There are countless and mounting problems with our so-called “perfect” technology having to deal with imperfect real life. Here are two example problems with my use of Apple technology during the coronavirus pandemic here in the US.
I own a number of Apple devices; several Macs , an iPhone and a pair of iPad Pros. The biggest issues I have are shared by the iPhones and the iPads having to do with their biometric locking features, TouchID and FaceID.
If I use TouchID on the iPads, over time after initial configuration both iPads will exhibit increasing failures to unlock when the front button is pressed. No matter how many times I erase and then reprogram touch ID, unlock remains erratic. I believe the problem is TouchID wants to decode “perfect” fingerprints, and mine are now far from perfect.
Based on recommendations due to coronavirus, I wash my hands extensively, both in length of time and number of times during the day, especially if I (seldom, as it turns out) have to leave home. To clean up I use Softsoap manufactured by Colgate-Palmolive. It does a great job, but one problem is it dries out the skin, especially on my fingertips. This leaves a lot of cracking and tiny flakes of skin. That detritus makes fingerprint matching with TouchID very difficult, requiring multiple touches/button presses to open the device. Either it eventually succeeds, or else I just punch in the password (which for me is at least eight characters because I’m that paranoid). In the end I decided that, because I don’t go out with either device, I would disable both password and TouchID. Now, when I want to open the iPads I just push the button and I’m into the device. With a cover on it, when the cover is opened I’m immediately into the device. Securing the device with TouchID while at home is too much of a burden.
Which leads us to the latest and greatest iDevice biometric lock, FaceID. In practice, on my iPhone 11, it works better than TouchID does on any Apple device that supports TouchID. When my phone can see my face at the position where I would comfortably look at the screen, it rarely fails. When it does fail it’s because I’ve either pushed my glasses up onto my forehead or I’ve taken them off completely. The second time I push up on the bottom edge of the screen, it succeeds. So only two attempts to unlock, which is far better than TouchID. The only time I have to enter my password is when it explicitly asks due to its built-in security timeout (after so many days you have to enter the password to get into the iPhone).
FaceID worked great until coronavirus and having to wear a mask. Now it fails repeatedly while I wear my mask. The only saving grace of failure is that FaceID fails fast. And now that I’m used to it failing with my mask on, I’m ready to enter the password. The time to fail first and then enter the password isn’t that much longer than two successive failures, where the second FaceID attempt normally succeeds. Failure due to a face mask underscores the limitations of the existing system.
I say existing, because rumor has it that a better FaceID that can work with a mask on is being engineered and will be installed with the next point release of iOS. In the meantime I can live with the way it behaves because, well, I really have no choice. I won’t remove the device lock because I always travel away from my home with my iPhone and I do want it locked down as much as possible.
I look at the failure of both of these security systems as yet another example of many. These failures illuminate the deep limitations of machine learning that backstops these features running on the devices. It also points to a broader level of poor quality and too many failures in our current technology in general. We spend too much on tech that doesn’t work nearly as well as advertised, and we’re putting it into critical parts of our system of systems, leading to overall fragility in our daily lives. I’ve begun to look at our world as a house of cards. Trouble is, you can’t go anywhere else in the world without running into the same issues. If we ever do have a machine apocalypse it’ll be because the machines themselves malfunctioned, not because they rose up in revolt.
Personal Postscript
Back in October 2019 I spent a premium sum on this particular iPhone 11 Max Pro just for the privilege of trying it out in real life. After seven months of continuous use I now consider that purchase a mistake, especially given how little I’ve come to use a number of its vaunted features. I won’t spend that amount of money again on an iPhone, or any smart phone for that matter, especially an Android. My next iPhone will be “down market”, a much cheaper iPhone 12. I’d considered the latest iPhone SE, but it has TouchID, which I don’t care for all that much anymore. I’m looking at a “low end” iPhone 12 with a smaller screen but with as much battery as possible inside the device. Now all I have to do is bide my time until October of this year.
Radarwitch
My wife’s Samsung phone will unlock when the cat looks at it. No, the Mrs. doesn’t have a beard.
The AI that this and other technology uses is so far from an acceptable level of reliability that it shouldn’t even be out of Alpha testing yet. But we rely on it to monitor Internet posts and even drive cars.
I don’t use a smartphone of any kind, and get laugh at because of this. By people who spend their whole day swearing at the technology they think is so wonderful as it repeatedly refuses to perform the requested function. Such as unlocking so you can get at the other functions.
Just go with the password right from the start and save yourself the hassle.
LikeLiked by 1 person
As for the password I believe you’re right. I’ve already sworn off TouchID, and FaceID may be next. As for so-called “smart” phones, they really don’t need to be smart. They just need to be repeatably reliable, which means as few functions as possible to minimize failure modes and attack surfaces in the software. Here’s a short list of what my hand holdable computer should have:
Need-to-haves
* A big fat battery for 48 hour normal usage
* A fucking 3.5mm headphone jack
* A universal charging port, USB 3
* WiFi
* Voice with ability to record conversations and a whitelist feature to block robocalling
* Phone book to support whitelist robocall blocking
* Text (SMS) messaging
* Email
* Clock (for setting up alarms)
* Simple text pad (raw text only) that supports more than one text document
Nice-to-haves-but-won’t-be-missed
* Web browser that is standards compliant with ability to disable JavaScript
* Video streaming app (Netflix in particular)
* Camera app that supports phone camera
* Bluetooth
That’s it. No AI. No games. No social media support at all. No Google (YouTube especially). Everything should be encrypted. All electronic messaging on and off the device should be encrypted. Sounds dull; dull is great. It doesn’t have to have the fastest SoC. It doesn’t need 5G because 5G is a power hog as well as easily attenuated by damn near everything in its path.
5G doesn’t spread coronavirus or cause cancer, but it does suck money out of your wallet a lot faster than everything that’s come before it. It costs more up front for the handset and the carriers will charge a monthly premium for the privilege of using it on their network, if it’s installed. Here in the US it’s not installed anywhere of significance.
All software on that phone should have no need whatsoever to call back to the mothership unless you need to make a phone call, check the local time, or download mail. For web and video, only reach out to retrieve web and video data.
It will never be built because there’s no way to monetize such a device by the people who want to sell it to you unless you pay for the true cost of such a device up front.
LikeLiked by 1 person
Yup. We see it all the time in tech: build it because we can, and sell it to make money. Never mind need or usefulness or efficiency. Just the-latest-is-the-greatest, even if it’s useless.
We have people burning cell towers in Quebec because they think they’re 5G and dangerous. No joke. Meanwhile out here in the bus we have NO cell service because who needs it? Well anyone who gets stranded out here far away from cities on roads that have low traffic volume. You know; exactly the situation where a cell phone would be useful, as opposed to in cities where landlines make more sense.
As an engineer one common thread I’ve seen throughout the years and across many types of tech is that multi-purpose means “doesn’t do anything very well”.
LikeLike