Jump to content


Level 1
  • Posts

  • Joined

  • Last visited

About DeAtrament

Recent Profile Visitors

724 profile views

DeAtrament's Achievements



  1. @Dave-in-Decatur Part of the problem for me is that De-Linking doesn't really assure any safety. Linkage Attacks are common and easy (that "un-de-link" - if you will) AND sensitive information can be in the note content itself (patent ideas, medical history, etc). As you said, I don't see what Evernote is purposing to offer in exchange of the elevated access to our data. And I am specifically picking on the technical details because I feel like the communication from Evernote so far has aimed to placate - just throwing out two technical terms to make us feel safe.
  2. No worries - I appreciate your knowledge of the product and community. I think things are just getting heated because privacy is such a hot topic right now (as it should be!) and asking companies questions can be like talking to wall.
  3. Ahaha - I see the confusion.... In machine learning / big data / data science community, the idea of "Learning from Data" is done by mining data to build models. This is called machine learning. In a supervised model (the type they would use if they are validating with user data), they require a training set that performs a specific way when run through the model. Consequently, when they discuss their plans for machine learning, my concern is the mechanics of how they are building their models. The mechanics of their tools would gauge how worried we should be as a community. My questions were to an evernote employee and meant to address those concerns. Instead you are arguing on their behalf without understanding the fundamentals for the technology in question.
  4. @DTLow Yes. I know of several large corporations who clone server back-ups for data mining. This is fairly common practice in the machine learning world. What makes you think I am making things up? I am a Data Science consultant and have seen exactly what I describe above. Evernote has had an impeccable reputation and simultaneously been hold notes have sensitive data. They have not described a Policy (with a capital P) that details what data in notes are used. On a side note - your tone is incredible! I am asking for details about what is happening with Evernote and pulling in industry standard information. I don't see why you would be so accusatory.
  5. @DTLow And which of these removes them from server backups? I have already deleted all notes but I want to be wiped from existence. I have opted out but as I stated in my last post (and linked), evernote employees can still access data outside of ML. Or If they create a supervised learning training set from a backup will my data be part of it? My health records are not for consumption.
  6. @benmc Interesting response... but de-linking is not sufficient and I am surprised that this is not obvious to your team. The note content may have sensitive data regardless of which account it came from. When I got my masters degree, there were several courses on ethics with User Data and Predictive Analytics. It really sounds like your team could use a refresher.... Sorry to be grumpy, but I trusted Evernote. The change in service and now the change in privacy policy is particularly scary. Being unable to opt out of human eyes (https://help.evernote.com/hc/en-us/articles/235660588) on my data is silly (yes- I understand law enforcement compliance is necessary and that is fine but your privacy policy allows for much broader usage). So, is there a way to remove all my data from Evernote and prevent all employees from looking at it?
  7. Wow.... Well, I have only read the first 9 pages (so forgive me if the other 9 pages answer these questions - still reading).... This is a pretty disturbing event for me. I had been moving away slowly anyway (missing features and lackluster community response to requests - see dark mode....). As many cryptographers (like the preeminent bruce schneier) state: the Cybersecurity and data privacy are all about trust. My trust in Google consumer products is zero and the last two years with Evernote have losing trust with their consumers. I understand that Google CloudServices is a different business model altogether, but the ability for one disgruntled employee (from either company) to dump unencrypted data is frightening. Especially since the long-requested feature of notebook level encryption is a "low-priority" Finally, I read earlier in this thread that Evernote will be data mining my data, first page Rich Tener: ... why is this needed? Surely existing indexing techniques are sufficient. Data mining has roughly four tasks: clustering, classification, dimensional reduction and regression. The last three are entirely impractical for search optimization and Evenote does not need to use clustering to provide search power.... so again, why in the world is this happening? The "three laws" of data privacy are nice fluff but do not define a transparent data security policy. Can Evernote please describe their internal data policy and how my (our) data is being used in ML?
  • Create New...