a computer that has no brain. its power is in the cloud. peno-logic ceo john tish working to make pcs dumber. crowd sourcing with one of napster's earliest ceos. clouds and crowds this week with our reporter john schwartz of usa today, john schwartz of fortune magazine, this week on "press here." good morning. i am tempted to show i a tape of ants working together in some

common cause because that's what we're going to start with. the power of the crowd. not a crowd of ants, of course, but a crowd of people. people all over the world working together on a big project. crowd sourcing. what usc needed to find thousands of lost alumni, it used crowd-sourcing, hiring thousands of people to search for former students on-line. and where do you find a temporary work force of thousands of people all willing to take on a small job? well, the internet, of course, through services like cloud crowd and employment agency. you come to it with a massive task. say, translate a popular book into hindi. cloud crowd offers the task to thousands of people. each may just get a few pages, but the entire book is

translated in minutes. jordan ritter is one of founders of cloud crowd. he knows the power of the many because he is also one of the earliest developers of the most disruptive services on the internet ever, napster. joined by my colleague john port of fortune, john schwartz of usa today. we will get to your start-up in aç few minutes, but it's not every day that you get one of the early engineers of napster. the number three guy in the company. where did you fall? >> well, i was one of the guys working on the back end with sean before it was even a company, so depending on how you count, i'm actually the second guy. >> and the guy who built the -- some of the original software that made the whole thing go? >> yeah. originally sean wrote -- back when it was called music net. some people don't know that. it was sort of a side project, and he had written both the windows client and the service software, but sean was 16, and

he was being taught by some smart friends, but it was not stayable. it wasn't working well. very shortly thereafter he became very interested in the windows side of things, and he needed somebody to take over the back end, so it was back again before it was even a company, before we knew we were raising money, before it became what people know it now as napster. i was the back end guy. >> what did you think you were building? >> to be frank it all originated -- a lot of it originated in the underground hacker security subculture. there were -- >> not reality -- >> the old school version back before we had all these im clients, and it was just -- we were part of a group that was about 60 people strong, and we all had our own professional jobs or we were enthusiasts. whenever we had a problem, we would solicit the group for help.

there was guaranteed somebody knew what you were talking about. >> you were interacting via chat. it was a precursor to what you were doing now. we were collaborating. you were outside philadelphia and sean is in boston. >> i was in boston at the time, and he -- we did not know this, but he was literally a few blocks away. we went three months without ever actually knowing that. >> did you at the time have any idea what -- how big that was going to get? we talked to developers or people who have done viral videos who said, you know, i went to bed on tuesday, and i got up on wednesday, and i was the most famous person on the planet. did you have any idea how fast it was going to grow and how big it was going to grow? >> absolutely not. i would loveç to sit here and y we were gods and we planned every part of that, but we were just a bunch of really talented software engineers and hackers who wanted -- who were frustrated by what we found on the internet and we wanted something better, and it was a great product. that's a lot of us were actually

involved in the very beginning in helping debug it, make it better, make suggestions, add little pieces to it before it actually became a company. >> besides what you're working on now, the science cloud crowd, what do you see that's going on the internet space that's interesting to you as napster was back then? >> of course, what i'm doing now. >> that assumed, yes. >> gosh, hmm. that's a tough question. >> i know it's kind of out there. >> location-based services things like cloud computing. >> i'm probably going to take a different opinion that i think most people would offer. of course, there's lots of great things. we see computer services, computer technology, software scaling, cloud architecture, all these great keywords and things that sound awesome, but i'm cynical. i see a lot of things that already twisted that have just been rebranded. there are absolutely

improvements, and i think your next guest will talk about that a little bit, but by and large, i just see, you know, as new communities rise, as new cultures rise, the old ones fall, and they're literally just handing off these concepts and being reinvented. like cloud computing has been around for a while. zoog sure. >> it's only been called cloud computing for the last, i don't know, couple of years. >> let's talk about cloud crowd. it is the -- the concept is you are able to farm out jobs to all kinds of different people. one of the ways to get in contact those people are through facebook. you figured out that go to the place where the people are, and we have some pictures of the facebook website and some of the jobs available. is that the best place to find the biggest crowds? >> well, there's 400 million people now on facebook. 2 billion minutes a day being wasted. >> when they could be earning money with you. >> a little extra dollars on the

side doesn't hurt. i think by that estimation, it's a great place. is it the best? we'll see. >> is that primarily where your business is coming? when you are farming out jobs. >> that's the first place we've gone to because it was an obvious one. there's a massive underutilized labor pool of people interested in making a few dollars on the side, and with networking capabilities and sort of the built-in viral aspects, it's a great first platform. >> i think what's interesting about cloud crowd is it's so similar to what's happening with cloud computing, the idea of taking a big job and breaking it up into smaller, manageable tasks and distributing them to different processors. i mean, do you see parallels there, and -- >> of course. total parallel. as i mentioned earlier, that sort of abstract architectural idea has been around for a while. it's just either the taming has been bad or the execution has been bad or technology has not caught up to the point to make

it separate at the scale that we see it today. it's -- i guess i should differentiate against sort of first generation crowd sourcing solutions and crowd computingç solutions. i think there's been this rush that's driven this technology to make -- to bring -- to use technology to replace cost-ineffective or cost-inefficient slow human intelligence operations. i start seeing that as plateaued. google is probably the gold standard of how search results. i think there's a new plateau there. now we're to a point where even though that's plateaued, it's come up to a point where we can bring it back into the equation at scale, and that is what, you know, cloud crowd and second generation crowd sourcing is all about. sfoo which tend to be the projects that are most common? i have heard of people actually writing books based on this

concept. there's a company called guru forms, which is through facebook, where they basically put a company in touch with experts and pay those experts based on what they bring for projects. what type of things are we talking about? software development? creative works? >> we're focussing on the mass market, so the mass market is about under utilized labor. specialized labor tends to be far more employed. they have jobs. they have careers. they have degrees. although we do make use of those people as well, and effectively so. we're focused more on the let's break these tasks down into their most simplistic forms and -- >> that really lends itself to the crowd -- to the crowd, yes. the -- here are 18,000 pictures. some have dogs in them, some don't. separate them out into -- that's a horrible job for somebody to do, but if 18,000 people are doing it all at once, it's done in under a minute.

we'll talk more about the jobs on cloud crowd with jordan ritter coming up in just a moment.

sxwrirchlgs if are you joust joining us, we're talking with jordan ritter about cloud sourcing. send it out over the internet and having people work on that. we were talking about, say, identifying picture or sorting things. they can be simple jobs, but they're so overwhelming to one person that they have many people do it. they can do it very quickly. if i contract out to your company and say i need this done, are there any legal ramifications as to who is doing it? i mean, what if a kid is on facebook and he is you should 15 or whatever and he helps sort out a pile of pictures or if he is overseas, what does that mean

to me as a company? >> well, all of our workers are contracted to us, so you as a client -- you're not -- it doesn't impact your business. that's not your concern. we manage our own -- we manage them as contractors, and when they sign up, they basically sign forms that, you know, that confirm how old they are and whether they have ability to work and all those other details. that's not something the clients have to worry about. >> bottom line, it's your problem. >> it's our problem. >> amazon has tried this with something called mechanical turk, which is a historical name, but it's a terrible name for a service because no one knows what it is. >> true. >> amazon did it, and now you're doing it. does that make you late to the game. >> we're doing something quite different. i characterize them as a version 1.0 approach. it's -- it's a single task, sing the hit things that's all self-serve. that's the first point of distinction.

we're not self-serve. we work directly with our clients and take our factory floor, which is the second distinction. we treat our platform as a virtual factory, and shane launches a complex set of tasks that through distinction has quality control points. one of the core vaulsz of what we have is a reputation service or credibility system that every time a piece of work is done, it's peer-reviewed. in other words, things can't get through from beginning to the end until everyone agrees that it's actual -- >> it's some sort of filter or management. people are increasing the possibility of inaccuracy or conflict. you basically have a control set. >> that's something that first generation systems like$c@r(t&hc mechanical turk don't have. they weren't designed for that, and, frankly, it's hard to add that to platform that's been established and has a very, you know, simplistic way --

interface and simple listic interaction model. >> you can't get nine women together and have a baby in one month. referring so the software program, more people doesn't necessarily make it better. there are some tasks that services like this will be good for and some taskses that it won't at all be good for? do you have a clear distinction of what cloud crowd is good for and what it's not good for? >> well, so i guess academically there might be some things, but by and large with the companies that we've been working with and sort of the problems that they come to us with, we have been able to abstract things down to atomic tasks that given the right structure and given the proper factory configuration on the floor, produces high quality results. >> like designing a logo would not be a job for cloud crowd? >> probably not. it's not different to leverage what we have as a platform to do that, but our focus is on the mass market. our focus is on -- >> how about something creative

like writing a book? i know a ceo who wants to ghost write a book with a reporter, and they've been asked to outsource certain chapters to different reporters, and you almost would have multiple voices. you wouldn't have one unifying ois voice. it would be a difficult task. do you think that would be more difficult than, say -- >> sure. i mean, crowd sourcing is about leveraging the mass crowd. when you are reading a book, you are looking to -- >> you're looking for a distinctive voice. >> a couple of voices. >> kind of the neat thing with crowd sourcing, too, is that the more ridiculous the job, the more mind-numbing the job sshgs and i was using example, i have 18,000 photographs, five of which have a dog in it, and i need only those dog pictures, that's a horrible job for somebody, but it's very digital, and the worse the job is, the more it lends itself to crowd sourcing in the sense thatç somebody will say i won't look through 18,000, but he'll look

through 18, and if you can get your 1,000 to do that, you have it. >> it highlights the point of what's wrong with the first generation slewings. there's this philosophical limit to what you'll put through those systems. because you have to review everything that comes out of it, at some point it's basically the same time cost as it would have been to do it yourself. >> jordan ritter is one of the founders of cloud crowd. the more information at cloud crowd.com. thanks for coming in. >> my pleasure. >> up next, why computers are getting dumber and why one silicone valley company wants to joop that tr hn john tish when "press here" continues.

welcome back. just a few years ago if i asked you to predict the future of computers, would you probably predict the average computer would continue to get more and more powerful, but you would be

wrong. in fact, the future of computers is in dumb. computers like this, made through a partnership between fujit and pannologic. it's just a screen hooked up to an internet connection. by itself it's useless. there's no hard drive. hooked up to a distant server, the zero client as it's called is infinitely expandible and upgradable because all the heavy lift issing done in the cloud. >> production is automatic. >> ironically, if are you old enough, you know the dumb terminal model is how computers got their start. >> which, in turn, feeds the electronic computer. >> in the late 1950s and early 1960s the terminal was a technoligical necessity. now it's a cheaper and more flexible alternative for business.

>> and john, kisch, ceo of panologic joined by john ford of four tune and john schwartz of usa today. i always call them thinç clien. the idea that there's a computer somewhere that didn't really do anything on its own. it depended on a bigger computer judge else. you're calling them zero clients. is it a marketing tool, or is that -- explain zero -- defend zero client. defend it. >> sure. zero client is actually quite a bit different than a thin client in that there's actually no processor in it, no memory, no software, no firmware. >> i call it a screen. >> well, pretty much. i mean, the magic here really is that the device that we have and that we're marketing is really nothing more than an appliance. it's not a computer actually. the computing is being done back in the network, whether that's in a data center or in the cloud someplace. what this does is really allow you to think about the end point of a more diz posable form of

consumption point. end users are simply sitting in front of their computer but their computer is actually running someplace very different. >> ironically, it's back to the future, much that we saw -- >> that's how computers started. you couldn't afford lots of them. you put one in the basement and you put terminals or teletypes or something else throughout the office. >> sure. in point of fact, this is even dumber than an old dumb terminal in that dumb terminals had processors running in them, and, you know, away we're able to do is to contain the desk top in its own bubble that runs in the network, and we're able to use this device to access that bubble and make it easy for people. >> pardon the dumb question, but what is the end user doing in front of the zero client? what tasks are they performing? what are they doing? >> sure. they're doing anything they would do in front of a pc. in point of fact, ub like a lot of thin clients, can you treat this device exactly the same way. let's say you have a camera and you want to download pictures from that camera, the zero

client, you just plug it in. actually, behind the scenes back in the cloud someplace windows is loading the drivers for your camera and enabling you to go ahead and take care of the work you're trying to do. >> the point is -- it'sç kind what it comes down to. to anybody sitting in front of it and it just behaves. i log on. if you are running windows, it looks exactly like it does on any other computer, right? >> that's exactly the point. thin clients could get you about 60% of the way there. i know that because i used to run the largest thin client company in the world. why? and now what we've been able to do with desk top virtualization is actually build a device which is designed from the ground up for virtualization, for this idea that you run the computer in the bubble and the network someplace. >> and this is virtualization. this is you have somewhere out there in the cloud software that's pretending to be this computer and the broadband

connection is what enables this to happen. there are still the people who are going to want to have a full-fledged laptop setting and be able to access that virtual desk top through the web, so when i unplug this and i'm on my way home, it can actually do stuff in between broadband. stoo there's absolutely no doubt this technology today isn't applicable to every possible use case, but it is applicable for most of the uses that people have when they're sitting in front of a computer. that's currently connected to a network. this is being used in hospitals, in universities, in government offices, in manufacturing floors, in financial services companies, et cetera, and what's really driving this is actually the fact that if you do computing this way, you are able to lower the overall cost of providing computers -- of computing to the end user. why is that the case? well, not only does the device cost an awful lot less than a pc

or thin client, in point of fact you don't have to administer it because a lot of the computing is going on somewhere else, so when it breaks or if there's a problem, you can handle that centrally rather than having to dispatch someone to the end point to change a hard drive or reboot the computer or do something like that. >> the whole thick depends on the connection to that server out there somewhere. if that goes down, my whole office goes down. i suppose answering my own question, my whole office would go down anyway. there's very little we do in a news room that isn't connected to some distant server anyway. internet, e-mail. even the news program we use -- >> usa today west coast bureau, and i'm beholden to whatever happens on the east coast. >> you can fool around with paint shop for a while, and that's about it. >> that's exactly the point. five years ago we would hear this as an objection when people were still experimenting with this idea of -- >> that's -- there's a single point of failure there. i can understand the objection.

well, if somebody cuts this cord, we're gone. >> true. any time you have an end point that's relying upon a network to either get data or applications or in our case the entire computer, you, in fact, have a point of failure. the truth is over the last three or four years we've had tremendous advances in reliability in the network, and so consequently it's not really as much of a problem as it was, say, five years ago, but it's always going to be an issue as long as we're dependent upon a public network.ç >> do you have any sites that several thousand -- what's your largest installation, for example? >> largest installation is about 1,500, and that's running in a large hospital in new york. it's "acampus, it's being used from everything from surgery to pediatrics, to whatever. it's functioning exactly the way you would expect. quite frankly, the professionals love it because one of the

advantages of having your computer running in the network is you can get to it from different places very, very easily. if you are a nurse and you move from the fourth floor to the fifth floor all the time, you don't have to go in and relog -- sort of reboot your computer in order to get access to your information. >> what's your business model? it can't be selling hardware, right, because the hardware itself is so cheap, you can't be getting much margin on it? you machine selling the software and service that is make this work? >> well, today we actually sell both. we do sell hardware -- a piece of hardware, that i have in front of me here, but really, as i said before, the magic is in kind of how we enable this to interact with the network, and that's what we license out, and that's specifically what -- >> historically that's where you want to be in the long run is in the service of the software or the development, and that kind of thing. the hardware at some point becomes a race to the bottom. >> that's exactly right. every hardware company goes through a pretty standard

commoditization curve. >> pick up one of those things. what is this? that's a computer? >> well, no. >> sort of. i'm sorry. you're right. it would act like a computer. i would not know the difference. >> it's actually just an appliance. what you'll see here if you can look at this is you'll see three usb ports. you'll see a video connection. you'll see some audio on it, and you'll see an ethernet tap, and that's it. inside is just an fpga. we use it az abuffer to take the video suggest signals and audio signals, and the truth is that's our biggest advantage here because there isn't anything in here. i'm not paying a lot of license fees or a lot of taxes to other companies, and in point of fact, this will end up, i çthink, beg able to be be in a much more interesting way than anything else will. >> john kisch, ceo of panologic.

back in a moment.

my thanks to all the folks we had on this week. fp you would like to know more ways of thinking about the computer check out episode number 6 called cheap computer. thanks for making us a part of your sunday morning.