I think it's best summed up by a tweet I saw last year:

Non-magic users: collect crystals, call their pet a familiar, draw pentagrams

Magic users:the most magical things I keep in my house are rocks, and I keep a hammer next to them in case they act up

Source: https://twitter.com/TooMany_Monkeys/status/11205616344125849...

That's a play on the original[1]:

> Tech Enthusiasts: Everything in my house is wired to the Internet of Things! I control it all from my smartphone! My smart-house is bluetooth enabled and I can give it voice commands via alexa! I love the future!

> Programmers / Engineers: The most recent piece of technology I own is a printer from 2004 and I keep a loaded gun ready to shoot it if it ever makes an unexpected noise.

[1] https://biggaybunny.tumblr.com/post/166787080920/tech-enthus...

Thanks! I hadn't seen that, but yes, the magic phrasing is clearly derivative. I wonder why they bothered given that it was pretty clear what the metaphor was about.

Well, it was part of a Twitter thread on "what if magic was like IT?"

I am currently reading "Sourcery" by Terry Pratchett. Your remark is literally one of the plot points of the story.

I love his observations about the nature of magic and power. Terry is top of my list of people I wish I'd met while they were alive, he had a tremendously humane view of people and what makes for a good world.

The R in IoT stands for “Reliability”.

Tangential, but I think it’s funny so whatever. Some smart-arse at work came up with this one:

“Security: we put the ‘no’ in ‘innovation’.”

Just like the S for "Security".

See also: @internetofshit on Twitter

For what it's worth I have a Roomba and you can still press the button on the vacuum itself and it works just fine. It's just the (back-end of) the app that's down, since they host it on AWS. Doesn't seem too unreasonable.

When your Roomba and your phone is on the same network, yes it is.

It's ridiculous. I don't understand why so many of these things need to connect to cloud services just to work.

I mean, surely computing power is cheap enough these days that you can build all the intelligence you need into the device for day to day operation and simply use internet connectivity to get updates, such as security patches, and the like?

Maybe the compute side of things could be a pluggable module that could be upgraded independent of the rest of the device if you ever do need more processing power. This would work particularly well for larger devices such as robot vacuum cleaners.

I find it indefensible that some data centre going down in Australia (or wherever) could render something as basic as a doorbell, or even a vacuum cleaner, inoperative.

>I mean, surely computing power is cheap enough these days that you can build all the intelligence you need into the device for day to day operation and simply use internet connectivity to get updates, such as security patches, and the like?

I work on embedded systems that can not fail when the internet connection fails, and it's bonkers just how much you can accomplish with a <$5 microcontroller or even a <$10 chip running linux. But the tide of the industry is pulling toward remotely hosting as much as possible.

Even if it doesn't strictly need networking, the desire for a companion app, firmware upgrade workflow and/or business desire for surveillance often justify it. Once you have network connectivity for any reason, well, then you're not not going to be phoning home to a surveillance system running in AWS anyway. If you have a companion app, making sure you can get a connection between the phone and device is going to have a fall-back path that runs through your servers because even self-styled techy people don't always understand networking.

Now that your networked device is already sending all the data home anyway, why not do more work serverside? Add in the prevalence of developers that can't be arsed to learn to develop for a microcontroller. Or those that can only get by in the very-handheld embedded environments. Now your design is starting to take shape, and you need to do some more number crunching. Do you invest in learning to leverage esoteric features of your microcontroller? Add a DSP to the BOM? Or just say it's easier to do on the serverside with comparatively infinite processing resources and the ability to rush development because features and bugfixes only require updating servers rather than pushing firmware to the fleet?

My theory is that consumers already have wifi networks and know how to enter a wifi password (or push a wps button). They don’t need to buy some new base station and then worry about if it speaks the right protocol as all their other devices. Historically there wasn’t a ubiquitous, push button, cross-vendor wireless protocol that fit the bill. I’m really hoping that everything actually moves to local comms over zigbee.

People don't refuse to buy them for doing it, ... in part because there is seldom any disclosure of it.

So “smart”