It's all machine-written.

Associated Press (AP), an American nonprofit news agency, announced in Friday that it had published 3,000 articles in the last three months of the past 2014 years, and that in the past they had been able to publish only 300 articles. The 10 times-fold increase in the number of posts has no growth in human costs, all of which come from an algorithm.

Last summer, Associated Press, in collaboration with a company automated Insights, which provided real-time content automation, transformed the data accumulated in the past into articles, visual images, and so on. In the past 3,000 articles, the article generated by the algorithm is almost no mistake, you can not even put these articles with the human writing of the article separate. All of the articles are based on the AP style, and the content is rich and varied far beyond the ability of a normal author.

Not only the high yield and high quality, but more importantly their timeliness in the news advantage. Last year, just 3 minutes after the Los Angeles earthquake, AP reported the news of the quake to the public, and in recent days, after Apple released a few minutes of last year's earnings, AP wrote about Apple's 2015 Q1 quarterly earnings forecast. This means that machines have begun to learn, and even provide analysis, insights into areas that add more brain thinking.

Narrative science is another company offering similar services, and their CTO, Kristian Hammond, says that by the year 2030, 90% of the news will be written by machine. Currently, Forbes also has some news media, has used narrative science company's house arrest writing e-mail and so on documents. It can be imagined that future machine writing will replace more and more scenes, and their writing news will become more and more intelligent.

PS: This article by the machine generation, and the author Lai min have no relationship.