Technology marches on, but it sometimes just isn’t intuitive enough: there’s a surprise.
There’s an article on The Telegraph online headlined ‘Twitter’s machines call on low-paid humans in battle to keep up with hashtags’. It says that Twitter has said it’s developed a “computation engine” made up of low-paid human workers to categorise topics and hashtags enjoying sudden spikes in popularity in order to help it sell advertising against them.
It seems that computer algorithms are currently unable to understand the meaning of some topics, says Twitter. Well, I have to say, I don’t understand many of the topics on Twitter and I am happy for my own personal algorithm to stay that way. Twitter is good in certain respects, but I find much of it banal and a sap and waste of human energy.
Back to the Telegraph article: an example of a trending hashtag that was not understood by Twitter came during the recent US presidential debates, during which Mitt Romney said he had “binders full of women” at his disposal when he was choosing his senior staff as Governor of Massachusetts. The odd turn of phrase was immediately mocked online and led to a #bindersfullofwomen hashtag among Twitter users. This became so popular it became a ‘trending topic’.
If you are not familiar with Twitter, trending topics are the most popular hashtags, phrases or names being talked about on Twitter. They, therefore, account for a large proportion of traffic at any given time. When you are on Twitter you are able to see displaying trending topics, which can change from minute to minute.
Computer algorithms have never seen hashtags such as #bindersfullofwomen before so they are unable to put them into any context.
Of course, Twitter needs to understand these trending topics to optimise its growing advertising business.
So Twitter, it seems, has hired a new “human computation engine” through an Amazon service that links employers with home workers who are willing to do typically mundane jobs that computers are still incapable of – including accurately transcribing speech.
“We’ve built a real-time human computation engine to help us identify search queries as soon as they’re trending, send these queries to real humans to be judged, and then incorporate the human annotations into our back-end models,” it explains.
Twitter hasn’t revealed how many people are employed doing this, but says it relies on an elite “small custom pool”.
The Telegraph article estimates that the workers will be on a rate of $3.73 per hour.
While I wonder at many advances in technology I really do worry about #theworldtoday
Has #communication improved at all? It’s not just computer algorithms that are having problems keeping up.