13 ms·
While I understand the need for reliable and granular information on your users, it should technically possible to get most of that without selling your users o
by atoav 8y ago
While I understand the need for reliable and granular information on your users, it should technically possible to get most of that without selling your users off to third party services whose practies you never checked for yourself.
Furthermore it should be possible to sell ad space without directly embedding potentially malicious stuff from foreign whose content might change on such a granular level even you as the site owner cannot verify what is served to your visitors.
It is possible to use webfonts without loading it from a foreign servers.
It is possible to verify real users without training the AI models of monopolist webcompanies.
It is possible to link to your Facebook page without embedding a Facebook tracker in an icon.
It is possible to get comments without giving these to other parties.
These are all decisions you take when you decide whether you really value your users privacy.
Of course sometimes you don’t have much of a choice, but I saw sites which decided against all points raised above, and list about 30 trackers that are enabled per default and still claim they value their users privacy. So not only do you not value my privacy, but you also lie to me.
I avoid sites like these like the pest and will close that tab before reading anything.
- majewsky 8y ago> It is possible to verify real users without training the AI models of monopolist webcompanies. Honest question: How?
- kilburn 8y agoI'm pretty sure that "verify real users" was about captchas. Solving that problem in general is probably impossible, I grant you that. However, reaching a good-enough solution for a particular site is often doable. In increasing order of strictness and complexity: - Use hidden/visible field shenanigans - Ask questions your audience should be able to answer (chess-captchas, maths-questions, etc.) - Require registration with e-mail validation - Require registration with SMS validation - Make that part of the site invitation-only - Use some kind of trust-based system (e.g.: users can invite other users) - Manually approve stuff - Ask for ID scans and manually check them - Combinations of the above Unless you are a juicy enough target (not many sites are), just a few measures will get you to that good-enough point. Of course, implementing any of the above will be harder than slapping a recaptcha and calling it a day ;)
- Kalium 8y agoI find it interesting that the proposed solution to "How do I not use a CAPTCHA?" is "Reinvent a CAPTCHA". Although "Handle very private information instead" is an uncommon twist.
- atoav 8y agoYou misrepresented the problem. The problem was “How do I not use a CAPTCHA that grabs your user’s data and is a blackbox to you”. There are multiple solutions to this: - find a CAPTCHA solution you trust and host it yourself - build your own If you say you trust google, you should write: We value your security and privacy, but seriously we have no idea what parts of our page really do
- jimmy1 8y agoIf you can think of it, someone can make a computer automate it.
- TeMPOraL 8y agoThey can, but unless you're a big and juicy target, they probably won't - because why bother?
- jakelazaroff 8y ago"To prove you're human, read this source code and determine whether the program will halt" :)
- atoav 8y agoSure. On the other hand, I run a blog with a simple to answer question for multiple years now and I didn’t receive a single Spam comment that wasn’t of obvious human origin. Once that stops to work I can bump it up a notch.