A long-time staple of the pub philosopher's diet, the infinite monkey theorem states that a monkey randomly typing for an infinite amount of time would almost surely compose the complete works of Shakespeare. Uncylopedia, a vital resource for those not in need of the right answer, holds up The Da Vinci Code as evidence that this has already been put into practice.
While a mathematical proof exists for the theorem, it's not really applicable in the real world. A four-week experiment with six Sulawesi Crested Macaques at Paignton Zoo in 2003 resulted in just five pages of text, most of which was the letter ‘S'.
It's an illustration of a practical constraint. If we had enough resources - particularly intellectual ones - think how many of those knotty challenges in life and business we could address.
In game shows, asking the audience is almost always the best option; in the real world we lack access to a big enough pool of talent, and a way to communicate with and pay it.
However, the internet has changed this. Crowdsourcing, a term first coined by Jeff Howe, then editor of Wired, in 2006, describes ‘the act of taking a job traditionally performed by a designated agent, usually an employee, and outsourcing it to an undefined, generally large group of people in the form of an open call'.
A wide variety of organisations has put this into action, from the Texas Border Sheriff, who puts feeds from Mexican border video cameras online and invites citizens to report suspicious activity, to those wobbly passwords that restrict access to websites.
CAPTCHA images are often used by websites to ensure only humans can access information. ReCAPTCHA, a clever implementation of this, uses images generated by book-scanning projects. If the system does not recognise a word, the reCAPTCHA system, displays it as a password test to users
on Facebook and other sites. In their transcriptions, users are effectively digitising the book word by word - 10m words a day, to be precise.
Both these projects tap into low-skill activities, creating something valuable that would be costly by other means. Yet most tasks don't warrant setting up dedicated systems such as this.
Amazon's Mechanical Turk was developed to create a market for low-skill jobs. A company needed a list of sales contacts in a structured format, but had thousands of documents with contacts buried in them. It published the documents on the Mechanical Turk as a HIT, or ‘Human Intelligence Task', and offered12 cents for each reponse.
There are more than 40,000 HITs currently on the system, ranging from translation to categorising images. And these crowdsourcing marketplaces are effective for complicated tasks as well.
Innocentive was established in 2001 to bring research and development challenges to inventors and scientists across the world. Procter & Gamble, Eli Lilly and several not-for-profit organisations use the site's 175,000 registered ‘solvers' to find solutions to their development needs. The site has paid out more than $4m in rewards so far.
Crowdsourcing is a creative and
cost-effective way to access resources and innovation skills that organisations cannot carry in-house. It enables the completion of tasks that simply wouldn't have been economic before, and has even found its way into the outsourcing of marketing tasks.
But while costs might be slashed, as development risk is transferred out of the company, budding crowdsourcers should bear in mind Paignton Zoo's experience. If you pay peanuts, you will get Sulawesi Crested Macaques.
Andrew Walmsley is co-founder of i-level
30 seconds on CAPTCHA
- The concept of a test to prove that a web user is a person, rather than a programme, was devised in 1997. It was intended to prevent internet bots from indexing URLs.
- The acronym ‘CAPTCHA' was based on the word ‘capture'; it stands for ‘completely automated public Turing test to tell computers and humans apart'.
- In 2000, programmers at the School of Computer Science at Carnegie Mellon University developed the ‘wavy text' format that is now in use.
- Luis von Ahn, a professor who was part of the original team, estimated in 2008 that web users fill out 200m CAPTCHAS a day.
- Programmers predict that it will be two to four years before internet bots can accurately read CAPTCHAS.
- reCAPTCHA is a free public service, the use of which helps to digitise old books.
- It is also working on making back issues of newspapers such as The New York Times available online.
- reCAPTCHA accurately digitises 99.5% of text.