The press typically describes DeepMind as "highly secretive," but actually they publish a ton of their research — including this paper — in all the usual venues: NIPS, arxiv, etc.
Highly secretive sounds cooler (to most of their audience). Personally, I find open science exciting.
The paper.
Discusses the technical aspects of one of Googles AI projects. According to a pcworld the system "apes human memory and programming skills" (this article seems pretty solid, also contains link to the paper).
The abstract:
(First post here, feedback on the appropriateness of the post appreciated)