When a virtual entity, programmed with artificial intelligence - a ‘machine’ - receives unstructured input, like text, pictures, music, or voice, it will split it up and convolutionally convert it into small numeric representations, called vector embeddings. When this input is systematically offered in different ways (like people would experience this in training), layers of interconnected vectors - a ‘machine mind’ - will start to emerge. Eventually this ‘mind’ will be able to:

Detect probable relationships between items presented to it;
Determine the context (or lack of it) these items exist in;
Distinguish the levels of significance of such items in time and space;
Derive possible intentions and scenarios regarding these items.

This is why intelligent machines can ultimately represent us, people, in a territory populated by machines only.

However, if machines need to interact with each other, like in jobs, transactions, or projects, they need a protocol how to make this work. Transreality.com coined the word ‘vector merging’ for this. It reflects a technical protocol that permits different machines (each one deploying its own vector database), to temporarily connect by merging their vector databases into a solution-response mechanism for a short period of time.

If you want to skip this tech stuff, please check our fields of application page.