Hacker News new | ask | show | jobs
by ljd 4080 days ago
This has been around for several years. I tried to get a large company I worked at to use it to complement our internal machine learning algos.

We decided not to send data to google because we weren't sure about how it would be used or stored. I wonder if that was paranoia from management or if anyone else on HN feels the same way about using a third-party for intellectual property initiatives.

5 comments

This is exactly why Google should be researching fully homomorphic encryption like crazy, yet I haven't heard anything about them ever considering researching it. I think at the very least FHE is a slightly more practical research area than say researching quantum computers, and could have a bigger positive impact on Google's cloud business in the near future.
Homomorphic encryption available on Google BigQuery, if you need: https://code.google.com/p/encrypted-bigquery-client/

For others, strong certifications of privacy are enough, HIPAA among others: https://support.google.com/work/answer/6056694

Probably paranoia, unless that large company was a competitive cloud computing business. The TOS [https://cloud.google.com/terms/] has two sentences that pretty well eliminate any risks about the use of data and it also has a few paragraphs about storage.
You should take a look at Azure Machine Learning, a product I work on at Microsoft: http://azure.microsoft.com/en-us/services/machine-learning/

The Azure organization, just like all of MS, takes privacy and data management super seriously (http://azure.microsoft.com/en-us/support/trust-center/privac...). There's tons of documentation on our policies which are clearly laid out.

For those of us too lazy to read TOS, why is Azure's product less of a privacy or security concern than Google's?
>The Azure organization, just like all of MS, takes privacy and data management super seriously

Hah! So is that why Microsoft fired its chief privacy adviser in 2011 for telling a group of MS' National Technology Officers that “If you sell Microsoft cloud computing to your own governments then the FISA law means that the NSA can conduct unlimited mass surveillance on that data”, because they take privacy seriously?

The other worry of course is that google suddenly decide to shut it down like they did with quite a few APIs such as the translation API (After some back and forth they actually kept translation API but charge quite a lot for it).
Google didnt shut down translation API. They just started charging for it and people who were used to getting it for free considered it dead.

Personally, I would prefer that Google charge for its services. Its certainly better than advertising or trying to monetize users' data.

Hmm... if I recall correctly, originally they were going to shut it down completely. (I certainly see some old articles about that, from a brief search.)

I think there was enough pushback from people willing to pay that they decided to keep it open as a paid service instead.

Correct.

> UPDATE June 3: In the days since we announced the deprecation of the Translate API, we’ve seen the passion and interest expressed by so many of you, through comments here (believe me, we read every one of them) and elsewhere. I’m happy to share that we’re working hard to address your concerns, and will be releasing an updated plan to offer a paid version of the Translate API. Please stay tuned; we’ll post a full update as soon as possible.

http://googlecode.blogspot.com/2011/05/spring-cleaning-for-s...

There is no reason to believe they don't monetize the data coming in from paid services. In fact, that data is even more valuable than random freebie users.
The contract/EULA is a reason.
Is it in there, then? That they won't use their data? Even if it is, it is proprietary, freedom-restricting software so you wouldn't have a clue whether they are coming up with their part of the aggreement.
If you obfuscated the labels and stored those elsewhere how could Google make use of it, really?