“Internet of Things”: Probably Say “No” For Now

There is much excitement these days about the coming of the “Internet of Things” (a term coined by Kevin Ashton in 1999, but now coming true.

I’ve been interested in this stuff since before Ashton coined the phrase,

There is so much cool that can be done with smart environments, how could be not be excited?

However, along the way, things languished in unimaginative and short sighted apps–honestly, I don’t need to automate the light switches, they work just fine–and a lack of contact with the real world.

My favorite rant in this area is about the plethora of “intelligent” rooms that sense the wishes of the inhabitant.  Not the singular.  One person per domicile.

So, it knows your schedule and anticipates when to start the coffee, etc.  I adjusts lighting and music to the tastes of the (one) user. And so on.  (Also, these systems tend to be given the personality of a servant–a psychologically and sociologically troubling fantasy.)

In real life, for most real people, there are more than one person in the family.  So you can’t talk about “optimizing” for the one user, that makes not sense.  And optimizing for multiple people is hard–and in any case is what relationships are about.  Not really what your thermostat should be doing.

Example of why this is hard:  many services offer to take my collection of music recordings, crunch on them, and then make recommendations for me.

Here are some real world problems with that concept.

First, my collection extends decades back.  My tastes have changed over the years.  (Essentially, I’m not “one person” as far as the algorithm would have to assume.)

Second, the collection was merged with my spouse decades ago.  She has her own tastes, which also have evolved.

So how would machine learning deal with this?  I have no idea, but I’m sure I don’t care.

Where was I?  Oh, right.  Internet of Things.

What’s coming out this summer are a bunch of things from Internet/Mobile app people in alliance with appliance maker and marketers.  The technology is based on home scale internet things:  moderate bandwidth, light security, almost no local storage or cycles.

For some reason, people seem to think that hooking all this up to the Internet, kind of like a phone,  will be OK.  Are you nuts?

Aside from the sheer insanity of letting Google anywhere near my thermostat (to pick one example), there are so many reasons I don’t want my house full of low grade computing.

For example, we hear about cases such as a hack attack that dropped malware on home storage devices that secretly mined cryptocoins.  This went on for months because-wait for it-random home owners in Taiwan do not take time monitoring CERT bulletins and emergency patches for their home appliances.

When I read this story I realized that, even though I am experienced and well informed, I have no idea how many of the devices in my house may have bluetooth or wifi, nor if they could be hacked.  How would I know?  Why would I want to know?

Glancing at the web documentation from Google Nest, I found that I was supposed to be reassured by the fact that they use openSSL (!) and will ask before sending data to an app on your phone.  Problem solved!

If you want and even more thorough walk through, I refer you to recent books on this topic.

Until there is an application I really need or want, and a system that is really isolated from the Internet, I’m saying “no” to the Internet of Things.