Cats, dogs starve as web-connected chow chute PetNet plays dead

There are some interesting points this brings up expanding beyond the whole internet-of-things/robot-ification of the world.

a) Google, Amazon, and Facebook mean really bad products can get global reach instantly. Even Walmart had pretty big hurdles which products had to cover. There are some controls in place, but the problem has to be pretty large to register.

b) Products accidentally killing humans let alone animals is a very common occurrence, see global vehicle deaths. Plenty of stories of very dangerous products litter Amazon reviews.

c) The companies making a lot of money from bad products are the online marketplaces. Somehow this hasn't hurt their brands too much yet, which is interesting.

d) There is a market need for a better rating and review process for everything. Given the more reviews are viewed publicly the greater their perverse effects are so maybe it won't even be publicly viewable.

I spend a lot of time thinking about the d) problem. The scarier thing is that this already exists to a large extent in many businesses using machine learning methods and labeled fraud detection. What happens in a future IOT where you product doesn't function on purpose?

A pet feeder that needs Internet access? Come on. This is something that's a job for a dumb timer, or, at most, a simple microcontroller.

But leaving more than enough food lying around to be eaten ad libitum is a recipe for a fat cat, as they will not stop when they've had enough.

Possibly including the mice and rats attacted by all that food? let alone flies.

I mean presumably they might have had some web-ui functionality to change feed-settings - but why that wouldn't just update a cron-tab is beyond me.

I wonder if security systems work the same way. Robbers want to take in a smart home so they just cut the coaxial line and the whole thing goes 404, unlocks all the doors, turns off monitoring and can't alert emergency services.

Hopefully they have the decency to feed the dog.

Even though this would be more reliable, you couldn't control it from your phone, which seems to be a selling feature for all of these badly designed IOT devices.

Somebody should design a timer module that can be controlled via bluetooth and (optionally) via Wifi.

Then you could program the timer from your iPad or whatever, but if connectivity is lost, the local controller would retain the settings and continue to work until connectivity is re-established.

Why are they using a software development process that optimizes for time to market when reliability is critical to their business? The former is somewhat acceptable for a startup in the web but incredibly irresponsible for developing hardware.

It's not like it's impossible to ship software that rarely has major bugs even though there are no consequences to the real world compared with the space shuttle. We've done that during the entire console era until consoles became internet connected devices.

That company would also be the one that gets all the surveillance data and would be the one that gets to make all of the money.

This is part of the problem of IOT: lock-in is critical to the business model.