3 ms·

The data is meaningless if it can't be trusted. Email addresses are stored encrypted and will always be kept confidential.

swombat 16y ago
That's nice, but you get to know all the details and link them up to people. You might end up meeting those people in networking events, or being asked about those people. What's to stop you from making use of that data? I'm not attacking your character or suggesting you will make use of the data, but it's a pretty tempting pool of data to sneak a look into. "Oh, so-and-so at Twitter earns that much, interesting. Holy crap, I met this guy last week, I had no idea he got screwed so badly by his current startup. Huh, looks like startup X has a fair bit of money." A lot of this is actionable data. If the startup has IPOed, you might even fall under insider trading regulations and need to make sure you don't trade on it. On a purely ethical level, you might want to consider what sort of level of access you should have to the data. Finally, there is, of course, the possible issue that someone might offer you a tidy sum of money to buy that data off you, once there's enough of it. I would not advise anyone to put their data in there as this stands.
- jbryanscott 16y ago
  "What's to stop you from making use of that data?" The vision of Ackwire is to make startup compensation data transparent to prospective employees and employers. Email addresses are encrypted and stored in a different database. It's true that they aren't hard to manually access, but this presents a reasonable barrier. The same argument applies to any web service that stores personal information. I'm sure there are administrators at Facebook, Gmail, etc. who could read all your private messages, view your photos, etc. But we assume they don't, and most of us use these services anyway. "A lot of this is actionable data. If the startup has IPOed, you might even fall under insider trading regulations and need to make sure you don't trade on it." I'm not a securities lawyer, but this doesn't seem to pass the material non-public test. Significant shareholders in a public company already have to disclose their holdings, publicly. "Finally, there is, of course, the possible issue that someone might offer you a tidy sum of money to buy that data off you, once there's enough of it." I think this is jumping the gun a bit, but any buyer would have to keep the current user agreement intact (posts are confidential and anonymous) or explicitly ask users to opt into the new terms.
  - roel_v 16y ago
    I think the point of the OP was that there is no mechanism in place to enforce your promises. It basically asks people to just trust some guy on the internet that they've never seen or met and most likely never will, and who may or may not be who he says he is. I don't have an answer, I just think it's a valid concern and if your only response is 'oh but I won't do that', well to me that shows that you don't understand the actual concern.
    - rdl 16y ago
      If the operator of Ackwire used this data in sufficiently malicious ways, I think there would be potential civil liability (at least enough to bring a credible suit; maybe not to win). I wouldn't trust a random person with this data, but I'd trust a member of a somewhat-meritocratic community like hackernews or quora, university alumni association, etc. It's fairly clear why this data is being collected, what is being done with it, and my potential downside risk is fairly capped. At some point, being overly paranoid is too much of a cost. Yes, a technical solution to this class of problem would be preferable to social trust. (and is one of the things I'm working on, actually...)
      - swombat 16y ago
        I think you're not being paranoid enough. The "HN member" in question has a karma of 18 (yes, karma does count), and I don't recognise his username. The questions you need to ask yourself is: 1) is there any potential downside whatsoever to submitting my data here? If this data is on the front page of the new york times tomorrow, and everyone I know knows it, do I stand to lose anything because of it? 2) is there any real upside to submitting my data here? (seeing other people's data has obvious upsides) Unless 2 is clearly bigger than 1, submitting your salary info here is an irrational decision.
        rdl 16y ago
        Actually, I mixed up the submitter and the creator (892 vs. 18); I agree I'd trust someone I recognized as a member of the community more than others, but I still think hn-member alone is worth something. So, I don't disagree with you.
- nolite 16y ago
  "What's to stop you from making use of that data?" This could https://ssl.trashmail.net/ https://ssl.trashmail.net/ been using it for years..
- cookiecaper 16y ago
  Probably the bigger risk is that another person will find this data appetizing and hack the server. Seeing that this is a relatively small service, it's probably not very hard for a determined person to get in, even if that means waiting until a zero-day is released and hitting before the owner gets a chance to patch.
- jacquesm 16y ago
  > I would not advise anyone to put their data in there as this stands. Second that.
jacquesm 16y ago
Encrypted email addresses don't need to be broken, all you need to do is run the addresses using the same hashing algorithm against a set of known addresses to see which ones pop up. Plenty of HN'ers have their email addresses out in the open. If email addresses are stored 'encrypted' how can they be kept 'confidential', you store them for a reason or you don't need to store them at all, hashed or not. What gain do you have from storing the hashed address?
- rdl 16y ago
  If I were creating the service, I'd want the email or some other "is an actual credible person" credential, but wouldn't store it afterward. Referrer being hackernews might actually be sufficient to not require email. The best credential would be something like "is a member of FB or LinkedIn with a sufficiently filled in profile" but nothing personally identifying. For the traditional version of this (alumni association stats), you have a trusted third party (the association), or restricting access to people on 18.0.0.0/8 or who have kerberos credentials, although logging anything personally identifying.
dean 16y ago
By collecting email addresses and verifying them, users are explicitly _not_ anonymous. If it was truly anonymous, even you wouldn't know who they are. Anyway, it's the easiest thing in the world to create a throwaway email account, so the data still can't be trusted.

swombat 16y ago

That's nice, but you get to know all the details and link them up to people. You might end up meeting those people in networking events, or being asked about those people. What's to stop you from making use of that data? I'm not attacking your character or suggesting you will make use of the data, but it's a pretty tempting pool of data to sneak a look into. "Oh, so-and-so at Twitter earns that much, interesting. Holy crap, I met this guy last week, I had no idea he got screwed so badly by his current startup. Huh, looks like startup X has a fair bit of money." A lot of this is actionable data. If the startup has IPOed, you might even fall under insider trading regulations and need to make sure you don't trade on it. On a purely ethical level, you might want to consider what sort of level of access you should have to the data. Finally, there is, of course, the possible issue that someone might offer you a tidy sum of money to buy that data off you, once there's enough of it. I would not advise anyone to put their data in there as this stands.

jbryanscott 16y ago
"What's to stop you from making use of that data?" The vision of Ackwire is to make startup compensation data transparent to prospective employees and employers. Email addresses are encrypted and stored in a different database. It's true that they aren't hard to manually access, but this presents a reasonable barrier. The same argument applies to any web service that stores personal information. I'm sure there are administrators at Facebook, Gmail, etc. who could read all your private messages, view your photos, etc. But we assume they don't, and most of us use these services anyway. "A lot of this is actionable data. If the startup has IPOed, you might even fall under insider trading regulations and need to make sure you don't trade on it." I'm not a securities lawyer, but this doesn't seem to pass the material non-public test. Significant shareholders in a public company already have to disclose their holdings, publicly. "Finally, there is, of course, the possible issue that someone might offer you a tidy sum of money to buy that data off you, once there's enough of it." I think this is jumping the gun a bit, but any buyer would have to keep the current user agreement intact (posts are confidential and anonymous) or explicitly ask users to opt into the new terms.
- roel_v 16y ago
  I think the point of the OP was that there is no mechanism in place to enforce your promises. It basically asks people to just trust some guy on the internet that they've never seen or met and most likely never will, and who may or may not be who he says he is. I don't have an answer, I just think it's a valid concern and if your only response is 'oh but I won't do that', well to me that shows that you don't understand the actual concern.
  - rdl 16y ago
    If the operator of Ackwire used this data in sufficiently malicious ways, I think there would be potential civil liability (at least enough to bring a credible suit; maybe not to win). I wouldn't trust a random person with this data, but I'd trust a member of a somewhat-meritocratic community like hackernews or quora, university alumni association, etc. It's fairly clear why this data is being collected, what is being done with it, and my potential downside risk is fairly capped. At some point, being overly paranoid is too much of a cost. Yes, a technical solution to this class of problem would be preferable to social trust. (and is one of the things I'm working on, actually...)
    - swombat 16y ago
      I think you're not being paranoid enough. The "HN member" in question has a karma of 18 (yes, karma does count), and I don't recognise his username. The questions you need to ask yourself is: 1) is there any potential downside whatsoever to submitting my data here? If this data is on the front page of the new york times tomorrow, and everyone I know knows it, do I stand to lose anything because of it? 2) is there any real upside to submitting my data here? (seeing other people's data has obvious upsides) Unless 2 is clearly bigger than 1, submitting your salary info here is an irrational decision.
      - rdl 16y ago
        Actually, I mixed up the submitter and the creator (892 vs. 18); I agree I'd trust someone I recognized as a member of the community more than others, but I still think hn-member alone is worth something. So, I don't disagree with you.
nolite 16y ago
"What's to stop you from making use of that data?" This could https://ssl.trashmail.net/ https://ssl.trashmail.net/ been using it for years..
cookiecaper 16y ago
Probably the bigger risk is that another person will find this data appetizing and hack the server. Seeing that this is a relatively small service, it's probably not very hard for a determined person to get in, even if that means waiting until a zero-day is released and hitting before the owner gets a chance to patch.
jacquesm 16y ago
> I would not advise anyone to put their data in there as this stands. Second that.

jacquesm 16y ago

Encrypted email addresses don't need to be broken, all you need to do is run the addresses using the same hashing algorithm against a set of known addresses to see which ones pop up. Plenty of HN'ers have their email addresses out in the open. If email addresses are stored 'encrypted' how can they be kept 'confidential', you store them for a reason or you don't need to store them at all, hashed or not. What gain do you have from storing the hashed address?

rdl 16y ago
If I were creating the service, I'd want the email or some other "is an actual credible person" credential, but wouldn't store it afterward. Referrer being hackernews might actually be sufficient to not require email. The best credential would be something like "is a member of FB or LinkedIn with a sufficiently filled in profile" but nothing personally identifying. For the traditional version of this (alumni association stats), you have a trusted third party (the association), or restricting access to people on 18.0.0.0/8 or who have kerberos credentials, although logging anything personally identifying.

dean 16y ago

By collecting email addresses and verifying them, users are explicitly _not_ anonymous. If it was truly anonymous, even you wouldn't know who they are. Anyway, it's the easiest thing in the world to create a throwaway email account, so the data still can't be trusted.