awful things that people will say who object to his theory
that he's worked on for decades. Some of those objections were real. He was socially-- He
was out among his colleagues and his social network. But some were also him thinking,
"One might say. One might object." It's a totally suboptimal way of doing this, but
it's what you had to do because you're writing. That's how you wrote: in private.
Now we don't have to do that so much. Another problem with the long-form, book-based long-form
is that you're trying to get your readers to Z. You're trying to keep them on the tour
bus. Books, in some ways, are too short. long-form is too short, or it's not wide enough. You
have to try to keep them focused, keep them focused, even though there's a whole landscape
full of interesting ideas and diversions and possibility. "No! Stay on the bus! You're
on a tour! I'm in charge of the tour! We're going to go like this, and we're going to
get here. And then later, you can go explore. But it's my damn tour bus." A third problem--.
And now that's a lost opportunity, that's a paring down of information and of knowledge.
The third is, that even though we love and adore long-form works of knowledge, they very
rarely work. It's very rare that you'll actually start at A and get all the way to Z. Origin
of Species, yeah. At least for me, I think most people in this room, you get to Z. That's
a great work. And it happens occasionally. But it doesn't work nearly as often as we
think it does, which, I think, maybe should make us think that this is a poor way of trying
to reconstitute the world. That the world maybe doesn't work as short steps that lean
in a long chain. That maybe God doesn't think in long-form arguments. Because they don't
work, generally. They often just feel like arbitrary constructions. So I'm certainly
not saying that books are going away, going to be replaced by Twitter or whatever. [laughter]
But it seems pretty clear to me that books are losing their pride of place as the epitome
of knowledge. The widest and where knowledge lives. Knowledge lives on networks now. The
opposite of long-form is not short-form, it's in fact web-form. It's the web. If Darwin
were alive now, one assumes that he would be tweeting from the Beagle as it goes around
the Galapagos, and writing his blog, and posting into archive dot org, and getting things wrong
and right. There would be a web of people, some of whom he knew, most of whom he doesn't,
who are correcting him, arguing with him, bringing in their own evidence to support
or to deny, to explain his works to people who otherwise wouldn't understand them, to
make money off of him as travel agents on tours of the Galapagos. "See the finches yourself!
Go on a wild finch hunt!" A web would grow up. Darwin would publish, perhaps, his work,
his long-form work. But it would have more value embedded in the web than alone on a
bookcase. Knowledge actually, now, lives in this network. It's at the level of a network,
of people corresponding, disagreeing, agreeing, enhancing, differing in every possible way.
That's where knowledge is. You can see this in action. This is an example
I'm borrowing from Michael Nielsen, who's Reinventing Discovery--. It came out about
a month ago. Really good book about network knowledge in science. He, in conversation,
points so the faster than light neutrinos, about which I know nothing. But the original
article was posted in archive dot org, prepress, no peer review, no guarantee that it's been,
just put it up there as is. Within a couple months, there had been 80 other articles posted
there, plus countless other ways that people wade in all across the web. Expanding, expounding,
proposing ideas, explaining to people like me who couldn't otherwise understand it. This
web is where the knowledge was. It did not go through peer review. If it went through
peer review, it would be two years before it got published. Maybe a year and a half
if they hurry. Then somebody would write a letter, and three months later the letter
would be published. It's too slow. You can't scale science this way. We already know this,
for example, from genetics. Very similar sort of thing. Open access could not be more important.
The knowledge lives in the web. One of the reasons I think this is really
important is that previous media were other people deciding what's interesting to us.
They'd be right or wrong, and frequently, they'd be right. But frequently, they'd be
wrong. The web consists--. It's not that it's a crowd that's deciding what's interesting
to us. It's that the web is a direct reflection of what's interesting to us. Every link is
an expression of human interest. A direct expression of human interest. The web overall
is an expression of what we as a species care about, at least, people on the web. It can
be very distressing, because apparently about 50% of what we care about is porn. Nevertheless,
that's what it is. Each of these links is something human and live. It is a direct expression
of what matters to us. This is a transformative of knowledge.
Why did knowledge turn out to be so fragile, to go back to that original question? I promised
you I can't answer it. But I want to point to one, well actually, two points. The first
is that if you go through the traditional list of characteristics of knowledge, that
there's only one knowledge, it's singular, it's settled, it's unchanging, neat and organized,
it's relatively scarce, it's impersonal in that anybody, whatever is true is true for
everybody. You look at the characteristics of network knowledge. These characteristics
are characteristics, I believe, not only of knowledge, but also of the internet, because
that's my hypothesis. The new medium is giving knowledge its new properties. These are also
characteristics of the world that we live in. What it means to live in a world as a
human being. It's overwhelming, it's unsettled, it's unresolved, it's messy, it's deeply,
endlessly connected, but sometimes loosely connected. It's held together by the fact
that the world matters to us, that it's interesting to us, that we care about it.
The first conclusion, --I should maybe put conclusion in air quotes-- but the first conclusion
is that network knowledge may or may not be better knowledge, truer knowledge. I think
that in general, this is the greatest time in human history to be somebody who cares
about knowledge. So I would say, "Yeah, generally, it is truer. But leave that aside, let's not
have that argument." Nevertheless, network knowledge, the knowledge that we see emerging
on the web, is truer to the nature of knowledge. That is, to knowledge as part of the human
condition, as truer to the world in which we live. As opposed to the old that said,
"No, we can settle things with certainty by passing them through expert filters." No.
We couldn't do that. The second conclusion, the last thing I want
to say, is that, despite Moynihan and the Enlightenment, what we have in common is not
knowledge. That's not what's going to enable us to live together peacefully in the world.
We're not going to agree about things. What we have in common is not knowledge about which
we agree, but what we have in common is a shared world about which we disagree. What
makes me hopeful is that on the web and with knowledge in particular on the web, I do see
us getting far better, more sophisticated, and in some ways more tolerant of disagreement,
and in fact, enabling difference, using difference to advance us. If there's going to be peace
through knowledge, it's going to be a very, very noisy peace.
Thank you.
>>Female #1: So, questions at the mike.
>>Weinberger: Questions at the mike. [clears throat] I'm sorry to have gone so long.
>>Female #1: Oh, it's still early.
>>Male #1: Thanks for your presentation. But it made me curious: why did you decide to
write a book to express all of that?
>>Weinberger: Why did I write a book? A couple reasons. I'm old fashioned. My parents would
be proud if they were here. And the web doesn't yet pay advances. [laughter] I'm sorry, that's--.
Unfortunately, those are honest answers. And they're not very good answers, which is why
it's unfortunate. Maybe saying in my defense, I do a lot on the web, too. [pause]
>>Male #2: Even if the--. Well, taking that what you're saying is right, that the knowledge
is across the web in the network, how do you think about--. There's still, presumably,
even if people have different opinions, relatively more valuable information knowledge there
and a lot of crap, for lack of a better word. Do you form any judgment about differentiating
stuff that may be valuable to different people versus what's less valuable to everybody?
>>Weinberger: Yes. I mean, we have to. We have to, otherwise we [pause] we'll die, and
we'll kill others inadvertently. So yeah, we have to do this. It seems to me the only
process is basically a political one. I don't meant through elected officials, though that
plays a part, but rather, it's what it has, in fact, always been as the postmodernists--.
And all of this, by the way, postmodernists say, everything I just said. Except that now,
it's actually real on the web. They said that too, by the way. As they have long said and
recognized, knowledge has been imbued with struggles for power. That was hidden to us
to some extent because the victors were so absolute in their victory. You couldn't hear
all the disagreement. All you could hear was what was coming through the broadcast towers,
and they're pretty much in agreement. So it was hidden from us. But it was there, because
it was a solid victory for those who gained control of the means of knowledge. We are,
I think, destined forever to be in a continuous negotiation and political struggle to try
to get what we believe with good cause is the truth to carry policy and to inform people's
minds. There's no--. The old technique of establishing a canon and basically removing
access to everything outside of that canon simply won't work anymore. There was an enormous
price to that old technique as well, so I don't think this is all bad. This, to me,
is a more realistic--. We now have a more realistic assessment of how knowledge works.
To put this more positively, I think this is very close to how scientists actually think
and work. They come to important conclusions. They base important decisions on those conclusions.
But they always talk about hypotheses and evidence, and are aware that it could be wrong.
That's a more adult view of knowledge than the "It says it in the Encyclopedia on the
page, so it's got to be right" view, which is only a bit of a caricature of our previous
paper-based approach. [pause]
>>Male #3: Hi. In another life, I'm a math and science teacher. Much of our time is spent
disseminating. The debate is around "What should we teach?" as opposed to "Why?" or
"How?" Even when we use social media and web, it's to disseminate or to help students to
find information. Based on your research, what do you see as helpful going forward for
using the web, that knowledge base?
>>Weinberger: I would always defer to a teacher who's actually teaching. Having said that,
I'm about to say things that will be foolish, at best. [pause] [sighs] I have nothing interesting
to say. There are obviously levels of things that we want taken as fact, and are fact.
No reason to dispute them. We want our students to know those, and not to have serious doubts
about them, always to be slightly skeptical. That's what we've always done as part of the
educational system. We clearly need to continue doing that. It's not, "Well, just go out,
build a Wiki, and whatever you believe is true", because it's not true that whatever
you believe is true. The phrase "Well, that's true for you" actually drives me insane. Because
in that case, we're not talking about truth. And we do need to be talking about truth in
the educational system. There are a couple things I think may be happening.
I think they're all obvious. One is I think we're beginning to understand through the
web and somewhat through our children that learning is not simply about enriching the
individual. That's nice, that's important, but that's a selfish act. You get educated
and you become a better person, and that's grand. Maybe you become a better citizen and
you help others. But so much--. You look at the software development community, the environment
that they, or perhaps I should say you, have built for themselves. It is so generous and
so amazingly productive. It's reasonable to argue that the software developers have built
the best environment for rapid learning we've ever had in our history. You have a question,
you'll get it answered. You need a code fragment, it's there. It's there for free. When you
do your own, you'll improve that one or do another, you'll make it free. This culture
of learning in public as an act that is not only bettering you, but is bettering the public,
that is an amazing thing within the software world. I hope that it will become an amazing
thing within the educational system overall. It's not simply about me making you smarter,
my student. It's about us learning together, A, of course, and B, doing this in public
as a way of improving the world in which we all live. Helping students to be literate
about the web, that would also be really, really helpful as well. That's beginning to
happen. But it seems a little slower than I would have expected. [pause] I told you
I had nothing. [laughter]
>>Male #4: Part of the value of pursuing a one definitive truth is that, at least in
theory, it should have a correspondence with an objective reality. If you understand gravity,
you understand the motions of the planets. If we lessen our focus on the one truth, does
it lessen our ability to be able to optimize our reality? For example, coming up with economic
policies that actually grow economies or safety policies that keep us safe.
>>Weinberger: First of all, we absolutely should be trying to head towards the one truth.
We just call it "the truth." The world is one way, and it is not another. I firmly believe
that. And we should be, of course, heading as far as, doing our best with that. I'm going
to reject your metaphysics, but accept your conclusion. I have-- Maybe the shorthand for
this is that I'm basically a Heideggerian, and take a more phenomenological approach,
so I don't--. The correspondence idea of truth, that it's a set of statements that corresponds
to how the real world is, for me, leads out the way in which truth is an aspect of how
we engage with the world, not an independent picture of it.
We can leave that argument aside. I'm not sure that it actually makes a difference.
And still think that yes, we absolutely have to be trying to figure out what's true about
the world so that we can make good policies that will then affect the world. That's crucial.
I want to be clear about what I was saying before, because we need scientists, we need
mathematicians, we need sociologists, we need everybody to be working on this. What I was
saying before is we should give up on the hope that we're all going to agree about the
facts. But there are still facts, and we still need to be pursuing them for precisely the
reason you say, because the real world, reality, our fates, depend upon them. I've compressed
the metaphysics and accept the beginning and end of your statement. [laughs]
>>Female #2: I want to thank you for coming to Google today.
>>Weinberger: Thank you for having me. Oh, >>Female #2: and that's it! [laughter]
>>Weinberger: Thank you very much.