3 ms·

>Useful features attract users. Users are a good thing. I think its a bit of a stretch calling this feature useful but I suppose you can re-define anything to

>Useful features attract users. Users are a good thing. I think its a bit of a stretch calling this feature useful but I suppose you can re-define anything to be useful. Ads are useful too, because you can save money by buying the exact thing that you need. Its too bad that people are downloading ad-blockers by the millions. Google snooping on emails and extracting purchase related information to stick inside a dossier on you, is in no way the same as you voluntarily, creating a label for an email which they have already scanned anyway. An advertising company wants to know what you're buying, how frequently, and how much you spend on things. That is like a goldmine for them. The labels, are almost noise because its data that is an abstraction away from telling advertisers what will get you to open your wallet - which is the end goal. >Like I can apply the exact same argument to email labels: I don't think you can, and I could not follow the argument you made.

joshuamorton 7y ago
So I've rehashed this argument a lot recently: Google has to snoop on your emails. Every email provider does. Spam classification, automatic labelling/filtering, things like the "we think this is phishing" banners, etc. all require some form of analysis. You could (and seem to be) calling some of that "snooping", but some of it is acceptable. So the question is where is the line drawn? > I think its a bit of a stretch calling this feature useful but I suppose you can re-define anything to be useful. It's used to alert you on upcoming packages you may receive or bills you need to pay (if you have an android phone/google assistant). I think that's useful. > The labels, are almost noise because its data that is an abstraction away from telling advertisers what will get you to open your wallet - which is the end goal. Statements like this amaze me! You really think that manually curated labels would be less useful than automated inference about an email? Supervised learning is basically always easier than unsupervised, and email labels are, quite literally, labels for the data. It's a model-maker's dream! > I don't think you can, and I could not follow the argument you made. That's because you personally find labels useful and non-invasive, but feel that this other feature is not useful and is invasive. That's the only difference. If you felt you got value from the purchase history tools, which again, some people do, or if you didn't have an unfounded belief that everyone was lying to you and this data was being used in ways it isn't, you wouldn't find what is essentially automatic labelling of some emails invasive. The argument is that labels are useful to some people. Google could misuse the data derived from labels. Similarly, purchase history info is useful to some people. Purchase history data could be misused by Google. If you trust Google to not misused labels, then why don't you trust them to not misuse purchase history? If you do trust them in both cases, this doesn't matter, and if you don't trust them at all, why are you using an email client that you think is completely lying about how they're using your data?
- la_barba 7y ago
  >So I've rehashed this argument a lot recently: Google has to snoop on your emails. Every email provider does. Spam classification, automatic labelling/filtering, things like the "we think this is phishing" banners, etc. all require some form of analysis. You could (and seem to be) calling some of that "snooping", but some of it is acceptable. So the question is where is the line drawn? Its not about a line. Advertising functionality is simply not a core function of an email service. It is something that is tolerated in exchange for a free service. That is not the same as ceding control and letting a vendor do anything they wish simply because a vendor claims its useful. >It's used to alert you on upcoming packages you may receive or bills you need to pay (if you have an android phone/google assistant). I think that's useful. Well, how does storing the item purchased and the dollar amount and other purchase details help with tracking anything? AFAIK Amazon for e.g. has stopped putting in tracking information in their emails. There could be a useful function as far as bills are concerned, but its weird to take an email about a bill and then generate another wasteful notification about the same bill. Or to take a package has shipped email and generate another notification. But that is a separate discussion. Also, if Android and google assistant users have chosen to enable such features with informed consent, I don't see why every single gmail user has to have this this enabled. I mean that in a rhetorical way, I do see why an advertising company would do such a thing. Anyway, I'm simply registering a protest and I encourage other users to do the same. Several invasive Google policies were reversed due to privacy push-backs, and so all hope is not lost. >The argument is that labels are useful to some people. Google could misuse the data derived from labels. Similarly, purchase history info is useful to some people. Purchase history data could be misused by Google. If you trust Google to not misused labels, then why don't you trust them to not misuse purchase history? If you do trust them in both cases, this doesn't matter, and if you don't trust them at all, why are you using an email client that you think is completely lying about how they're using your data? That to me, is needlessly polarizing a nuanced issue. "Either you trust us completely or you don't" is not really an argument worth having is it?

joshuamorton 7y ago

So I've rehashed this argument a lot recently: Google has to snoop on your emails. Every email provider does. Spam classification, automatic labelling/filtering, things like the "we think this is phishing" banners, etc. all require some form of analysis. You could (and seem to be) calling some of that "snooping", but some of it is acceptable. So the question is where is the line drawn? > I think its a bit of a stretch calling this feature useful but I suppose you can re-define anything to be useful. It's used to alert you on upcoming packages you may receive or bills you need to pay (if you have an android phone/google assistant). I think that's useful. > The labels, are almost noise because its data that is an abstraction away from telling advertisers what will get you to open your wallet - which is the end goal. Statements like this amaze me! You really think that manually curated labels would be less useful than automated inference about an email? Supervised learning is basically always easier than unsupervised, and email labels are, quite literally, labels for the data. It's a model-maker's dream! > I don't think you can, and I could not follow the argument you made. That's because you personally find labels useful and non-invasive, but feel that this other feature is not useful and is invasive. That's the only difference. If you felt you got value from the purchase history tools, which again, some people do, or if you didn't have an unfounded belief that everyone was lying to you and this data was being used in ways it isn't, you wouldn't find what is essentially automatic labelling of some emails invasive. The argument is that labels are useful to some people. Google could misuse the data derived from labels. Similarly, purchase history info is useful to some people. Purchase history data could be misused by Google. If you trust Google to not misused labels, then why don't you trust them to not misuse purchase history? If you do trust them in both cases, this doesn't matter, and if you don't trust them at all, why are you using an email client that you think is completely lying about how they're using your data?

la_barba 7y ago
>So I've rehashed this argument a lot recently: Google has to snoop on your emails. Every email provider does. Spam classification, automatic labelling/filtering, things like the "we think this is phishing" banners, etc. all require some form of analysis. You could (and seem to be) calling some of that "snooping", but some of it is acceptable. So the question is where is the line drawn? Its not about a line. Advertising functionality is simply not a core function of an email service. It is something that is tolerated in exchange for a free service. That is not the same as ceding control and letting a vendor do anything they wish simply because a vendor claims its useful. >It's used to alert you on upcoming packages you may receive or bills you need to pay (if you have an android phone/google assistant). I think that's useful. Well, how does storing the item purchased and the dollar amount and other purchase details help with tracking anything? AFAIK Amazon for e.g. has stopped putting in tracking information in their emails. There could be a useful function as far as bills are concerned, but its weird to take an email about a bill and then generate another wasteful notification about the same bill. Or to take a package has shipped email and generate another notification. But that is a separate discussion. Also, if Android and google assistant users have chosen to enable such features with informed consent, I don't see why every single gmail user has to have this this enabled. I mean that in a rhetorical way, I do see why an advertising company would do such a thing. Anyway, I'm simply registering a protest and I encourage other users to do the same. Several invasive Google policies were reversed due to privacy push-backs, and so all hope is not lost. >The argument is that labels are useful to some people. Google could misuse the data derived from labels. Similarly, purchase history info is useful to some people. Purchase history data could be misused by Google. If you trust Google to not misused labels, then why don't you trust them to not misuse purchase history? If you do trust them in both cases, this doesn't matter, and if you don't trust them at all, why are you using an email client that you think is completely lying about how they're using your data? That to me, is needlessly polarizing a nuanced issue. "Either you trust us completely or you don't" is not really an argument worth having is it?