My current research focus is in recurrent Neural Networks known as Cell Assemblies. This was implemented in the EPSRC funded CABot project to develop a Cell Assembly roBot that view the environment, takes natural language commands from the user, and maintains its own goal. Cell Assemblies are a computational model derived from mammalian neural, and psychological evidence. They are also a novel computational medium that generate active symbols. This mechanism gives a reasonable (though of course incomplete) explanation on how people think. Moreover, it is mechanism that could eventually be applied to a wide range of real world tasks. I have moved into this area as a way of implementing semantics for Natural Language Processing, but it could be used for a wide range of other problems. We are exploring the use of CAs for categorisation, speech recognition, cognitive models, and agent technology.