- cross-posted to:
- [email protected]
- cross-posted to:
- [email protected]
Some Samsung smart fridges stopped working on Tuesday due to a firmware update.
Korean outlets were the first to report the problems, which affect Samsung’s Bespoke AI line of fridges. Star News Korea, per a Google-provided translation, said that most of the affected devices were four-door fridges from 2024 or later.
They could absolutely seperate the controls of the fridge to a nice physical dial that is an accurate thermostat and keep it separate from all the smart functions.
I can’t think of any reason for needing to remotely change the temperature
OK, following scenario: you forgot to drink the verification can for your heater, but you’re on vacation. So the house gets cold and your goldfish are about to die of cold.
You dial in your Smart Fish Feeder™ to increase their rations, so that they can build up some fat. But then they’re unhealthy, and your Smart Fish Feeder™ dropped its Wifi connection anyway.
Instead you dial in your Smart Fridge™ and lower its temperature, so that it functions as a space heater now. Your fish are safe now.
But now your Smart Lock™ opens for safety reasons, as it misinterprets the sudden rise in temperature as a house fire.
When you come home your goldfish left to buy some cigarettes but never came home.
So, a very crucial feature, I’d say.
You don’t even need separate controls. Just have the compressor and temperature controlled by a simple microcontroller which gets commands from the touch screen. Thus if the touch screen crashes the microcontroller keeps the food cold.
That of course requires adding a microcontroller rather than just putting a couple of relays attached to the touch screen…
It doesn’t matter.
If they vibe code their software, even the most stupidly stable parts of the code that once made nobody should touch, ever, like say the part that decides when to turn on the compressor to reach the desired temperature, can be changed by the code generator everytime they decide to add a totally unrelated feature, and end up with bugs, possibly the stupidest shit imaginable of the “nobody would ever make this mistake” kind.
If they can push fresh firmware into it externally, they can fuck up ANYTHING at ANY TIME. It’s the ability of replacing the code in it at any time that’s the problem - if firmware can be updated by the manufactures at any time your food and your health (if you end up eating food that spoiled without you noticing) has a constantly active risk vector which is linked to the quality of the software development process of that manufacturer.
If shit isn’t exposed to the Net, shit doesn’t need firmware updates for security reasons and if a user wants to update the firmware to get some new feature then they will do it willingly in their own time, knowing that the firmware was updated and thus much more likely to notice things not working properly (and much less likely to keep on doing firmware updates in the future if the software development process of the device maker is shit and their firmware updates keep fucking up their devices).
Agree 100% on all.
There’s very basic, obvious, best-practices type stuff here that’s obviously not being done. Nothing wrong with vibe coding if you like unmaintainable spaghetti code, but whatever AI you use should have its access restricted to ONLY the part you’re modifying.
Firmware releases especially for something like a fridge should have a QC test (IE, automated unit tests) and an alpha/beta test program. That would have caught this.
USERS should be in control of software updates always. Even if the default is ‘keep my shit up to date’ user should have the ability to manually update, roll back, block updates, etc.
Plus which, to adapt an old quote-- having a refrigerator connected to the Internet say they’re adding security features is like a kid’s teacher saying they’ll now always wear a condom while educating students- strictly speaking it’s better than the alternative, but something is still horribly horribly wrong with this picture.