Edge; DIGITAL MAOISM: the Hazards of the New Online Collectivism By

Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

Home About Edge Features Edge Editions Press Edge Search Subscribe

The hive mind is for the mos t part s tupid and boring. Why pay attention to it?

The problem is in the way the Wikipedia has come to be regarded and us ed; how it's been elevated to s uch importance s o quickly. And that is part of the larger pattern of the appeal of a new online collectivis m that is nothing les s than a res urgence of the idea that the collective is all-wis e, that it is des irable to have influence concentrated in a bottleneck that can channel the collective with the mos t verity and force. This is different from repres entative democracy, or meritocracy. This idea has had dreadful cons equences when thrus t upon us from the extreme Right or the extreme Left in various his torical periods . The fact that it's now being re-introduced today by prominent technologis ts and futuris ts , people who in many cas es I know and like, does n't make it any les s dangerous .

DIGITAL MAOISM: The Hazards of the New Online Collectivism [5.30.06] By Jaron Lanier

An Edge Original Essay

Introduction

I n "Digital M aos im", an original es s ay written for Edge, c omputer s c ientis t and digital vis ionary Jaron Lanier finds fault with what he terms the new online c ollec tivis m. H e c ites as an example the Wikipedia, noting that "reading a Wikipedia entry is like reading the bible c los ely. T here are faint trac es of the voic es of various anonymous authors and editors , though it is impos s ible to be s ure".

H is problem is not with the unfolding experiment of the Wikipedia its elf, but "the way the Wikipedia has c ome to be regarded and us ed; how it's been elevated to s uc h importanc e s o quic kly. A nd that is part of the larger pattern of the appeal of a new online c ollec tivis m that is nothing les s than a res urgenc e of the idea that the c ollec tive is all-wis e, that it is des irable to have influenc e c onc entrated in a bottlenec k that c an c hannel the c ollec tive with the mos t verity and forc e. T his is different from repres entative democ rac y, or meritoc rac y. T his idea has had dreadful c ons equenc es when thrus t upon us from the extreme Right or the extreme Left in various

1 of 11 01/27/2009 12:56 PM Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

his toric al periods . T he fac t that it's now being re-introduc ed today by prominent tec hnologis ts and futuris ts , people who in many c as es I know and like, does n't make it any les s dangerous ".

A nd he notes that "the Wikipedia is far from being the only online fetis h s ite for foolis h c ollec tivis m. T here's a frantic rac e taking plac e online to bec ome the mos t "M eta" s ite, to be the highes t level aggregator, s ubs uming the identity of all other s ites ".

Where is this leading? Lanier c alls attention to the "s o-c alled 'A rtific ial I ntelligenc e' and the rac e to eras e pers onality and be mos t M eta. I n eac h c as e, there's a pres umption that s omething like a dis tinc t kin to individual human intelligenc e is either about to appear any minute, or has already appeared. T he problem with that pres umption is that people are all too willing to lower s tandards in order to make the purported newc omer appear s mart. Jus t as people are willing to bend over bac kwards and make thems elves s tupid in order to make an A I interfac e appear s mart (as happens when s omeone c an interac t with the notorious M ic ros oft paper c lip,) s o are they willing to bec ome unc ritic al and dim in order to make M eta-aggregator s ites appear to be c oherent."

Read on as Jaron Lanier throwns a lit M olotov c oc ktail down towards P alo A lto from up in the Berkeley H ills ...

—JB

THE REALITY CLUB:Res pons es to Lanier's es s ay from Douglas Rus hkoff, Q uentin H ardy, Y oc hai Benkler, C lay Shirky, C ory Doc torow, Kevin Kelly, E s ther Dys on, Larry Sanger, Fernanda V iegas & M artin Wattenberg, Jimmy Wales , George Dys on, Dan Gillmor, H oward Rheingold

PRESS: Slashdot, Sueddeutsche Zeitung, perlentaucher.de, Die Welt, Inside Higher Ed, Sueddeutsche Zeitung, Arte & Letters Daily, BoingBoing, New York Times, Open Democracy, The Guardian [click here]

DIGITAL MAOISM

(JA RO N LA N I E R:) M y Wikipedia entry identifies me (at leas t this week) as a film direc tor. I t is true I made one experimental s hort film about a dec ade and a half ago. T he c onc ept was awful: I tried to imagine what M aya Deren would have done with morphing. I t was s hown onc e at a film fes tival and was never dis tributed and I would be mos t c omfortable if no one ever s ees it again.

I n the real world it is eas y to not direc t films . I have attempted to retire from direc ting films in the alternative univers e that is the Wikipedia a number of times , but s omebody always overrules me. E very time my Wikipedia entry is c orrec ted, within a day I 'm turned into a film direc tor again. I c an think of no more s uitable punis hment than making thes e determined Wikipedia goblins ac tually watc h my one s mall old movie.

T wic e in the pas t s everal weeks , reporters have as ked me about my filmmaking c areer. T he fantas ies of the goblins have entered that portion of the world that is attempting to remain real. I know I 've gotten off eas y. T he errors in my Wikipedia bio have been (at leas t prior to the public ation of this

2 of 11 01/27/2009 12:56 PM Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

artic le) c harming and even flattering.

Reading a Wikipedia entry is like reading the bible c los ely. T here are faint trac es of the voic es of various anonymous authors and editors , though it is impos s ible to be s ure. I n my partic ular c as e, it appears that the goblins are probably members or des c endants of the rather s weet old Mondo 2000 c ulture linking ps yc hedelic experimentation with c omputers . T hey s eem to plac e great importanc e on relating my ideas to thos e of the ps yc hedelic luminaries of old (and in ways that I happen to find s loppy and inc orrec t.) E dits deviating from this s et of odd ideas that are important to this one partic ular s mall s ubc ulture are immediately removed. T his makes s ens e. Who els e would volunteer to pay that muc h attention and do all that work?

T he problem I am c onc erned with here is not the Wikipedia in its elf. I t's been c ritic ized quite a lot, es pec ially in the las t year, but the Wikipedia is jus t one experiment that s till has room to c hange and grow. A t the very leas t it's a s uc c es s at revealing what the online people with the mos t determination and time on their hands are thinking, and that's ac tually interes ting information.

N o, the problem is in the way the Wikipedia has c ome to be regarded and us ed; how it's been elevated to s uc h importanc e s o quic kly. A nd that is part of the larger pattern of the appeal of a new online c ollec tivis m that is nothing les s than a res urgenc e of the idea that the c ollec tive is all-wis e, that it is des irable to have influenc e c onc entrated in a bottlenec k that c an c hannel the c ollec tive with the mos t verity and forc e. T his is different from repres entative democ rac y, or meritoc rac y. T his idea has had dreadful c ons equenc es when thrus t upon us from the extreme Right or the extreme Left in various his toric al periods . T he fac t that it's now being re-introduc ed today by prominent tec hnologis ts and futuris ts , people who in many c as es I know and like, does n't make it any les s dangerous .

T here was a well-public ized s tudy in Nature las t year c omparing the ac c urac y of the Wikipedia to Encyclopedia Britannica. T he res ults were a tos s up, while there is a lingering debate about the validity of the s tudy. T he items s elec ted for the c omparis on were jus t the s ort that Wikipedia would do well on: Sc ienc e topic s that the c ollec tive at large does n't c are muc h about. "Kinetic is otope effec t" or "V es alius , A ndreas " are examples of topic s that make the Britannica hard to maintain, bec aus e it takes work to find the right authors to res earc h and review a multitude of divers e topic s . But they are perfec t for the Wikipedia. T here is little c ontrovers y around thes e items , plus the N et provides ready ac c es s to a reas onably s mall number of c ompetent s pec ialis t graduate s tudent types pos s es s ing the manic motivation of youth.

A c ore belief of the wiki world is that whatever problems exis t in the wiki will be inc rementally c orrec ted as the proc es s unfolds . T his is analogous to the c laims of H yper-Libertarians who put infinite faith in a free market, or the H yper-Lefties who are s omehow able to s it through c ons ens us dec is ion- making proc es s es . I n all thes e c as es , it s eems to me that empiric al evidenc e has yielded mixed res ults . Sometimes loos ely s truc tured c ollec tive ac tivities yield c ontinuous improvements and s ometimes they don't. O ften we don't live long enough to find out. Later in this es s ay I 'll point out what c ons traints make a c ollec tive s mart. But firs t, it's important to not los e s ight of values jus t bec aus e the ques tion of whether a c ollec tive c an be s mart is s o fas c inating. A c c urac y in a text is not enough. A des irable text is more than a c ollec tion of ac c urate referenc es . I t is als o an expres s ion of pers onality.

For ins tanc e, mos t of the tec hnic al or s c ientific information that is in the

3 of 11 01/27/2009 12:56 PM Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

Wikipedia was already on the Web before the Wikipedia was s tarted. Y ou c ould always us e Google or other s earc h s ervic es to find information about items that are now wikified. I n s ome c as es I have notic ed s pec ific texts get c loned from original s ites at univers ities or labs onto wiki pages . A nd when that happens , eac h text los es part of its value. Sinc e s earc h engines are now more likely to point you to the wikified vers ions , the Web has los t s ome of its flavor in c as ual us e.

When you s ee the c ontext in whic h s omething was written and you know who the author was beyond jus t a name, you learn s o muc h more than when you find the s ame text plac ed in the anonymous , faux-authoritative, anti-c ontextual brew of the Wikipedia. T he ques tion is n't jus t one of authentic ation and ac c ountability, though thos e are important, but s omething more s ubtle. A voic e s hould be s ens ed as a whole. Y ou have to have a c hanc e to s ens e pers onality in order for language to have its full meaning. P ers onal Web pages do that, as do journals and books . E ven Britannica has an editorial voic e, whic h s ome people have c ritic ized as being vaguely too "Dead White M en."

I f an ironic Web s ite devoted to des troying c inema c laimed that I was a filmmaker, it would s uddenly make s ens e. T hat would be an authentic piec e of text. But plac ed out of c ontext in the Wikipedia, it bec omes drivel.

M ys pac e is another rec ent experiment that has bec ome even more influential than the Wikipedia. Like the Wikipedia, it adds jus t a little to the powers already pres ent on the Web in order to ins pire a dramatic s hift in us e. M ys pac e is all about authors hip, but it does n't pretend to be all-wis e. Y ou c an always tell at leas t a little about the c harac ter of the pers on who made a M ys pac e page. But it is very rare indeed that a M ys pac e page ins pires even the s lightes t c onfidenc e that the author is a trus tworthy authority. H urray for M ys pac e on that c ount!

M ys pac e is a ric her, multi-layered, s ourc e of information than the Wikipedia, although the topic s the two s ervic es c over barely overlap. I f you want to res earc h a T V s how in terms of what people think of it, M ys pac e will reveal more to you than the analogous and enormous entries in the Wikipedia.

T he Wikipedia is far from being the only online fetis h s ite for foolis h c ollec tivis m. T here's a frantic rac e taking plac e online to bec ome the mos t "M eta" s ite, to be the highes t level aggregator, s ubs uming the identity of all other s ites .

T he rac e began innoc ently enough with the notion of c reating direc tories of online des tinations , s uc h as the early inc arnations of Y ahoo. T hen c ame A ltaV is ta, where one c ould s earc h us ing an inverted databas e of the c ontent of the whole Web. T hen c ame Google, whic h added page rank algorithms . T hen c ame the blogs , whic h varied greatly in terms of quality and importanc e. T his lead to M eta-blogs s uc h as Boing Boing, run by identified humans , whic h s erved to aggregate blogs . I n all of thes e formulations , real people were s till in c harge. A n individual or individuals were pres enting a pers onality and taking res pons ibility.

T hes e Web-bas ed des igns as s umed that value would flow from people. I t was s till c lear, in all s uc h des igns , that the Web was made of people, and that ultimately value always c ame from c onnec ting with real humans .

E ven Google by its elf (as it s tands today) is n't M eta enough to be a problem. O ne layer of page ranking is hardly a threat to authors hip, but an

4 of 11 01/27/2009 12:56 PM Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

ac c umulation of many layers c an c reate a meaningles s murk, and that is another matter.

I n the las t year or two the trend has been to remove the s c ent of people, s o as to c ome as c los e as pos s ible to s imulating the appearanc e of c ontent emerging out of the Web as if it were s peaking to us as a s upernatural orac le. T his is where the us e of the I nternet c ros s es the line into delus ion.

Kevin Kelly, the former editor of Whole Earth Review and the founding E xec utive E ditor of Wired, is a friend and s omeone who has been thinking about what he and others c all the "H ive M ind." H e runs a Webs ite c alled Cool Tools that's a c ros s between a blog and the old Whole Earth Catalog. O n Cool Tools , the c ontributors , inc luding me, are not a hive bec aus e we are identified.

I n M arc h, Kelly reviewed a variety of "C ons ens us Web filters " s uc h as "Digg" and "Reddit" that as s emble material every day from all the myriad of other aggregating s ites . Suc h s ites intend to be more M eta than the s ites they aggregate. T here is no pers on taking res pons ibility for what appears on them, only an algorithm. T he hope s eems to be that the mos t M eta s ite will bec ome the mother of all bottlenec ks and rec eive infinite funding.

T hat new magnitude of M eta-nes s las ted only amonth. I n A pril, Kelly reviewed a s ite c alled "popurls " that aggregates c ons ens us Web filtering s ites ...and there was a new "mos t M eta". We now are reading what a c ollec tivity algorithm derives from what other c ollec tivity algorithms derived from what c ollec tives c hos e from what a population of mos tly amateur writers wrote anonymous ly.

I s "popurls " any good? I am writing this on M ay 2 7 , 2 0 0 6 . I n the las t few days an experimental approac h to diabetes management has been announc ed that might prevent nerve damage. T hat's huge news for tens of millions of A meric ans . I t is not mentioned on popurls . P opurls does c lue us in to this news : "Student s ets s imultaneous world ic e c ream-eating rec ord, wors t ever ic e c ream headac he." M ains tream news s ourc es all lead today with a s erious earthquake in Java. P opurls inc ludes a few mentions of the event, but they are buried within the aggregation of aggregate news s ites like Google N ews . T he reas on the quake appears on popurls at all c an be dis c overed only if you dig through all the aggregating layers to find the original s ourc es , whic h are thos e rare entries ac tually c reated by profes s ional writers and editors who s ign their names . But at the layer of popurls , the ic e c ream s tory and the Javanes e earthquake are at bes t equals , without c ontext or authors hip.

Kevin Kelly s ays of the "popurls " s ite, "T here's no better way to watc h the hive mind." But the hive mind is for the mos t part s tupid and boring. Why pay attention to it?

Readers of my previous rants will notic e a parallel between my dis c omfort with s o-c alled "A rtific ial I ntelligenc e" and the rac e to eras e pers onality and be mos t M eta. I n eac h c as e, there's a pres umption that s omething like a dis tinc t kin to individual human intelligenc e is either about to appear any minute, or has already appeared. T he problem with that pres umption is that people are all too willing to lower s tandards in order to make the purported newc omer appear s mart. Jus t as people are willing to bend over bac kwards and make thems elves s tupid in order to make an A I interfac e appear s mart (as happens when s omeone c an interac t with the notorious M ic ros oft paper c lip,) s o are they willing to bec ome unc ritic al and dim in order to make

5 of 11 01/27/2009 12:56 PM Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

M eta-aggregator s ites appear to be c oherent.

T here is a pedagogic al c onnec tion between the c ulture of A rtific ial I ntelligenc e and the s trange allure of anonymous c ollec tivis m online. Google's vas t s ervers and the Wikipedia are both mentioned frequently as being the s tartup memory for A rtific ial I ntelligenc es to c ome. Larry P age is quoted via a link pres ented to me by popurls this morning (who knows if it's ac c urate) as s pec ulating that an A I might appear within Google within a few years . George Dys on has wondered if s uc h an entity already exis ts on the N et, perhaps perc hed within Google. M y point here is not to argue about the exis tenc e of M etaphys ic al entities , but jus t to emphas ize how premature and dangerous it is to lower the expec tations we hold for individual human intellec ts .

T he beauty of the I nternet is that it c onnec ts people. T he value is in the other people. I f we s tart to believe that the I nternet its elf is an entity that has s omething to s ay, we're devaluing thos e people and making ours elves into idiots .

C ompounding the problem is that new bus ines s models for people who think and write have not appeared as quic kly as we all hoped. N ews papers , for ins tanc e, are on the whole fac ing a grim dec line as the I nternet takes over the feeding of c urious eyes that hover over morning c offee and even wors e, c las s ified ads . I n the new environment, Google N ews is for the moment better funded and enjoys a more s ec ure future than mos t of the rather s mall number of fine reporters around the world who ultimately c reate mos t of its c ontent. T he aggregator is ric her than the aggregated.

T he ques tion of new bus ines s models for c ontent c reators on the I nternet is a profound and diffic ult topic in its elf, but it mus t at leas t be pointed out that writing profes s ionally and well takes time and that mos t authors need to be paid to take that time. I n this regard, blogging is not writing. For example, it's eas y to be loved as a blogger. A ll you have to do is play to the c rowd. O r you c an flame the c rowd to get attention. N othing is wrong with either of thos e ac tivities . What I think of as real writing, however, writing meant to las t, is s omething els e. I t involves artic ulating a pers pec tive that is not jus t reac tive to yes terday's moves in a c onvers ation.

T he artific ial elevation of all things M eta is not c onfined to online c ulture. I t is having a profound influenc e on how dec is ions are made in A meric a.

What we are witnes s ing today is the alarming ris e of the fallac y of the infallible c ollec tive. N umerous elite organizations have been s wept off their feet by the idea. T hey are ins pired by the ris e of the Wikipedia, by the wealth of Google, and by the rus h of entrepreneurs to be the mos t M eta. Government agenc ies , top c orporate planning departments , and major univers ities have all gotten the bug.

A s a c ons ultant, I us ed to be as ked to tes t an idea or propos e a new one to s olve a problem. I n the las t c ouple of years I 've often been as ked to work quite differently. Y ou might find me and the other c ons ultants filling out s urvey forms or tweaking edits to a c ollec tive es s ay. I 'm s aying and doing muc h les s than I us ed to, even though I 'm s till being paid the s ame amount. M aybe I s houldn't c omplain, but the ac tions of big ins titutions do matter, and it's time to s peak out agains t the c ollec tivity fad that is upon us .

I t's not hard to s ee why the fallac y of c ollec tivis m has bec ome s o popular in big organizations : I f the princ iple is c orrec t, then individuals s hould not be

6 of 11 01/27/2009 12:56 PM Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

required to take on ris ks or res pons ibilities . We live in times of tremendous unc ertainties c oupled with infinite liability phobia, and we mus t func tion within ins titutions that are loyal to no exec utive, muc h les s to any lower level member. E very individual who is afraid to s ay the wrong thing within his or her organization is s afer when hiding behind a wiki or s ome other M eta aggregation ritual.

I 've partic ipated in a number of elite, well-paid wikis and M eta-s urveys lately and have had a c hanc e to obs erve the res ults . I have even been part of a wiki about wikis . What I 've s een is a los s of ins ight and s ubtlety, a dis regard for the nuanc es of c ons idered opinions , and an inc reas ed tendenc y to ens hrine the offic ial or normative beliefs of an organization. Why is n't everyone s c reaming about the rec ent epidemic of inappropriate us es of the c ollec tive? I t s eems to me the reas on is that bad old ideas look c onfus ingly fres h when they are pac kaged as tec hnology.

T he c ollec tive ris es around us in multifarious ways . What afflic ts big ins titutions als o afflic ts pop c ulture. For ins tanc e, it has bec ome notorious ly diffic ult to introduc e a new pop s tar in the mus ic bus ines s . E ven the mos t s uc c es s ful entrants have hardly ever made it pas t the firs t album in the las t dec ade or s o. T he exc eption is A meric an I dol. A s with the Wikipedia, there's nothing wrong with it. T he problem is its c entrality.

M ore people appear to vote in this pop c ompetition than in pres idential elec tions , and one reas on for this is the ins tant c onvenienc e of information tec hnology. T he c ollec tive c an vote by phone or by texting, and s ome vote more than onc e. T he c ollec tive is flattered and it res ponds . T he winners are likable, almos t by definition.

But John Lennon wouldn't have won. H e wouldn't have made it to the finals . O r if he had, he would have ended up a different s ort of pers on and artis t. T he s ame c ould be s aid about Jimi H endrix, E lvis , Joni M itc hell, Duke E llington, David Byrne, Grandmas ter Flas h, Bob Dylan (pleas e! ), and almos t anyone els e who has been vas tly influential in c reating pop mus ic .

A s below, s o above. The New York Times , of all plac es , has rec ently publis hed op-ed piec es s upporting the ps eudo-idea of intelligent des ign. T his is as tonis hing. The Times has bec ome the paper of averaging opinions . Something is los t when A meric an I dol bec omes a leader ins tead of a follower of pop mus ic . But when intelligent des ign s hares the s tage with real s c ienc e in the paper of rec ord, everything is los t.

H ow c ould the Times have fallen s o far? I don't know, but I would imagine the proc es s was s imilar to what I 've s een in the c ons ulting world of late. I t's s afer to be the aggregator of the c ollec tive. Y ou get to inc lude all s orts of material without c ommitting to anything. Y ou c an be s uperfic ially interes ting without having to worry about the pos s ibility of being wrong.

E xc ept when intelligent thought really matters . I n that c as e the average idea c an be quite wrong, and only the bes t ideas have las ting value. Sc ienc e is like that.

T he c ollec tive is n't always s tupid. I n s ome s pec ial c as es the c ollec tive c an be brilliant. For ins tanc e, there's a demons trative ritual often pres ented to inc oming s tudents at bus ines s s c hools . I n one vers ion of the ritual, a large jar of jellybeans is plac ed in the front of a c las s room. E ac h s tudent gues s es how many beans there are. While the gues s es vary widely, the average is

7 of 11 01/27/2009 12:56 PM Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

us ually ac c urate to an unc anny degree.

T his is an example of the s pec ial kind of intelligenc e offered by a c ollec tive. I t is that pec uliar trait that has been c elebrated as the "Wis dom of C rowds ," though I think the word "wis dom" is mis leading. I t is part of what makes A dam Smith's I nvis ible H and c lever, and is c onnec ted to the reas ons Google's page rank algorithms work. I t was long ago adapted to futuris m, where it was known as the Delphi tec hnique. T he phenomenon is real, and immens ely us eful.

But it is not infinitely us eful. T he c ollec tive c an be s tupid, too. Witnes s tulip c razes and s toc k bubbles . H ys teria over fic titious s atanic c ult c hild abduc tions . Y 2 K mania.

T he reas on the c ollec tive c an be valuable is prec is ely that its peaks of intelligenc e and s tupidity are not the s ame as the ones us ually dis played by individuals . Both kinds of intelligenc e are es s ential.

What makes a market work, for ins tanc e, is the marriage of c ollec tive and individual intelligenc e. A marketplac e c an't exis t only on the bas is of having pric es determined by c ompetition. I t als o needs entrepreneurs to c ome up with the produc ts that are c ompeting in the firs t plac e.

I n other words , c lever individuals , the heroes of the marketplac e, as k the ques tions whic h are ans wered by c ollec tive behavior. T hey put the jellybeans in the jar.

T here are c ertain types of ans wers that ought not be provided by an individual. When a government bureauc rat s ets a pric e, for ins tanc e, the res ult is often inferior to the ans wer that would c ome from a reas onably informed c ollec tive that is reas onably free of manipulation or runaway internal res onanc es . But when a c ollec tive des igns a produc t, you get des ign by c ommittee, whic h is a derogatory expres s ion for a reas on.

H ere I mus t take a moment to c omment on Linux and s imilar efforts . T he various formulations of "open" or "free" s oftware are different from the Wikipedia and the rac e to be mos t M eta in important ways . Linux programmers are not anonymous and in fac t pers onal glory is part of the motivational engine that keeps s uc h enterpris es in motion. But there are s imilarities , and the lac k of a c oherent voic e or des ign s ens ibility in an es thetic s ens e is one negative quality of both open s ourc e s oftware and the Wikipedia.

T hes e movements are at their mos t effic ient while building hidden information plumbing layers , s uc h as Web s ervers . T hey are hopeles s when it c omes to produc ing fine us er interfac es or us er experienc es . I f the c ode that ran the Wikipedia us er interfac e were as open as the c ontents of the entries , it would c hurn its elf into impenetrable muc k almos t immediately. T he c ollec tive is good at s olving problems whic h demand res ults that c an be evaluated by unc ontrovers ial performanc e parameters , but it is bad when tas te and judgment matter.

C ollec tives c an be jus t as s tupid as any individual, and in important c as es , s tupider. T he interes ting ques tion is whether it's pos s ible to map out where the one is s marter than the many.

T here is a lot of his tory to this topic , and varied dis c iplines have lots to s ay. H ere is a quic k pas s at where I think the boundary between effec tive

8 of 11 01/27/2009 12:56 PM Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

c ollec tive thought and nons ens e lies : T he c ollec tive is more likely to be s mart when it is n't defining its own ques tions , when the goodnes s of an ans wer c an be evaluated by a s imple res ult (s uc h as a s ingle numeric value,) and when the information s ys tem whic h informs the c ollec tive is filtered by a quality c ontrol mec hanis m that relies on individuals to a high degree. U nder thos e c irc ums tanc es , a c ollec tive c an be s marter than a pers on. Break any one of thos e c onditions and the c ollec tive bec omes unreliable or wors e.

M eanwhile, an individual bes t ac hieves optimal s tupidity on thos e rare oc c as ions when one is both given s ubs tantial powers and ins ulated from the res ults of his or her ac tions .

I f the above c riteria have any merit, then there is an unfortunate c onvergenc e. T he s etup for the mos t s tupid c ollec tive is als o the s etup for the mos t s tupid individuals .

E very authentic example of c ollec tive intelligenc e that I am aware of als o s hows how that c ollec tive was guided or ins pired by well-meaning individuals . T hes e people foc us ed the c ollec tive and in s ome c as es als o c orrec ted for s ome of the c ommon hive mind failure modes . T he balanc ing of influenc e between people and c ollec tives is the heart of the des ign of democ rac ies , s c ientific c ommunities , and many other long-s tanding projec ts . T here's a lot of experienc e out there to work with. A few of thes e old ideas provide interes ting new ways to approac h the ques tion of how to bes t us e the hive mind.

T he pre-I nternet world provides s ome great examples of how pers onality-bas ed quality c ontrol c an improve c ollec tive intelligenc e. For ins tanc e, an independent pres s provides tas ty news about politic ians by reporters with s trong voic es and reputations , like the Watergate reporting of Woodward and Berns tein. O ther writers provide produc t reviews , s uc h as Walt M os s berg in The Wall Street Journal and David P ogue in The New York Times . Suc h journalis ts inform the c ollec tive's determination of elec tion res ults and pric ing. Without an independent pres s , c ompos ed of heroic voic es , the c ollec tive bec omes s tupid and unreliable, as has been demons trated in many his toric al ins tanc es . (Rec ent events in A meric a have reflec ted the weakening of the pres s , in my opinion.)

Sc ientific c ommunities likewis e ac hieve quality through a c ooperative proc es s that inc ludes c hec ks and balanc es , and ultimately res ts on a foundation of goodwill and "blind" elitis m — blind in the s ens e that ideally anyone c an gain entry, but only on the bas is of a meritoc rac y. T he tenure s ys tem and many other as pec ts of the ac ademy are des igned to s upport the idea that individual s c holars matter, not jus t the proc es s or the c ollec tive.

A nother example: E ntrepreneurs aren't the only "heroes " of a marketplac e. T he role of a c entral bank in an ec onomy is not the s ame as that of a c ommunis t party offic ial in a c entrally planned ec onomy. E ven though s etting an interes t rate s ounds like the ans wering of a ques tion, it is really more like the as king of a ques tion. T he Fed as ks the market to ans wer the ques tion of how to bes t optimize for lowering inflation, for ins tanc e. While that might not be the ques tion everyone would want to have as ked, it is at leas t c oherent.

Y es , there have been plenty of s c andals in government, the ac ademy and in the pres s . N o mec hanis m is perfec t, but s till here we are, having benefited from all of thes e ins titutions . T here c ertainly have been plenty of bad reporters , s elf-deluded ac ademic s c ientis ts , inc ompetent bureauc rats , and

9 of 11 01/27/2009 12:56 PM Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

s o on. C an the hive mind help keep them in c hec k? T he ans wer provided by experiments in the pre-I nternet world is "yes ," but only provided s ome s ignal proc es s ing is plac ed in the loop.

Some of the regulating mec hanis ms for c ollec tives that have been mos t s uc c es s ful in the pre-I nternet world c an be unders tood in part as modulating the time domain. For ins tanc e, what if a c ollec tive moves too readily and quic kly, jittering ins tead of s ettling down to provide a s ingle ans wer? T his happens on the mos t ac tive Wikipedia entries , for example, and has als o been s een in s ome s pec ulation frenzies in open markets .

O ne s ervic e performed by repres entative democ rac y is low-pas s filtering. I magine the jittery s hifts that would take plac e if a wiki were put in c harge of writing laws . I t's a terrifying thing to c ons ider. Super-energized people would be s truggling to s hift the wording of the tax-c ode on a frantic , never-ending bas is . T he I nternet would be s wamped.

Suc h c haos c an be avoided in the s ame way it already is , albeit imperfec tly, by the s lower proc es s es of elec tions and c ourt proc eedings . T he c alming effec t of orderly democ rac y ac hieves more than jus t the s moothing out of peripatetic s truggles for c ons ens us . I t als o reduc es the potential for the c ollec tive to s uddenly jump into an over-exc ited s tate when too many rapid c hanges to ans wers c oinc ide in s uc h a way that they don't c anc el eac h other out. (T ec hnic al readers will rec ognize familiar princ iples in s ignal proc es s ing.)

T he Wikipedia has rec ently s lapped a c rude low pas s filter on the jitteries t entries , s uc h as "P res ident George W. Bus h." T here's now a limit to how often a partic ular pers on c an remove s omeone els e's text fragments . I s us pec t that this will eventually have to evolve into an approximate mirror of democ rac y as it was before the I nternet arrived.

T he revers e problem c an als o appear. T he hive mind c an be on the right trac k, but moving too s lowly. Sometimes c ollec tives would yield brilliant res ults given enough time but there is n't enough time. A problem like global warming would automatic ally be addres s ed eventually if the market had enough time to res pond to it, for ins tanc e. I ns uranc e rates would c limb, and s o on. A las , in this c as e there is n't enough time, bec aus e the market c onvers ation is s lowed down by the legac y effec t of exis ting inves tments . T herefore s ome other proc es s has to intervene, s uc h as politic s invoked by individuals .

A nother example of the s low hive problem: T here was a lot of tec hnology developed s lowly in the millennia before there was a c lear idea of how to be empiric al, how to have a peer reviewed tec hnic al literature and an educ ation bas ed on it, and before there was an effic ient market to determine the value of inventions . What is c ruc ial to notic e about modernity is that s truc ture and c ons traints were part of what s ped up the proc es s of tec hnologic al development, not jus t pure opennes s and c onc es s ions to the c ollec tive.

Let's s uppos e that the Wikipedia will indeed bec ome better in s ome ways , as is c laimed by the faithful, over a period of time. We might s till need s omething better s ooner.

Some wikitopians explic itly hope to s ee educ ation s ubs umed by wikis . I t is at leas t pos s ible that in the fairly near future enough c ommunic ation and educ ation will take plac e through anonymous I nternet aggregation that we c ould bec ome vulnerable to a s udden dangerous empowering of the hive

10 of 11 01/27/2009 12:56 PM Edge; DIGITAL MAOISM: The Hazards of the N... http://www.edge.org/3rd_culture/lanier06/l...

mind. H is tory has s hown us again and again that a hive mind is a c ruel idiot when it runs on autopilot. N as ty hive mind outburs ts have been flavored M aois t, Fas c is t, and religious , and thes e are only a s mall s ampling. I don't s ee why there c ouldn't be future s oc ial dis as ters that appear s uddenly under the c over of tec hnologic al utopianis m. I f wikis are to gain any more influenc e they ought to be improved by mec hanis ms like the ones that have worked tolerably well in the pre-I nternet world.

T he hive mind s hould be thought of as a tool. E mpowering the c ollec tive does not empower individuals — jus t the revers e is true. T here c an be us eful feedbac k loops s et up between individuals and the hive mind, but the hive mind is too c haotic to be fed bac k into its elf.

T hes e are jus t a few ideas about how to train a potentially dangerous c ollec tive and not let it get out of the yard. When there's a problem, you want it to bark but not bite you.

T he illus ion that what we already have is c los e to good enough, or that it is alive and will fix its elf, is the mos t dangerous illus ion of all. By avoiding that nons ens e, it ought to be pos s ible to find a humanis tic and prac tic al way to maximize value of the c ollec tive on the Web without turning ours elves into idiots . T he bes t guiding princ iple is to always c heris h individuals firs t.

Jaron Lanier is a film director. He writes a monthly column for Dis c over Magazine.

Jaron Lanier's Edge Bio P age

On "Digital Maoism: The Hazards of the New Online Collectivism" By Jaron Lanier

Res pons es to Lanier's es s ay from Douglas Rus hkoff, Q uentin H ardy, Y oc hai Benkler, C lay Shirky, C ory Doc torow, Kevin Kelly, E s ther Dys on, Larry Sanger, Fernanda V iegas & M artin Wattenberg, Jimmy Wales , George Dys on, Dan Gillmor, H oward Rheingold

[ P roc eed to the dis c us s s ion... ]

|T op|

11 of 11 01/27/2009 12:56 PM