3 ms·
I'm firmly in the "hell no" camp. I realize how these devices can benefit me. As an embedded designer, I eagerly thought about these types of devices 15 years a
by mindslight 12y ago
I'm firmly in the "hell no" camp. I realize how these devices can benefit me. As an embedded designer, I eagerly thought about these types of devices 15 years ago.
The problem is that we, the technical community, got hoodwinked away from open protocols into thinking that closed systems based on dumping data to someone else's server is an acceptable basis for much of anything. Manufacturers are quite happy to foist these proprietary solutions on us because it lets them create lock in, differentiate on superficial features, keep hiding sloppy development, and derive licensing fees for any third parties that think they can help analyze your data (with your consent, or not). The "average" user sees no difference because it's all magic boxen to them, and we've done them a great disservice by shirking our professional duty and being "yes men" for this backwards environment.
My rejection doesn't come from "fear" of "risks" - it comes from the rational assessment that once this data is out there, especially in such collated form, you've lost control of how it is used against your interests. The only way to brush that off as "fear" is to say the future where cloud-big-brother knows more about me than I do is inevitable. So rather than being scared, I should learn to love the rats on my face. I am hoping this is wrong.
Furthermore, reliance upon any technology makes it de facto mandatory which becomes much harder to extricate yourself from to switch to something with a less hostile architecture. Especially as more open architectures will be less polished due to less funding as creating freedom isn't lucrative.
And it's not just the NSA - there's the much larger and better poised threat from the commercial side as well, via insurance / background check companies. How long until "I have to do 500 more steps today so I don't have to pay $50 extra on my health insurance" ?
(PS Never mind getting on the upgrade treadmill for fixtures who's lifetime has traditionally been decades, just for software bugfixes and upgrades. And the input response time on many of these products - looking at you Nest - show that it's amateur hour)
- justincormack 12y agoThe insurance thing us already happening eg http://www.citeworld.com/article/2450823/internet-of-things/appirio-fitbit-experiment.html http://www.citeworld.com/article/2450823/internet-of-things/...
- mindslight 12y agoAnd that's the pitfall of trying to come up with something that sounds concerning yet plausible.. I've explored this subject from first principles, but that comment will probably trip too many mind virus defense systems to be easily received. https://news.ycombinator.com/item?id=7928484 https://news.ycombinator.com/item?id=7928484
- darkmighty 12y agoHealth insurance is going to advance it's ways of knowing about your risks wether you like it or not. And denying them ways of estimating your risk will likely just boost your cost beacuse of high-risk outliers. If you don't think it's fair that some people will have a much higher health charge you should not argue against companies gathering insurance data but the fundamental problem -- probably by introducing universal health care legislation. It's also a fallacy to suppose that companies have to know your data at all (assuming they would even have such a huge interest in selling your stool sample data to risk their reputation). Encryption was created to solve exactly this privacy problem, and homomorphic encryption goes even further (although it's still early in development). Worst case, you just run the analysis on your own desktop. As technolgy progresses it's only natural our inference and survaillance prowess will increase and the battle can't be won by restricting technology itself. The only way out is for society to put proper value on privacy as a right and put proper regulations and legislation to defend it. (although the NSA scandals have shown most of the US doesn't care)
- icebraining 12y agoIt's also a fallacy to suppose that companies have to know your data at all (assuming they would even have such a huge interest in selling your stool sample data to risk their reputation). Encryption was created to solve exactly this privacy problem, and homomorphic encryption goes even further (although it's still early in development). Worst case, you just run the analysis on your own desktop. But that's exactly the point. The "hell no" comment wasn't directed at any toilet with sensors - it was directed at a toilet implementing the Internet of Things. If you run it on your own desktop, you're not talking about the same thing at all.
- mindslight 12y agoYour second paragraph is exactly what I'm getting at - companies do not need to know your data, and analysis should be run locally with software under your control. You would then be free to choose from various non-exclusive analysis options via informed consent. However, I think it's deluding oneself to sign up for today's proprietary share-everything services, and expect them to magically transform into privacy respecting ones as time goes on. For these type of applications, mentioning homomorphic encryption comes from that same wishful thinking that a third party will work to preserve your privacy more than you will.