Microsoft is collaborating with researchers at the University of Cambridge to develop an AI which can turn your descriptions into working code in seconds.
Called DeepCoder, the software can take requirements by the developer, search through a massive database of code snippets and deliver working code in seconds, a significant advance in the state of the art in program synthesis.
“The potential for automation that this kind of technology offers could really signify an enormous [reduction] in the amount of effort it takes to develop code,” says Armando Solar-Lezama of the Massachusetts Institute of Technology, who was not involved in the work. “All of a sudden people could be so much more productive. They could build systems that it [would be] impossible to build before.”
DeepCoder learns which combinations of source code work and which ones don’t as it goes along, and improves every time it tries a new problem. It created working programs in fractions of a second, whereas older systems take minutes to trial many different combinations of lines of code before piecing together something that can do the job.
Developers however do not have to worry just yet. At the moment DeepCoder is only capable of solving programming challenges that involve around five lines of code, but given the pace of development in AI in recent years it is probably only a matter of time before being a good developer would be more about telling the PC clearly what you want rather than being good at functions and inheritance.
The full paper can be read here.