interface

I agree to some extent with Manovich's commparison of Worf-Sapir hypothesis to the idea that an interface affects how the viewer takes in and assesses the information, but it is a bit extreme. The interface, in my opinion, can only have a minimal affect on how the information is encoded by the viewer. By stating this, is the author suggesting that the creators of these interfaces have possibly an equal or greater influence on the information presented than the creator of the information presented? I don't agree.
Also, I may need to just think about it more, but how can something created with the AL approach possible? If you assume that, for example, if you create a program, such as a Tamagochi, you need to create the almost infinite outcomes for every possible action/reaction, how can that kind of program be considered to be impossible to predict beforehand?