- The connected home is not smart
I only noticed this now; but, it seems that, when Google merged the Home and Nest divisions, it rebranded the connected home from ‘smart’ to ‘helpful’. The announcement said:
Today, we’re committing to that goal by bringing the Home products under the Nest brand. Our first step as Google Nest is to go beyond the idea of a “smart home,” and to focus instead on creating a “helpful home.”
The VP for Google Home and Nest, describes the helpful home, as:
It’s a home where products work together to help you feel comfortable and safe, keep an eye on things when you’re away, and connect you to friends and family.
With this move, Google changes the emphasis from what the product is to what the product does for the customer. So, Google is not tied to a specific product, and has more scope to find ways in which to compete and add value.
In addition, with this move, Google downgrades customers’ expectations. ‘Smart’ means being intelligent, while ‘helpful’ means being ready to give help. If the connected device doesn’t understand your accent, can’t answer your query, or gets the temperature wrong. it’s not very smart. But it can still be helpful.
2. Data is not the new oil
Clive Humby, the mathematician behind the first mass customisation loyalty programme (Tesco’s Clubcard), once described data as the new oil. He said that data could be a huge source of value for those that mined, analysed and found new insights from data; much like oil could bring wealth to those that extracted it, processed it, and produced goods (e.g., fuel, plastics…).
This description has been repeated many times since then, and has become a generally accepted way of looking at data and their use.
However, recently, Tim Berners-Lee challenged this comparison. Talking at the annual summit of the Open Data Institute, Berners-Lee said that:
Describing data as the new oil is really, really wrong. With oil, there is a certain price per barrel and if I give you my oil, then I do not have any. (…) If I give you my data, I still own it, and I can give it to other people, unless I come to an exclusive arrangement where I only give it to one company, and that’s really bad – because the value of data is when you share it.
If my data is something that is intrinsic mine; something that I can grant access to but over which I continue to have control, then:
Instead of talking about the value of data as though it were oil, we talk about the rights to it. I should have the right to have my data because it is a human right. I should have the ability to be able to do things with it.
3. Artificial intelligence is not machine learning
When people talk about Artificial Intelligence (AI), what they are often referring to is “Machine Learning” (ML). And many descriptions of ML are actually talking about supervised ML algorithms. A crude description of supervised ML is that it is a computer programme that uses data to learn by itself. For instance, the computer would use a dataset about adverts shown and a dataset about purchases made, and find the rules connecting the two.
Supervised ML is powerful, but is only one type of ML; and ML, in turn, is just one tiny part of AI.
So, when someone is talking about close we are to AI, how good / dangerous it is, etc… it is important to clarify what exactly people are talking about. Are they talking about the narrow case of supervised machine learning? Then, yes, we are very much there. But if it is the latter, then we still have a long way to go.