According to the article, Google are in process of making mahcines to think more like humans. As technology develope, people expect more and more from the machine. People want machine to understand what they are trying to ask by providing minimum information such as just giving a picture, typing up some words, etc. When Google finishes develpoing the programm so that it can be used in all kind of product they are providing, they will start updating this program to ALL the product from their company. The example they made is for example. if someone look for the picture of thanksgiving, there will be people, turkey, tree, and ect. But what the program we are using right now will read out of it is turkey, tree, or people, not thanksgiving.
I personally don't think it is really smart to develope such program. The reason is most of people I know, such as my family, friends, teachers, etc, don't use the "simple" search service. It is really hard to use, because if it tells us to say something for them to write down, they write down some random words that is compeltely different from what I said. For example, when I say Apple, they will read axe or something.
I saw similar article from few years ago that windows are working on the same project. The result of the developemnt is Cortana that window recently introduced. It is also very not useful. I never use that component since I first tried it when I got it and it didn't work at all. When I also look at those advertizements on tv, they are using the voice search thing very well, which never worked for me.