The game is afoot! Or a-tread!
Very. Very. Interesting.
Any thoughts on this @Vlad? What is cool, exciting unique, intimidating about this?
Honestly it’s hard to tell. The current info is all marketing-speak filled with nonsensical buzzword bullshit so it’s hard to actually decipher what will be there and what won’t. Also, it’s Google, so it may very well be that 6 months from now they drop this all together because they have the attention span of a dog chasing squirrels.
Now I can hypothesize what could be really cool if they do it right: fully integrated local and cloud computing utilizing their new Edge TPU modules as well as cloud resources to provide a seamless array of robotic capabilities, such as mapping, object/people recognition, NLP, etc. Sounds like they are placing lots of emphasis on sharing of data between robots such that what one robot learns others will benefit from (ie different robots mapping different parts of a building and all getting combined together). If they make that process easily expandable and available to stuff you develop too not just Google specific features it could enable some really cool collaborative robotics applications.
A lot of it will depend on what platforms they support etc. It also depends how much they depend on the “cloud”. If they make it that everything is all cloud based and try to force you to be all in the cloud it’s not going to be good. If they do a smart scaling of local and cloud stuff then it could be awesome.
We shall see.
Don’t hold back
Also, note that there have already been efforts (commercial and academic) to achieve these goals.
E.g., there is C2RO (http://c2ro.com/). They actually pivoted. The word “robot” is not even on their homepage now, but compare with their website 19 months ago: https://web.archive.org/web/20170316001725/http://c2ro.com/
E.g., in academia, RoboEarth | What is Cloud Robotics?