Wow, that's cool, being able to manage my home without any remotes, which my kids tend to lose somewhere, anyways. Thinking of the kids: How about the ability to have an override on their TV usage ...
Home Automation is the Future
Home automation or smart homes are on the list of hot topics
for quite a while. The
intelligent fridge that automatically orders food items that are about to
run out is a decades old story. Its first mention was about 1998. Even my 12-year-old
house here in NZ is fully wired and has some (not so, admittedly – but pretty
expensive if you need to repair them) smart controllers which probably could
control more than they effectively do.
Technology like this is not only exciting but offers quite
some potential to make our lives far simpler and to improve the experience of using
the connected devices, well, with that of living in our homes.
Still, many attempts at smart homes have been done over the
years, most have failed or were not that successful.
Comes the cloud and massive improvements in computing power.
Comes the mobile phone and the tablet.
Comes working voice recognition.
Comes predictive analytics, artificial intelligence and machine learning. Well, AI is around since at least the mid 60s; Weizenbaum’s famous Eliza dates back to 1966. Alan Turing developed his famous test in 1950. AI was again a hip topic in the 90s (my thesis back in 92 was about knowledge based AI systems; but that is almost prehistoric …).
And now we have intelligent assistants like Siri, Google Assistant, Alexa, to name but a few – with Siri probably having some of its intelligence directly on the device. But notably, most of these systems have their intelligence in the cloud, or more precisely on servers owned by their makers.
The Way it Works
Let’s go back to home automation again. Having the
intelligence that controls my house on a remote server basically means that
whatever command is given goes there via my LAN, through my router, into the
cloud where it gets interpreted – and all the way back.
At the same time cybercrime is soaring (data from Symantec), reports on hacked cars are coming in at more than a drip rate. Try googling ‘hacked cars’! This query retrieved 780,000 hits.
Home networks are regularly the weakest protected networks. How many people do not even bother to change the default password of their Internet router? With all the hype and also all the admiration for the great tools that have been developed by Apple, Google, Amazon and all the other companies this sounds like a pretty big risk to me. Still none of the vendors says much about security; if anything at all it is about the communication between my gadgets and their intelligence happening via SSL.
A Better Paradigm
With all the hype and admiration of what these companies
have achieved there must be another way that exposes the user of these exciting
technologies somewhat less and makes her sleep better. And that still works
with the Internet connection down …
Think Edge Computing.
Let’s have a look at Tesla, Apple Car, Uber, Google Car. All of these models of self-driving cars rely on a very powerful AI, and on machine learning, too.
But they employ local intelligence in combination with swarm learning. This means that the intelligence is on the car, and for several good reasons that I do not want to discuss here.
Apply this to the smart home!
There is a local intelligence that talks to the devices. This communication is fully encrypted. This communication does not leave the home as long as I am in my home. Local learnings may still get sent to the main brains, in an anonymized way. New learnings get pushed back to my local brains. Fully encrypted, too.
This is far more secure, even in the scenario of me using my smart phone from wherever I happen to be on this planet.
And as a side effect there is increased privacy, too. But this is probably more of an obstacle to Google and Amazon.
The technologies are all there. The experience would be the same if not better. And the customer could sleep better, too.
What do you think?
Comments
Post a Comment