Page 32 of 43« First...1020303132333440...Last »

Two’s company, three’s a crowd conference: CrowdConf 2010

September 1st, 2010 by

It has never been easier to keep in touch and informed with the world around us. With email, newsfeeds and Twitter, the problem today is too much information, rather than too little.

Even so, I’m a firm believer that there is no substitute for good, old-fashioned face time. Actually meeting people in your field, having a chat and swapping contact details is still the best way to develop your ideas and grow your business. The problem is, of course, no one has the time to constantly travel around the place meeting people. Everyone is far too busy reading their news feeds.

Mingling with the crowd
This is why all of us at Microtask are so excited about the first annual crowdsourcing conference, CrowdConf, which is being held in San Francisco on 4 October.

In attendance will be researchers, technologists, outsourcing experts, legal scholars, and artists from all over the crowdsourcing world. Along with the all-important mingling, attendees will hear from a list of guest speakers that reads like a who’s who of the crowdsourcing and human computation industry. This list includes Jeff Howe (who first coined crowdsourcing in 2006), the bestselling authors Tim Ferriss, Jonathan Zittrain and David Alan Grier, and the CEOs of some of the most exciting new crowdsourcing businesses such as Crowdflower, Samasource, Keniks, uTest, Odesk, LiveOps and Victors and Spoils.

In addition to the guest speakers there will be a series of peer-reviewed presentation tracks and technology demonstrations. Topics for discussion include the past, present and future of crowdsourcing, quality assurance and metrics, social and economic implications of crowdsourcing, task design and worker incentives, as well as innovative projects, experiments, and applications. Basically, everything we talk about on this blog.

If the future of the working world sounds like something you might be interested in, don’t miss out (tickets are still available). Exactly what we end up talking about is of course anyone’s guess, but what is clear is that the ideas and relationships that this inaugural conference fosters will play an important role in shaping the world in years to come.

Microtask is a proud sponsor of the event, and both our CTO Otto Chrons as well as yours truly will be attending in person. If you’re visiting CrowdConf, please come and swing by our booth. And if you really don’t have time to make it, don’t worry too much – at least there’ll be plenty of blogging, vlogging, texting and tweeting to fill you in on what you missed.


Human Flesh Search Engines: The Most Dangerous Game

August 30th, 2010 by

Like a wolf in sheep’s clothing, evil is most frightening when in disguise. The scariest movies – the ones that send shivers up your spine and keep you awake at night – are those where evil manifests in an innocent form. As anyone who has seen The Exorcist, The Omen or The Ring (or looked after a two year old) knows, nothing is scarier than cute kids.

The Human Flesh Search Engine (HFS) sounds awfully sinister. But far from being a deadly disease or a machine from the Saw series, it is simply a term used to describe searches that are conducted with the help of humans (as opposed to the ones carried out through a search engine like Google).

A tiger in the grass?
A good demonstration of the power of HFS is now known as the South China Tiger Event. In 2007 a tiger species considered long-extinct in the wild was spotted and photographed by a hunter in the Shaanxi province of China. After the pictures were published in Science magazine, skeptical Chinese internet users started a challenge to establish whether they were authentic.

The human searchers who took up the challenge included professionals with backgrounds in zoology and botany, photography and geometry. In the end, however, the key was much simpler. One of the volunteers recognized an all too striking similarity between the pictures and a painting in a calendar, exposing them as a scam.

Just as in this example, HFS events generally involve both an offline element (such as checking photos from old calendars) and voluntary crowdsourcing.

Typically an event will start with the formation of a small seed community, which issues a task with a defined goal. Most appear first on an online forum, enabling searching by broadcasting requests and action plans, sharing online and offline search results, or offering specific rewards. Some can also be triggered – or widely publicized – by the mass media.

The taste for human flesh
Unfortunately, as anyone who has heard of the phenomena before will know (don’t feel bad if you haven’t, as HFS is mainly a Chinese thing), the applications of HFS are sometimes just as sinister as the name suggests. In 2008, for example, a young man started an HFS claiming he was sick and looking for help to fulfill his last wish: to see the woman he loved one more time (cue romantic sighs from the crowd).

Aided by the community, the man managed to trace his lost love. The problem was, his sickness was actually more mental than physical, a fact well known by his ex-girlfriend who had been hiding from him (cue violins from the shower scene in Psycho). After trying without success to win her back, he killed her with a knife.

As a result of this accident and other minor abuses, the HFS community started to self-regulate, carrying out background research checks on those who initiate searches. But while the case was unusual in some respects, much of what is wrong with this scenario is typical to HFS, and not addressed by the regulations. The reason why the name Human Flesh Search Engine is so apt is not just because people are involved in the search, but because most events involve hunting down people that the online community want to punish for some reason. This led The New York Times, to call it online vigilante justice.

But HFS need not be synonymous with petty vigilante justice and tragedy. By combining the wisdom of crowdsourced human intelligence and the speed of algorithms, these hybrid searches have enormous potential.

What is needed is a system which inspires people to work together for the greater good, instead of the current Hollywood style do-it-yourselves justice league. The question is, are we up to the task? Or will HFS remain the most dangerous game?


Time for our next step

August 24th, 2010 by

Cobija: Corporativa al atardecer - Flickr Meeting at Tusk by Milivoj Sherrington @ Flickr

It is often said that timing is everything. While this is probably an exaggeration, history is full of great ideas that didn’t work out because the time wasn’t right (just ask Galileo, Aristotle or Van Gogh).

Recently we told you how Microtask’s journey was beginning to gather momentum, with success at the Nordic Tech Tour.

Although only a short while ago, things can change quickly – especially in the world of tech start-ups. Back then double-dipping was still associated with chocolate and ice-cream rather than recession. Now instead of talking up success, when the media mentions new companies, it is more likely to discuss how little funding is available. On both sides of the Atlantic, the economic outlook seems gloomy.

The sun also rises
It is against this backdrop that we take extra pleasure in announcing our agreement with the highly regarded venture capital firm Sunstone Capital. Along with seed funding, this agreement means we will now have the considerable expertise of serial entrepreneur and all round great guy Nikolaj Nyholm on our board. (For extra details, check out this article from the ever-vigilant ArcticStartup).

In our opinion, timing could not be better. Despite gloom in much of the economy, momentum in the digital work market is rapidly building. Several prominent start-ups have recently raised their first venture capital financing rounds, including CloudCrowd and Crowdflower.

Rather than view them as a threat, we believe these US based start-ups will help us to develop the global market (both in terms of supply of digital labor and demand for it). With the current momentum, we think 2011 will be the year crowdsourcing goes mainstream, providing new opportunities for millions of people around the world.

Leaving aside the crystal ball, we are now working flat out with our first customers. Later this fall we are planning to announce more exciting news. Until then, if you are an engineer and want to help shape the future, we are now recruiting. There has never been a better time to get on board!


Winner takes it all?

August 17th, 2010 by

On this forum we spend a lot of time discussing the enormous potential of crowdsourcing, and how it is going to change all our lives for the better. At Microtask, we just love crowdsourcing. With that in mind, what I am about to say may shock you. I urge you to sit down.

Recently I read a rather disturbing article on crowdsourcing. The article focused on 99designs, a community for crowdsourced web pages and logo designs.

In theory it’s great. Say you want a new identity for your company. All you do is post a design brief on the website, and then sit back and watch the submissions roll in. At the end of the process you only pay for your favorite submission.

99designs has all the mechanisms you would expect to ensure chosen work is paid for, avoiding both copyright theft and exploitation of the community purely as a source of inspiration. Even so, there is much criticism of how it and similar sites (such as reDesignMe, MycroBurst, CrowdSpring) operate.

Minority report
Some claim that this kind of crowdsourcing incentivizes quantity over quality (note 99designs’ mantra “a new design uploaded on the site every 7 seconds”). There may be some truth to such claims, although presumably the fact that people are actively using the site suggests that they believe the results are worth the cost and effort.

People also argue that this “competitive crowdsourcing” exploits the contributors. As a user who commented on the article I mentioned earlier asked: “I want to know how many of you would show up at the office Monday morning if you had to compete with 92 co-workers for a single pay check at the end of the day?”. Note that the two criticisms are interrelated, in a way, given that the ability to only pay for one solution drives the cost down. (Of course an economist might argue that rational contributors factor in their chances of success as well as the possible reward when they make their decision to participate).

The idea that such communities lead to exploitation gathers weight when one considers that minors are prevalent amongst the contributors. Although 99designs is actively trying to stop people under 18 from accessing the community, their efforts are easily circumvented over the internet.

To this criticism one user replied “No one is forcing people to participate by sending in designs. This is all voluntary work. (…) For some of those 11-year-old designers, perhaps they really are showing an interest in art or design, and this gives them an avenue to exercise their nascent talents, or learn some designing skills, or software tools. Heck, I think it would be a great exercise for a school project in an Art/Design class to have every student work up a submittal as part of a class assignment.” (Without getting into this particular argument, personally I think that there is no doubt that a system like this would be a great teaching aid in certain circumstances).

Creative microtasks?
Leaving these debates aside, for me the underlying question remains: when a high quality, creative output is required, is competitive crowdsourcing the best solution? To my mind, the key problem with these crowdsourcing communities is that submissions are mutually exclusive. If one is selected all the others are automatically discarded or must be reworked and adapted for different projects. Many of the issues surrounding exploitation of contributors flows on from this fundamental problem.

Happily, there is another way. Crowdsourcing concepts like Microtask’s, which rely on a very low level of worker input, have an entirely different approach. Every submission (or “microtask”), when added to all the other small fragments, cumulatively contributes to the successful completion of the whole task. Everyone who completes a task to the standard required will get the reward they expect.

Such a crowdsourcing solution clearly has enormous potential for tasks involving recurring steps or well defined mechanisms, but its ability to deliver creative solutions has yet to be proven.

This is an area we are currently researching. We would love to hear your opinion on how creativity and crowdsourcing could coexist in the same sentence without sacrificing output (and life) quality. Don’t worry; all interesting opinions will be rewarded with equal amounts of our gratitude and respect.


Watching Big Brother

August 10th, 2010 by

The summer of 2010 will be remembered for searing temperatures and soaring government debt. With more whistle-blowing than the World Cup final, it was also the summer of Wikileaks, the whistle-blowing internet site, and its weird, white-haired front man Julian Assange.

Thanks to Wikileaks, I now know that even if I am dying of thirst in a heat wave, I am better off avoiding Helsinki’s ATMs to draw cash for a drink. When I am kept awake at night by the heat and ATM fraud, I can put myself to sleep reading 92,000 secret documents it has published relating to the war in Afghanistan.

Opening the flood gates
But even all the attention Wikileaks is receiving, historians looking back at this time may see it as just the exciting edge of a much wider trend. For when I finish reading the files about Afghanistan sometime in 2020, there are now vast databases of government information to get stuck into. Unlike Wikileaks, these online databases are not leaked documents, but rather information voluntarily released by governments around the world. Although less sensational than Julian’s masterpiece, this information may prove to be even more useful.

Pressure on governments to publish public information has been mounting for some time. Reflecting this mood, in 2005 a non-binding European Commission directive encouraged increasing openness by member states. But it seems like it is this year that things have started to get moving. On his first day in office, President Obama instructed the heads of federal agencies to release as much information as possible. This has led to a surge in information available for organizations like Public.Resource.Org. Similarly, in the UK, the Government has recently launched the massive COINS database, giving anyone who cares to look an incredibly detailed look at UK Government expenditure.

A wolf in Google’s clothing
All this is great, but just releasing this information does not necessarily change much. The problem is that the information available on any one of these sites is so vast, that even if I never slept another wink, I would never get through it. It means in practice that the information is almost as impenetrable as it always was.

Happily, to coincide with the emergence of these databases are systems which allow the public to collectively and individually scrutinize this data in a meaningful way.

One method, relying on the collective efforts of many concerned citizens, is to crowdsource the analysis. So, for example, when the MP expense scandal in the UK broke, rather than have teams of reporters sifting through almost half a million expense claims, the Guardian newspaper crowdsourced the scrutiny of them, with some interesting results. Broken down into bite sized fragments and shared amongst many thousands of people, analyzing government is more a microtask than a massive undertaking.

But a crowd is not always necessary. New systems, such as Wolfram Alpha, allow even individual citizens to scrutinize large amounts of data in a meaningful way. Rather than searching data like Google, Wolfram Alpha computes it, applying code from Mathematica to its own vast library of curated data (which makes it more like Wikipedia than Google, in a way). In theory, at least, this will make it much easier to analyze huge data sets, such as COINS.

Government by the people, for the people
All this access to information obviously makes it much easier for the public to keep an eye on what their elected (or unelected) officials are up to. It also has profound implications for how government works.

Rather than fear the public scrutiny, many in government are using it to come up with better ways to manage services. One of the first things the US federal government CIO did after taking office was to create an online dashboard, detailing the government’s $70b technology spending. Along with allowing extra scrutiny of what they were doing, the aim is to get the public to offer suggestions on how to do things better and more efficiently. It is not surprising then that cash-strapped governments like California’s are implementing similar programs.

Just how much Wikileaks and COINS type databases, Wolfram Alpha and crowdsourcing will change the way we are governed is hard to say. Transparency and scrutiny of government will always be a good thing, but there are perhaps practical limits to how much power can be devolved to the people. If California’s experiment with direct democracy is anything to go by, sometimes it is better to let elected politicians get on and do their jobs. As with much of what we discuss in this blog, only time will tell just what the future holds.


Page 32 of 43« First...1020303132333440...Last »