Some clarifications (please let me know if you see it differently!)…

In the chat as UI examples (Magic, Operator, M), the no UI part is the functionality that you don’t deal with because the service is automating it for you. You don’t have to interface at all with the flower shop to buy flowers when M does it for you (AI + humans you don’t see) after you request that through their UI (messaging). The chat UI itself is not no UI.

In both the Soli and Emotiv demos the devices are replacing existing input devices, not the UI. You can plainly see that the demo users are controlling a UI on a monitor to demonstrate that the input works and get feedback from their input (needed to complete the actions shown in the demos). When a person thinks a thought or rubs their fingers together and a garage door opens, that will be no UI. You could demonstrate that with any existing device though, so you see it’s the lack of UI that makes it no UI. :)

This doesn’t minimize the awesomeness and innovation of the technology you mentioned, it just isn’t accurate in terms of examples for no UI.