Molecules of Knowledge (MoK) is a coordination model supporting self-organisation of knowledge in Knowledge Intensive Environments (KIE). Usual approaches to knowledge management in KIE consider data as a passive, "dead" entity and rely on "brute force" approaches assuming an ever-increasing computational power and storage capacity (e.g. big data). This won't scale forever, thus alternative approaches should be explored. MoK promotes the vision of data as a "live" thing, continuously and spontaneously interacting and evolving-self-organising. Accordingly, MoK relies on features such as locality, probability and situatedness to tackle KIE challenges such as scale, openness and unpredictability. In this seminar, the MoK model is motivated and introduced, then some early "evaluation" described.