Telemetry can be done right by letting end-user to see an example of the data that's going to be sent and then also audit the actual reports that were sent. Transcribed into a human-readable format with some comments about what the values mean - I can see the binary data myself, thanks. And a way to opt out if the data is found to be actually sensitive.
Basically, an opt-out switch and "sure, but give me a full transparency report" mode that a "power user" can optionally access if they happen to care. This is rarely the case, and not the case with Windows.
I agree with the desire for complete transparency in data shared, but eventually this will run afoul of a number of practices which rely on cat-and-mouse games where one side doesn't have complete information (an obvious example being precisely what sorts of data Windows Defender uses/sends for malware found by heuristic, presuming it doesn't just ship the entire binary off), to say nothing of the conspiracy theories that would arise every time they ever added some kind of data collection.
I don't have a good answer for this friction in general - complete transparency would only fly until a piece of malware that circumvented every heuristic and strict rule caused a major incident, whereas nearly complete opacity is where we are now.
(There's also the conflict of their lack of recurring revenue for Windows 10, which is why I'd anticipate them eventually at least floating the option to disable the "end user data as payment" harvesting for an annual fee, but...)
https://www.microsoft.com/investor/reports/ar16/index.html
https://www.fastcompany.com/3064030/satya-nadella-on-microso...
I couldn't find the older articles about how Ballmer was complaining that they were missing out on the personal data gold rush that Google and Facebook were experiencing.
Google's entire business is built around selling the presentation of unwanted information around the organic content you want. That works great in a visual medium where the extra information requires no more of your time and only minimally more of your attention. Alexa and Cortana and Siri shift that interaction to audio where it becomes incredibly difficult to insert even a tiny fraction of the surrounding ad content search engines used to deliver visually. Alexa and Cortana and Siri break google's business model because they break the ability to spam customers with ancillary information, and they can do that because neither Amazon nor Apple nor Microsoft derive the bulk of their revenues from ad sales. This is a battle of business models, not a quest for the other guy's business model.
So yes, Microsoft is building Cortana, and no, Cortana isn't about becoming an advertising powerhouse it's about blocking their two ad-fueled competitors (the best discussion I'm aware of of google's efforts to fight back against this agent-driven threat to their ad-based business model is [0] )
[0] https://stratechery.com/2016/google-and-the-limits-of-strate...
Is there an article where I can read more about that?
There were earlier articles that talked about the frequency of audio dumps (every 30 minutes) but I wouldn't call them as reliable sources as Ars or Microsoft.
https://en.wikipedia.org/wiki/Windows_Error_Reporting#Privac...
- You can't disable telemetry completey. Some people want it, some don't. Let the user decide.
- Microsoft does collect an insane amount of data.
https://docs.microsoft.com/en-us/windows/configuration/basic...
I would be curious how many users want telemetry sent to Microsoft. It seems the two largest buckets, by far, would be those who don't care and those who don't want it.
The problem with opt in is that most people, where most crashes occur, would never read what the prompt says asking them to send the data, they would just press the x in the corner because a dialog box popped up they didnt ask for. "What did that box say. I dont know. Why did you close it. Because I didnt want it."
You're describing a UX issue with the worst possible way this could be implemented. What chafes me is not that I'm not asked for permission for every crash report, but that I don't have the option to disable crash reports at all.
That's right, because all these unsolicited prompts and notifications are a waste of people's valuable time. A well-designed OS should aggressively keep quiet, stay out of the way, and err on the side of user privacy.
Your Windows crashing is an issue strictly between you, Microsoft, and the makers of whatever software and hardware you were using when the crash occurred. Absolutely nothing to do with anybody else.
I want your data sent to Microsoft too, for the same reason. I don't want my data sent to Microsoft, though.
Developed capacity may not equal intent, but if you can be ordered to do secret work and make it so anyway, the distinction is irrelevant.
As long as you're clearly informed about what kind of information it's collecting and you can turn it off (and ideally tune it) it's fine with me.
The opt in piece is important because you need large amounts of data to detect emerging issues. It lets you distinguish between flukes and trends. This enables the quality of service that the user expects.
It is simple, data is becoming worth something because you now have the ability to analyze all the data you could ever collect. And so, everybody naturally jumps on the wagon because of the inherent profit motive.
If a computer was an appliance, you would buy it and it would never need to phone home. Now we both know that there is some need for updates and it might therefore be a good idea to classify computers as something other than appliances.