What is Google tensor, and why you want to know


Here is a phone chip that may have a very interesting effect.

What is Google tensor, and why you want to know /img/google-tensor.jpg

The new chip in the Pixel 6 phones by Google, called Tensor, may be particularly interesting for one weird but powerful reason, if a journalist got it right: a successful and widely copied Tensor could be a setback for 5G mobile networks.

The thesis is that the real reason to make Tensor chip is not to sell huge number of Google phones. It is to make the whole Android ecosystem, that is all phones and tablets of every maker that run it, to “shift toward running more Artificial Intelligence (AI) and Machine Learning (ML)” software straight inside those devices.

This is good (for Google) because Tensor-enabled devices may do themselves, that is in the hands of their users, much of the data analysis that is needed to find which ads could be most profitable with each single user.

In other words, if enough Android devices get the same capabilities that Tensor provides, Google will be able to make as much money as today, without risking any privacy lawsuit, or regulation. Regardless of who makes those chips or devices!

The problem is that what is good for Google is not necessarily good for other interests. Phones so powerful to run AI applications themselves don’t need “high-bandwidth, low-latency” connections, that is the 5G networks on which so many governments and corporation desperately need to spend lots of money.

This could go very well

In my opinion, 5G already has enough fragilities and reasons to NOT exist that anything that can set it back is welcome, and may have to work less than many would imagine. And the idea of “in-phone” profiling may solve a lot of the surveillance problem caused by the social media platforms of today. If it is not controlled by Google, that is, and can be completely configured by each user.

If you get a Tensor phone, ask how you can “turn Tensor off”. Then let me know how it goes.

Stop at Zona-M   Never miss a story: follow me on Twitter (@mfioretti_en), or via RSS