sam palmisano: i am, quitehonestly, happy he's here. we are honored, quitehonestly, to have with us today. he's very, very busy. so please join me inwelcoming eric schmidt, chairman of google. eric schmidt: i'm delightedto be here. i started my computing careeron an ibm 360/91, which is a great machine.
water cooled, everythingyou wanted in a computer in the 1960s. when the ibm pc came outthat's what i had. now, of course, we uselots of think pads. and along the way i've alwayshad tremendous respect for the underlying architecture andtechnology folks at ibm. so one day sam calls and says,we'd like your thoughts about distributive computing. and i said, well, i thinkthis, this, this.
he said, he's a very, very niceguy, and he said, well, we've already done those. and all of a sudden i realizedthese guys are ahead of me. they've actually thought throughthe consequences of this new life thatwe're building. and we immediately organizedto partner with the technology, and thetechnologists, and the vision, from an extraordinarily wellled organization of ibm. so the reason i'm here is totalk a little bit about what
that means. what it means to us, to theworld, to you, to our joint customers, and so forth. and i wanted to start by talkingabout information. it is my personal belief thatthe role of information is under appreciated in politics,business, and in the world. the rate of innovation in theinternet is increasing, it's accelerating. we have huge new opportunities,for all of us
together, in mobile platforms,enterprise businesses, and in the role of informationand businesses. and this is a global audienceand, of course, ibm being the most global of all corporationswould have a very global reach. i like to think of four majorexperiments in innovation going on in the world. europe, the sound bites,educated, thoughtful, rule of law, empowering.
that's one model. the us, innovative, creative,surprising, individual, another model. china, monocultural, chaotic,controlling, confronting the individual truths,another model. a fourth model, india,following, accelerating, the merging of languagesand cultures. all of these are worthyexperiments in information. i also believe that the outcomeof each of those major
societies, which cover 80% or soof the world's population, will be determined by how theyembrace information, how they embrace the internet. how they use it, how they usethe tools that we're all together building, there are all sorts ofinteresting examples. rice is in short supply, whichis a terrible world crisis right now. one of the reasons rice is inshort supply is that the rice
farmers are using mobile phonesto discover the world price of rice and discoveringthat the middlemen are denying them their just profits. fisherman are on boats trying tofigure out which port to go into using, again, mobile dataconnections to say, can i get a price over here, or a betterprice over there. this is a profound change ineveryone's lives, and not just in corporations in buildings,but with people and the way they think about markets andinformation and so forth.
and we're on the edge of atruly great breakthrough. ibm has been working on this fora long time, much longer than google has existed. and ibm has figured out that theunderlying platform is the server, services platform, thatthey have literally been driving the modernization. i've been looking atbreakthroughs, and here's a quote for you. great breakthroughsare closer to what
happens in a flood plain. it's not one idea. a dozen separate tributariesconverge and the rising waters lift the genius high enough sothat he or she can see around the conceptual obstructionsof the age. so think of this as not asingle idea or a single company, but rather a set ofrivers that are flowing in which are getting stronger andstronger, and at some point it all coalesces into a model thatbecomes so obvious we
don't understand why wedidn't do it earlier. as a backdrop, the internet,of course, is growing very quickly. 1.3 billion users, on the orderof 200 million new users per year dependingon who you ask. and the underlying phenomenahere is what i'm going to call the technology basecase, moore's law. looks like we've have another10, 15 years of moore's law, cpu performance essentiallydoubling every 18 months.
even more frightening, from myperspective, is a term called crider's law, which says thatdisk drives, literally the disk capacity, is doublingevery 12 months. this is particularly good foribm's revenues, by the way, because they're going to bethe ones putting all this stuff together. and if you think about thosetwo compounding you have amazing consequences,architecturally, that will affect all of us in many,many different ways.
ibm, by the way, is the authorof exabyte, petabyte, [? xylobyte, ?] so forth and so on, all of thenew things, because ibm had to invent the names for theconcentration of storage that are projected over thenext 5 or 10 years. so this, the open approach,the open standard approach that the internet has pioneered,really has allowed us to go from the four networks,including one here in los angeles in 1967, to themore than 500 million networks
that comprise theinternet today. you go from 400 serversin 1983, 500 million servers in july 2007. that scaling is phenomenal. and that scaling occurs becauseyou have one good thing in technology. it can really replicatevery, very quickly. so let's do a littlebit of fun. if you assume current growthrates, which is a reasonable
assumption based in the last10 or 20 years, by the year 2019 it'll be possible to havea device that's on your belt, or in your purse, that will have85 years of video on it. so you'll never be ableto watch all of it. you'll be dead beforeyou've finished. to me this is the ultimatefrustration device. i really wanted that show,i really wanted to do that, but i'm done. or i was, my parents wouldn'tlet me watch it now, when i
was young, but now i'molder and i'm not interested in it anymore. you can imagine allthese dialogues. here's another example. 81% of the internet users injapan access the web from their mobile phones. and just to prove that thejapanese are really wacky, the three top books in 2007 in japanwere first delivered on mobile readers, mobile devices,and were only
subsequently printedin real bookstores. showing, once again, thejapanese are ahead of us in everything. so is this our future? maybe, especially if you spenda lot of time on subways. my point is that the technologyenables some things which are surprising. americans send 1.6 billiontext messages per day. god knows what they're about.
i never really know. in any case, we have theemergence of this massive shift of computing led by ibm,google, and a few others. it's truly a shift from thedesktop centric model to a network centric model. you need the server, you needthe desktops, and you need the network in between. we call it cloud computing,there are other terms for it, but to me this is the storyof our lifetime.
it is the story that will defineour careers and what we do, because this is going togo on for a long time. now how many devices are youall carrying right now? how many chargers did you bringon this trip, wherever you came from? how many remote controls doyou have sitting in your television room that you cannotquite figure how to work all together? that's the consequence of beinghalf way through this.
but eventually all that stuffshould be on the network. it should all be seamless,it all should just work. now you sit there and you go,well, why do i have all these different devices? well, it's combinatoric. the vendors, the manufacturers,build a device, but they can't make surethat everything works. it's an end to end problem. so again, if they have astandard, if they have a way
in which the network works, thenthey can just plug into the network and everythingworks just fine. and, again, you say, what'sthe cost of this? there have been a lot ofstudies about this. the estimate is that 6 out of100 products are returned simply because consumers can'tfigure out a way to work them. think about the efficiency, thewaste and so forth, from this design issues. so we do know that convergencewill occur in this cloud of
servers that is being builtby ibm and others. and the original termcame from a fellow named mark weiser. he called it ubiquitouscomputing, but it's the same principle. and you really do want to beable to watch things on any device, see informationanywhere, and so forth. here's an example. wouldn't you want to be able todo a map search on your pc
and have the same thingon your mobile phone? well, you can do that now. you couldn't do thatfive years ago. and in fact, the interestingterm is everybody kept saying about convergence. people concluded thatconvergence meant that everything went to one device. exactly wrong. convergence is everything endsup on one server, or set of
services in the cloud,and there are many different devices. and excuse the analogy,but it's the best i could come up with. women have multiple purses,and they spend their time, they have this purse for this,this purse for this, this purse for this, and then theymove their contents from one purse to the other. so the analogy here is thatyou have a set of devices,
each of which is very different,it makes perfect sense to you. you use this device here, andthis device here, and this device here, and they're quitedifferent from each other. but every time you turn oneon it has exactly the same contents as the other one. this is a much more profoundchange than you might have originally thought when youthought, well cloud computing sounds pretty interesting.
so when you think about it,and it depends on the generation. with our young employees atgoogle, i was in a meeting yesterday and said, how many ofyou have your friends and your photos and so forth allon a social network? every hand goes up. i suspect that wouldn't betrue in this audience. it's certainly not true of me. my hand was not up.
how many of you put allof your photos on photo sharing sites? google offers one. again, i suspect not likelyfor this audience. it's certainly reallynot true of me. so not only is this visioncorrect, but people of a certain age viewit as obvious. in other words, eric, why areyou giving this speech? that's how profound this changeis, and how important
it is for us to understand. so in this, you could think ofit as the electricity model, computing resources shiftingto electricity. 100 years ago people generatedtheir own electric power. eventually they started to relyon a grid, which is, in fact, remarkably reliable. and basically thesame thing has occurred around web computing. an example, we brought out aproduct called google sites in
march, which is basically basedon a company we acquired called jotspot. what you do is you plug andplay, you build your site, you press the button, and there itis, immediate authoring. it makes perfect sense if youthink about it, because everybody has to have a webpresence and it's all hosted on the google infrastructure. so the cost of innovation hasnever been lower, which means there'll be even moreentrepreneurs, even more
opportunities for you all to gohelp those innovators learn how to use this new model, whichmeans, of course, we all jointly need to understand itand understand it's both limits as well as scale. so to me, the consumers aredriving this and corporations are caught. and the corporations are caughtbecause consumers are demanding it and thecorporations are not operating quickly enough.
they haven't yet mastered howquickly to turn key things. ibm has done a particularly goodjob in the application space of trying to do that. there are other companies aswell, but it's very, very important that our corporatecompanies, many of whom you serve or represent, understandthat the consumer expectation is very, very high on thesesorts of issues. so you can do lots ofinteresting things. the consumer statisticsare quite interesting.
there are more than 70 millionblogs, more than 120 million blogs that are created daily. i looked at the average, andit appears that the average blog has exactly one reader,which is the writer. so not to worry. i would argue maybe it's two,the reader and the reader's mother, in honor ofmother's day. but nevertheless this phenomena is certainly growing.
the statistics are really quite alarming, at least to me. 700 million photos are uploadedto our picasa website photo sharing, which is justone of them, every day. and youtube is particularlydisturbing. there are 10 hours ofvideo uploaded to youtube every minute. so that's the scale of this. this is no longer web sites withtext and funny pictures.
this is integrated storytellingwith video and authoring and so forth to causesome consumer behavior, whether it's information oreducation or games or so forth and so on. the important thing here isthat this cloud computing model really does matterto enterprises. it departments need to focuson productivity. google and ibm have a set ofapplications which are web hosted which allow you to,essentially, let somebody run
it for you. if you look at thevirtualization model that ibm has, they got the model right. it's just much better if,architecturally, you could virtualize those services andlet a set of professionals write it so you canget back to work. the same principle appliesfor email, web authoring, underlying execution, and allthe vm work that they, literally, invented.
so there are many examplesof this. google is a very strange places,as you can imagine. we have meetings whereno one looks at each other, it's a rule. instead everyone looksat their computers. and we're talking while everyoneis typing and editing all the documents, and allof our products are sharable on the net. so one of the key componentsof this new network
architecture is sharing. so all of a sudden you discover,wow, that changed, this changed, and soforth and so on. my talks are written by peoplein some random order and then i have to look at themand say, god, what were they thinking? our business plans arewritten this way. it's just phenomenal, and it'sat least creative, if a little bit out of control.
this decision making processis also very different. you can now ask your customersthings that you couldn't before. you can trial a product andyou'll get the feedback immediately. all of a sudden thewhole product lifestyle is so much quicker. so all of a sudden the notion ofa three year plan and a one year plan and so forth can bethrown out of the window by
new technologies where youcan literally survey your customers the moment one comesout, and you can make the changes along the fly. so if you think about it, it'sno longer sets of separate applications, it'sintegrated apps. for example, you can do chat inour google docs, which is our web hosting services. you can chat inside thedocument all the time. there are things called googlegadgets, and i'll give you
does a search. so all of a sudden all of thatcorpus of information is now embeddable inside of a spreadsheet, and the spreadsheet auto loads and does all theright obvious things. what about calendars? anybody here think calendarsare boring? i've always thought calendarswere really boring. well, they're no longer boring,because not only can you cross link calendars, butwe can search all the
activities that you care aboutand dynamically update your calendar and then you'lldiscover you really are out of control and you really areoverloaded and there really are too many things to do in losangeles while you're here. my point here is that thecomputer can help because of sharing, and because of thedynamic nature of this new web model, and the kinds ofapplications you can make are very interesting. i have lots of exampleshere, flash animation
of data over time. somebody decided thatthey were upset. in switzerland everythingruns exactly on time. so the trains havegpss on them. so we give you a down to themillisecond time estimate of when the train will show up inthe train station and exactly where it is on google maps allthroughout switzerland. most other countries don't payattention that much, they're not that accurate, but in theswiss it's down to the
millisecond, done as a mash upon top of google earth and google maps by a programmerwho had nothing else to do while he was waiting, literally,for the train and annoyed because it was late. we have a platform calledandroid, which is the expression of all this in mobilephones, which is in the process of being delivered. and part of it is it'sa developer. and somebody said, what i'mgoing to do is i'm going to
take my gps location, andremember phones are very interesting. phones have cpus, memory,they have cameras, and they have gpss. so what this person did is hewrote an application which says where am i, what am ilooking at, and then let me tell you what you're seeing. so you can put in a mode whereit says not only this is what i see through my photo lens asa camera, but also i'll tell
you what everything is. because i know the landmarksbecause i'm integrated with maps. it can also, by the way,tell you where to go. you're lost, turn left,turn right. but it's a very different way ofvisualizing where you are. another example which isparticularly scary, there's something calledfriend finder. and what happensis, you have to
choose this, thank goodness. you choose to let other peopleknow where you are and it says where the other people are andhow to get there fastest. so this thing pops up and it saysall my friends are here, turn left for this, turnright for that. again, the off button hereis very important. and it shows you the path, showsyou the visual, shows you a picture of the person. this is this new geocentric,application centric, cloud
computing model before us. now ibm, of course, is busybuilding this architecture inside of a whole bunchof its products. partly because ibm was soforward looking, and it's going to linux, so forwardlooking, and it's coming to java. and all of the web services now,they have a head start. the websphere portalproduct is one. for example, you can basically,using websphere,
you build these web sites,you can embed all these interesting gadgets that i'mtalking about right in your application. now you sit there and you go,what would i use this for if you're doing somethingin manufacturing? well it turns out that gpscoordinates and sharing apply to simple consumer applications,but they provide real value in a businesscontext. that's the secret of ourcollaboration here.
and you can do fun stuff likethe weather, but if you're in a shipping company the weatheractually matters. in the case of lotus noteswe did the same thing. the email reading structurewithin lotus notes, we can now embed the sharing and gadgettags that i described earlier. so the point here is, anda lot of us are jointly developing applications forenterprises, that there's a new layer here. remember i talked about as abreakthrough, and this new
layer is both exciting, whati've done for my whole career and we're finallygetting there. it's just phenomenal. but it's also challenging fordevelopers, because if you think about it, the developernow has to think about not just the pc or the main frameor the unix server, but they have to think about buildingthings for internet enabled devices, game consoles, mobilephones, devices, music players, cars, appliances,and so forth and so on.
you're going to have 500million wi-fi chipsets sold in next year. think about the wi-fienablement of all of these devices. just remember that that settop box on top of your television is, in fact, acomputer with an ip address. people can talk to it, and theycan start having some fun with your television. so my point here is that thearchitecture that's being
built here has to enablemore than just one platform being built. and so developers have tobasically build an application that really anticipates these extraordinarily large data sets. again, go through the numbers,the amount of video, the amount of data. just think of the scale that adeveloper has to deal with. and this is true of yourdevelopers, the people who
work for you as well. and they have to basically bewilling to execute algorithms across computing clusters. they have to be able to writein languages like python and other sort of scripting. they have to have distributivedata storage, something which ibm literally invented. you have to have distributivedata storage where you can actually draw that and theinformation isn't just on one
server but on multiple servers,and you have to put it all into a single platform. and of course this all has tobe available through the mechanisms of the internet. so what happened after samcalled me and said, let's talk, we sat down and we said,what's the most important thing we need to do right now,between the two companies, to accelerate this? and they came back and theysaid, nobody understands this,
but scientists, basically innew york, the scientists at google, and a few otheruniversities, why don't we build a version of thisarchitecture, this is the clever part, and giveit to science? what we'll do is we'll give itto all of the universities, literally worldwide by theway, not just in the us. and we'll give it to thembecause, well it's obviously the right thing to do, pluslet's be blunt, it'll help ibm and google because all thosepeople will get out there and
they'll be able to program andbuild against this new model. so what happened, of course, asa result is that we, after a lot of joint scientific work,we announced an academic cloud computing initiativelast october. and what it does is it takesopen source, literally ibm and google do not own any of this,and it builds a set of data centers that are for the publicthat actually implement this whole computing model. and by the way, itdoes it at scale.
so we started at university ofwashington, we did carnegie mellon, mit, stanford, the usualsuspects, university of california berkeley, and theuniversity of maryland. we also then immediately wentto shinhwa university in beijing and are now going globalwith a whole bunch of other universities. and in fact, what's interestingabout the collaboration, google and ibmhardware and software architecture and tivoli as thenetwork management council
makes perfect sense. so what are some of theinteresting results of this? somebody figured out a wayto build, a 19 year old, literally, physicist, figuredout a way to simulate collisions related to the bigbang that could never have been done before because of thecomputing research that was put out literally throughthis mechanism. another showing you thatintelligence is distributed everywhere, a european 18 yearold developed the largest
collection of prime numbersever, using the same infrastructure, sitting in hishome in eastern germany. this scenario has been sopopular that actually it got out of control. and so we eventually asked thenational science foundation to run this and to actually figureout there are so many interesting applications forthis, we literally couldn't pick the winners anymore. it gives you a sense of thepower of the model that i'm
talking about, and the sensethat we're right on the edge of this taking overalmost everything. and i'm very, very proud of whatwe've done together here to provide for the publicsomething that can actually work on climate modeling, genesequencing, protein mapping, answering the fundamentalquestions of whether the earth is going to be around or arewe going to go kaput. the fundamental issues aboutcures for cancer and other sorts of diseases, all issuesabout viral and protein
mapping and so forth that arenow possible at this level, and you say, well, those aregreat, but how does that translate into my business? it's the same platform. that's what's so interesting. we get such leverage. so when i think about it, and ithink about my history with the ibm 360/91, which was agreat machine at the time, fast loading point processor,one megabyte, was this big,
and i think about wherewe've come today. the key idea here is thatcomputers are not alone. that computers actin networks. that if you can get thearchitecture right, this cloud computing architectures, if youcan get the development layers right you can buildthese platforms. and the platforms are morerobust than the one you did yourself, because youget the sharing. they're architected in a waythat provide tremendous
customer value. and they literally change theworld, and that's what's so exciting to be here, to do whati do, to do it with sam and his team, and franklyto be here today. so with that thankyou very much. sam palmisano: it's justyou in charge. male speaker: yes, so why don'twe make the most of the time we have and get startedright away if we could. and i was very intrigued by the
descriptions of cloud computing. and in addition to the specificapplications it seems to me that there's a new modelof collaboration here. it's different from the oldwe'll develop windows and then people will write softwareapplications for windows. it's different from linux,which is a non commercial entity trying to makethings available for the world at large. so one of the intriguingquestions that was in my mind
as i heard these descriptionsof this cloud computing initiative is how doescollaboration work when there are two organizations both withsignificant resources, somewhat different commercialinterests, trying to pool their efforts to get somethinglike this done? sam palmisano: i'll start. i find it interesting, eric'sheard this story, about how we got to me callingeric one day. it'll get a little bitto your point.
we have a meeting once a yearat ibm where we argue about the future of technology. research comes in and takesa provocative view and the businesses defend, becausethey're obviously grounded in operating income statements andresearch tends to be much more theoretical andfarther out. and we had this big debateabout the cloud. is it real, is it not real,we're doing all this work and research, we have all of thesetech capabilities that eric
has described. and so we had this big debate. and we did it differently thanyou do it, eric, because at ibm they look at each otherand yell at each other. so they don't look at theirmachines, they've got to look at their face. so there's one of their face. might have been better tolook at their screen. so i remember asking mycolleagues the question.
i said, well, are we better offwith working with someone like google or not? we're not really competitors,but it's different model, different business structurethan ours, what do you think? and they had a long discussionthat we all concluded that we should just call and see ifthey'd be interested in talking with us. and you'd say, well, what ledus to this conclusion? that's what you're looking for ithink, the logic that led us
to this conclusion. we really weren't lookingto sell google anything. that wasn't the thought, but thethought was if you saw the future as eric has describedit very, very well, and i won't waste the audience'stime redescribing it. but if you really believe thatthat was a future state and you believe that it wouldn'tbe everything, it never is, there'll always be other sortsof things that go on in this computing world thatwe live in.
and there's a still a lot ofbig large mainframes doing banking transactionsevery day. but this would compliment,if you believed it would compliment, who would be thebest company to partner with to get something like thisat least established? who would that be? and as we thought about it,and a little bit of the comments i made, it had to besomeone committed to the standards, clearly.
common points of view of thefuture that had to align or it wasn't going to work. and the other thing for us isa culture where the teams would come togetherand work together. because the engineers, at theend of the day, as much as eric and i might encouragethem, at the end if the engineers don't believe in thisstuff, i think in both cultures, i know in the ibmculture, it just won't go. the engineers have to believe,the scientists have to
believe, or it's very hard tomove them down this dimension. so that got us there, commonpoint of view of the future, committed to standards andopen architectures. actually, eric was very kind tous in the conversation, it was google's idea to the bestway to get this going was through the academiccommunity. that was really a google idea. eric schmidt: we thinkit was your idea. sam palmisano: historianswill write it was ibm.
perhaps you're beingkind to us. i'm going to pass to eric in asecond, because remember, you were saying that we have to getpeople to understand it and train and develop on it. and how can we do that? and maybe ibm responded bysaying, well, we can get these academic initiatives goingbecause we do these things all the time anyway, as you do. but i think it was you, youframed the problem that day on
the white board. you're going to frame thisproblem with me that day. that led to this next step. so it was that kind of a modelthat led to the collaboration. and i'll pass it tomy colleague. eric schmidt: i was going to saythat we've talked to other companies, and why did this onework and none of the other ones worked, i think isthe real question. and all innovationcomes bottoms up.
and so it's important thatyou have two things. first you have to have reallysmart people who are actually doing the work, which ibmclearly does and hopefully we do as well. and they have to like oneanother, that was easy. and the second thingis there has to be a compatible culture. and it turns out that theculture that ibm had set, surprisingly to us, based onprejudice and so forth, and
not really knowing it, wasreally quite compatible with trying to truly changethe innovation model. and that's because theleadership let it happen. other companies will have verysmart engineers, but the leadership is either suspiciousor they send large numbers of lawyers to visitus, or so forth and so on. and everything bogs down. sam palmisano: we have lotsof lawyers, though. male speaker: so asan academic--
sam palmisano: having just leftour annual shareholder meeting i've had a lot ofexperience with lawyers this week, eric, so i havea bias right now. male speaker: so as an academic,first of all it's great to hear that academiccustomers are considered desirable for certainpurposes. and the second thingi understand-- sam palmisano: exploitthem if they are. eric schmidt: it'sinteresting.
ibm is, again, much too modest.ibm has always been building specialized hardwarefor universities. sam palmisano: yeah,we have for years. eric schmidt: and so forexample, there was a chess challenge which ibm endedup winning big time. and ibm was willingto build the very best and fastest hardware. i have no idea why that was agood business for you sam, but it was really good from atechnological perspective.
sam palmisano: i thinkit's likely the parallelism of the cloud. we did it in this kind of aserver, it was a unix server, against kasparov. but the logicof it was, and if you think about it, it'sa blue gene. what we do in these massivesupercomputers, protein folding, we simulate the nucleararsenal, et cetera, and those kinds of capabilities,modeled, weather, what have you.
but the reason why we doit is for two reasons. one is, we believe, toadvance the science. why is it important for ibmto advance the science? not because we've always doneit, but beyond that because we do spend a lot of money doingit, we obviously want to get our returns. but we feel we're better offbeing in the forefront of advancing the science, or thestate of the art, or the physics, or the computing model,or the architectures,
in an open standards way thantrailing, in that debate. now, quite honestly, we wouldprefer to have others with us. universities are key to this,governments are key to this. there aren't as many commercialit companies anymore who really havefundamental research left. they've all gone down the modelof very oriented towards wall street, very shortterm oriented. a pc distributor, can pick, onehas 1% in r and d. we have 7 billion a year.
so it's a completely differentbusiness model, but fundamentally we've come to theconclusion that over time we're better off advancing thefield of these competing architectures, and then if we'resuccessful and get it right we get the economicbenefits associated, versus trailing and havingsomeone else lead. so we've always donethis in ibm. if we didn't the scientistswould leave. they'd go to google.
we have 205,000 engineers andscientists, and i'm convinced if we didn't give them theopportunity to create the future the people we've hiredwould find some place to go create the future, becausethat's the individual that's attracted to a placelike that. male speaker: so some of itseems to be the compatibility between engineering orientedcultures, values, interested in pushing things a little bitpast where purely commercial short term optimizationwould take you.
still, those seem likenecessary, but not sufficient, conditions given the differencesin physiognomy that sam was describing duringthe course of his talk. so what other lessons can youoffer the audience, me, in terms of how one makescollaboration work, even when the conditions on the groundare fertile, which is obviously a prerequisite foranything like this to work? sam palmisano: ericmade it real. he's very, very modest, but hegave both teams a challenge to
come back with a prototype. and mapping and shoppingwas the example as i recall, right eric? he defined the challenge to bothteams to come together, work collaboratively on it. he said, this the challenge,come back. he gave the team eight weeks,we were together. but i think because he had thevision of articulating that challenge it did acouple things.
one, it was a hard technicaltask to solve the problem, but it also congealed the teams.they got to know each other, they got to respect each other'scapabilities, they got to learn from each other. and to me, my observations oncewe came back with these prototypes, operational, infact you had to go off and meet with a political leaderthat day, as i recall. you had to leave the meetinga little early. but the teams came together.
they had a lot of respectfor each other. technically they had respect,and then quite honestly, i think we just got out ofthe way and it ran. we didn't have to do muchof anything after that. eric schmidt: i thinksometimes you have to show the future. so i'll be talking to somebodyand people are always very polite and they'll go,yes, yes, yes. and i discover they haven'treally used the internet very
much, they haven't really usedsocial networks very much. so everybody wants to look likethey're real experts, but in fact you have to be embeddedin it to really understand its implicationin your business. so i try to use everything allthe time, even if i hate it. and at my age a fair amountof this stuff makes no sense to me. so nevertheless i try toforce myself to at least understand it.
for people in the audience,you have got to immerse yourself in it. you have to to use your product,you have to use the e-commerce site, you have to buythe product, you have to understand it, you haveto next to it. and if you're not doing thecalls you have to sit next to the person in the call centerand listen to them and experience that with them. you're not going to get thebusiness insights without
doing that. it's really about immersion. male speaker: so just continuingon with cloud computing for just a minuteor two longer. i was wondering, one of thethings that i thought i'd heard was that this may lead toa convergence between the enterprise space andthe consumer space. now, historically ibm andgoogle have focused on largely, not entirely,on different
portions of that space. how did you think about thispossibility of intersecting worlds as you got intothis collaboration? sam palmisano: well, we thinkabout it two different ways. what happens is that we'velearned that many of these new technologies start in theconsumer space earliest, because there's lessrisk to deployment. eric wasn't here this morning,but you heard the cio panel talking about risk of deploymentwith all the things
they have to keep buttoneddown, sarbanes-oxley compliance and allthese things. they're very nervous aboutexperimenting too much with new technologies. so you see a lot of thesethings emerge. the internet emergedin the academic, in the consumer space. some might argue that the $1billion we put behind it made it ready for the enterprise,which was a rather modest
investment, but was viewed aslarge at that point in time. so we've always been attentiveto these sorts of things and try to see if they extendand they can apply. and that's a technologypoint of view. and then there's obviously thebusiness model to mention there as well. but in the technology point ofview, we're looking at this and saying, ok, is therea logical way that an enterprise, as they try to getscale in virtualization, to do
it in their world, we call itbehind the firewall, but in a secure way that they can protectthe transaction, fill out the information identity,et cetera, et cetera? you're not doing the publicnetworks, doing other things you aren't as sensitive to,is there a way to do that? so that was observationaltechnology. how does it apply? and quite honestly, as you lookat this little platform we're now announcing thathopefully our partners will
put their applicationmarketplace on, it is fundamentally a cloud. we will operate the cloud. it is what it is. we've learned a lottogether here. we'll make it available to ourpartners and hopefully they can sell on top of thattheir various kinds of services and offerings. so to us it's the firstinstantiation of taking
something that was prettymuch consumer. we're going to use own economicpurposes with our partners, and also looking atareas, especially a lot of these large enterprises havethousands and thousands of these traffic based servicesjust throwing things at a back end. and is there a way to applythe technology there. so that was how we thought aboutit, as a place to learn that we could thenapply elsewhere.
eric schmidt: my own predictionis that there will not be that much differencebetween the consumer architectures and the enterprisearchitectures. there's one fundamentaldifference between enterprise and consumers, which is that theenterprise customer will pay for things that the consumerwill not, and the enterprise customer has a veryreal need for greater security and greater qualityof service. and that you can build abusiness where you charge for
those and they're willing topay for it and it's a good business decision for them. and again, if i go back to thevirtualization that ibm did starting 15 years ago or so totraditional data centers done off site, you see that ibm hasfigured out a way to do this. google is also trying to dothis with some lower level services where, again, they'refree for unsupported but there's a modest feefor supported use. and that model seemsto be coming along.
i think it's a new modelfor the enterprise. male speaker: so we've talkeda little bit about the enterprise versus consumerworlds, now to switch to another kind of world. we spent some of the morningtalking about globalization, and i wanted to explore that setof issues as well, because one of sam's themes during histalk was accelerating global integration. on the other hand, we knowthat there continue to be
differences across countries. so what's ibm figured out abouthow you simultaneously meet the challenge of beinglocally responsive and relevant versus achieving thekinds of economies of scale that allow you to do better thanyour local competitors? sam palmisano: it's the businessanalogy to eric's comments on these computingarchitectures. and so what do i mean by that? well, the cloud ofthis is process.
all the back of ibm isglobally integrated. it runs for the entire world,accounting, hr, procurements, supply, all that stuff. we don't do it by regionanymore, we do it for the world, and that's howwe run ourselves. and we did it for obviousefficiency purposes, but we've been years doing it. we have a long way to go,but it's been very productive for us.
why is that important? it's important because then youcan spend your time, and we're going into these100 new countries. we've been in them for along time, but we're investing in them. they don't have to dealwith all that anymore. so now when we access both theorganization, probably at the end of the day we'll roll it outin july, 50, 60 countries in it today.
they've all been growingsomewhere between 10%, 30%, some 50%. some are big now, likechina and india. i know russia's small, 300 or400 million, maybe they'll sway to 600 million,so that's not big. others are billions,middle east. but the point of it is what itallows you to do is that those people just worry about whatit takes to serve the front end, the client.
what skills do you need? what partners do we need? because we distribute with alot of people in this room now, it's not just purely ibm. what are those keyrelationships? what government contextdo you have to have? what university relationships doyou have to have so you're viewed as a valued citizenin all those places? but the cloud takescare of the cloud.
you don't worry about an auditor a payable or a receivable and all that treasuryand all that stuff. and so as a result, what itallows you to do is scale a lot faster. it's a business modelthat we talked. eric schmidt: you haven't evenmentioned that you're doing all of our accountingand payables. and you refuse to tell me whichcountry you're doing them in, and i don't care.
i think it's india but i'mnot really quite sure. sam palmisano: i will tell youthis, the best audit we've ever had was in kuala lumpur. kuala lumpur for payableswe received the best ever in the world. eric schmidt: i'm pleased toreport to you that we're still happy with the dealwe did with you. sam palmisano: thankyou very much. so the other side of it is toyour point, the cultural.
so that's the businessmodel side of it. the cultural side of it is youhave to deal with each of these countries differently. you have to be able to do thatbased upon where they are on their societal agendas. and so we have these programsin china, add india, we just did vietnam out, etcetera, et cetera. we're doing middle east out. but we look at what are the mostimportant things for that
society on the ground. and it varies over stages oftheir economic development. and then based upon thatwe try to connect in. but again, we don't have thebenefit of a consumer brand. so we can't go watch a think padin all those countries, or whatever, anymore. so we try to connect withsocietal things, it's very, very important. and then embed ourselves inthat economic expansion.
part of the lenovo deal wasour ability to create a chinese global company thatcan be a success story for china, and it has been. the government's very happy. well, because they're veryhappy, obviously other good things happen forus as a result. so it's this localunderstanding, connecting to the societies, requirements thatpeople do on the ground. they're not burdened with theoverhead of the past. they
don't have to worry abouttreasury and data centers and education centers, and allthat stuff we had to put in 50 years ago. it's in the business processcloud with google and a lot of other people in that cloud doingother kinds of work. eric schmidt: i was going tooffer a simple rule, that people around the worldare exactly the same. they want exactlythe same stuff. they care about their families,they want better
health care, they want safety,they want lower taxes, they want cars, and soforth and so on. so once you understand that,that all of the six billion people inhabiting our fragileearth here, and growing precipitously, they allwant the same things. the only differenceis language and some aspects of culture. so once you understand that youunderstand that they are, as a group, going to want allthe crazy stuff that we have
invented or the thingsthat we've pioneered. they're going to want carsand they're going to want computers and they're going towant to do stupid stuff the photos and use youtube andso forth and so on. so subject to language, weapproach the globe as people are pretty much the same. now the societies differ, buttheir ultimate driving, the individual, very similar. male speaker: so let me justask a follow up on that,
because my impressionis that, say take something like orchid. does well in iran and brazil,and i've spent some time trying to figure out thecultural commonalities between iran and brazil. eric schmidt: and india. male speaker: and india. doesn't have quite those marketshares elsewhere. so what would accountfor that variation?
eric schmidt: we don'treally know. it's an interesting question. for those of you who don't know,social networks have evolved with leadersin each country. and social networks tend to benetwork affect businesses, because if all of your friendsare on one network you're likely to join that networkas opposed to another. my own view on social networks,and this is controversial, is that theyare still too closed.
and that the internetsays that all societal activity is open. and that eventually thesenetworks will become much more open and people will move inand out of the networks. so we don't really know why asocial network takes over in one country or another, but theimportant thing is that to say about those social networksis that they're real. a lot of people, especially ifyou have high school kids or their college kids, you assumethis is some toy that they're
playing with. certainly all my friendsthink that. but these are real phenomenonof people living their lives online, and it has legs. we're going to have to dealwith it as a society. male speaker: well, the clockreads zero, which i take as my cue for trying towrap things up. so it is very interesting thatwe started off by talking about network computing andare back to talking about
networks in some sense. and i think that's a little bitof a reminder of how the traditional monadic,monolithic view of enterprises, and of individuals,is something that we've gone past. and why agathering like this is so important to, obviously, ibm,but i assume also to all the partners in this audience. so with that, thankyou very much. eric schmidt: thank you.