kkm

This is very interesting, planning to write about it?

Also the vllm patch accompanying the blogpost: https://github.com/doublewordai/vllm-amd-blog-doubleword

I can also highly recommend reading: Touching Lives: The Little Known Triumphs of the Indian Space Programme, by SK Das (2007).

Excerpt: Touching Lives is not merely a chronicle of the community outreach of the Indian Space Research Organisation (ISRO). It is the story of journeys to far corners of India meeting people whose lives have been transformed by technology.

More books about ISRO: https://forum.nasaspaceflight.com/index.php?topic=33814.0

I can recommend material from Donella H. Meadows. The book “Thinking in Systems” is a very good one.

https://www.goodreads.com/book/show/3828902-thinking-in-syst...

Found a link on Slideshare, if someone does not want to download the PPT. https://www.slideshare.net/kfrdbs/peyton-jones

Big Big fan of mitmproxy. This is a wonderful tool, and with scripting functionality it helps achieve so much more. I have been introducing this tool to so many developers and quality assessment teams, to not just improve day to day things but also perform privacy assessments w.r.t GDPR etc.

Thank you so much for your work. What is the best way to donate to your project?

The price difference are huge, even when you compare from Germany, however, it’s important to note that iMessage and FaceTime are most likely disabled on iPhones bought from UAE. At least that was the case back in the time.

https://www.reddit.com/r/apple/comments/943iql/i_think_the_u...

I think so, but the data collection practices are governed by individual apps.

I have been testing Edge on iOS for Privacy practices. I think there are lot of things to improve. Also, it would be great for Edge to make it easier way to report and have discussions. Currently it’s a black box.

- https://twitter.com/konarkmodi/status/1258163915319640071

- https://twitter.com/konarkmodi/status/1258185278168223746

- https://twitter.com/konarkmodi/status/1262019416914644994

- https://twitter.com/konarkmodi/status/1258338835722887171

Thanks, for the article. Always good to refresh the common pitfalls.

There is a RFC which also details the best practices for JWT: - JSON Web Token Best Current Practices: https://tools.ietf.org/html/rfc8725

On the similar topic, some more interesting RFCs / Drafts from IETF on OAuth: - OAuth 2.0 Threat Model and Security Considerations https://tools.ietf.org/html/rfc6819

- OAuth 2.0 for Browser-Based Apps - https://tools.ietf.org/html/draft-ietf-oauth-browser-based-a...

Disclaimer: I work for Cliqz.

Good to hear that Mozilla changed their mind, when we proposed an open API for extensions to access the search area it was turned down

Ref: -https://bugzilla.mozilla.org/show_bug.cgi?id=1361327

https://bugzilla.mozilla.org/show_bug.cgi?id=1361327#c28

Thank you for taking the time to read the article and sharing your feedback.

Noise is added for what is called plausible deniability. Note that messages themselves do not contain any user-identifier. We take extra measures to strip request headers not needed by the server to avoid extra information that could be used for implicitly linking messages[1]. Possibility of linking messages based on network fingerprinting (ex: IP - which we do not log) still exists and is an open concern which we will solve in the next version. This at most makes it possible for us to learn that these 3 domains are visited by the same person, again - given that the list of domains are from shortlisted top-news domains, it is safe to assume that they do not contain any PII.

That said, it is not the strongest model we apply -- due to resource-constrains we have not updated to strongest models like we do on more sensitive data - via HumanWeb[2]. We will soon do the changes on a couple of dimensions: a) each domain as separate message, right now this introduces un-wanted spatial correlations, and b) send the domain through the proxy network HPN[3].

Ref: [1]: https://github.com/cliqz-oss/browser-core/blob/7679c40aec9fe... [2]: https://www.0x65.dev/blog/2019-12-03/human-web-collecting-da... [3]: https://www.0x65.dev/blog/2019-12-04/human-web-proxy-network...

Disclaimer: I work for Cliqz.

There is some similar work in the space of checking if the credentials being used by the user are compromised or not.

Example: https://arxiv.org/pdf/1905.13737.pdf

Hi dessant,

Disclaimer: I work for Cliqz.

Would be really interesting to know your concerns with FoxyProxy.FoxyProxy is legally bound to not log the IP or share it.

From the HPN protocol's perspective, data can be routed via any trusted party - in our case it's FoxyProxy.

Right now, there is no way to configure this in the Browser, but should be doable. It's actually one of the motivations to move to the newer version of HPN[1].

We do agree that sending data through Tor network is the gold standard for anonymity.

- We did a lot of work on getting Tor running in Cliqz browsers. It's a hard problem but definitely do-able, something we might pursue again in future[2]. - We also have experimented with WebAssembly version of Tor client to make it compatible for web extension[3].

Having the ability to use the Tor network in Cliqz products is also good, because we can actually leverage the anonymity guarantees by sending data via .onion services. You can also check more details under evaluation section of the paper[4].

In case you wish to check the network traffic you can also check the debugging section[5].

References:

1. https://www.0x65.dev/blog/2019-12-04/human-web-proxy-network... 2. https://github.com/cliqz-oss/browser-f/commit/12fcef8479d9c3... 3. https://github.com/cliqz-oss/browser-f/commit/12fcef8479d9c3... 4. https://arxiv.org/pdf/1812.07927.pdf 5. https://www.0x65.dev/blog/2019-12-03/human-web-collecting-da...

Hi,

Thank you for bringing this up. Although, this is not relevant in the context of the blog post, we are on that list by mistake. We do NOT collect any personal data in our browser: (more details e.g., here: https://0x65.dev/blog/2019-12-02/is-data-collection-evil.htm... and https://0x65.dev/blog/2019-12-03/human-web-collecting-data-i...) and we go a long way to make sure not even implicit indentifiers go through. We believe we ended up on that list for a bad Firefox experiment and we will reach out to the maintainers, make our case.

Disclaimer: I work for Cliqz.

First, let's check if you can open another domain on .dev TLD, like web.dev, if not then:

Seems like you have some mapping for .dev TLD. Assuming based on your mention of Safari, that you are using Mac.

Could you check if you have some setting in your /etc/resolver for dev TLD, or if you are using some service like dnsmasq which is trying to resolve .dev to a non-existent location.

Hi,

Interesting, could you tell what's the error to see.

Other ways you can reach the blog: If you use Tor browser can you try opening: http://cliqzdevxo33b4h6.onion/

Or if you use Beaker browser: dat://ee172d7cd9235b2cf86ea9481e8a40e48cea29c743036621edc79a4765aa0281

Disclaimer: I work for Cliqz.

Hi soumyadeb,

It is true that there needs to be a way to store the aggregate queries "per user". In the current approach of collection this place happens to be on Server, but aggregation per user can easily be done on client-side by leveraging Browser storage.

Approaches like k-anonymity or L-diversity are good but they tackle the problem from a different perspective - making sensitive data available for querying without revealing actual content. The approach suggested in this article talks about methodology which removes the need to collect such data in the first place.

You can also check our paper: https://www.0x65.dev/pages/dissemination-cliqz.html#GreenAna...

We will talk in detail about this methodology- Human Web and how we use to collect data for our search, without compromising users privacy and based on client-side aggregation.

Disclaimer: I work with Cliqz.

Hi, Disclaimer: I work for Cliqz.

Thank you the questions, we are always looking for constructive feedback on and off HN.

1. These random values are used for grouping partial queries together, and they reset when you press enter or start a new query. Source code on how it's generated: https://github.com/cliqz-oss/browser-core/blob/master/module... We actually take one additional precaution of using crypto random and not plain Math.random(), which could potentially be used to link multiple sessions together.

2. There is no feature to disable auto-suggest. But I will pass your feedback to the team.

3. No there is no shared ownership, we don't have access to their servers. We also do additional encryption with bucketing on the payload sizes that we route through foxyProxy, so that the proxy provider cannot learn anything about the content of the message. We will have a blogpost explaining this on Wednesday - 4th December. Also, we are looking to add an option, where user can choose their own proxy provider too.

4. There are two parts: a. You can select the option from Control Center (Q menu) icon in the toolbar -> search -> search via proxy. (Now the calls should go through FoxyProxy) b. All calls to api.cliqz.com go through proxy when in private mode. The only reason it's not default is latency. As to what goes through FoxyProxy by default is: all Human Web data.

Once again, we appreciate you looking into details, and please keep digging, we would be happy to answer, improve our documentation and if there are bugs specially related to privacy and security they are on our uttermost priority.

Hi Aorth,

Disclaimer: I work for Cliqz.

We are aware of this and it's unfortunate, although the lists are not the main ones from community like Easylist, uBlock Origin etc. Cliqz does need to collect data to build it's search engine and other features. Even though we do with by privacy-by-design in mind and without collecting any PII (we will be sharing more details on the process on data collection in the blogposts scheduled for Monday, Tuesday and Wednesday),still we do end up getting on such lists.

Of course, you don't need to trust what we say, all the code used for collecting data is open-source for transparency and auditing. Additionally, also happy to help setup network debugging incase you want check what exactly goes out. - https://github.com/cliqz-oss/browser-core

We are also willing to discuss with the maintainers of the list and explaining what and how data collection is done.

Hi Marcus,

Thank you for your questions.

Disclaimer: I also work on WTM from time to time, so I hope can answer these questions to some degree :)

1. WhoTracksMe is served via CloudFront, it only get's the IP that CF reports. To get the country, there are two ways that can be used. IP => Country mapping via a Database or additionally, because CF has multiple edge locations, if needed the you could see which edge location served the request, that will help you narrow down the location at a country level. Ofcourse, in both cases if the user is using VPN, you would only records the VPN location.

2. The problem we want to avoid is same IP with same hashing algorithm will produce same value. Which can then be used to co-relate user activity across days. Second, depending on what hashing scheme one is using, rainbow tables can be used to get back the original value. Therefore, to avoid, we used the approach of daily key. Now that I write, we could also use some hash + daily_salt. This should give the same guarantees. - Needs to be checked.

3. In this setup we are not using Cloudwatch but Cloudfront, which is the CDN provider from AWS.

In particular case of liberation.fr, anyone who has access to the value of ‘djazsession’ cookie can log in to the users’ account. This is one of the cookies being sent to Eulerian.

Here is a demo video:

https://twitter.com/konarkmodi/status/1198412297842184192?s=...

Websites need to really up their game, specially given the amount of third-parties they are using.

I've tried highlighting similar issues in the past, where even if there is no active breach, but they are leaking sensitive data to multiple third-parties when it's not needed in the first place.

https://dev.to/konarkmodi/watching-them-watching-us-how-webs...

https://news.ycombinator.com/item?id=16516687

Lot of companies are shooting themselves in their own foot by sharing critical data with a plethora of third-parties.

They put sensitive information like username, orderid in the URL which is then shared with all the third-parties on that page, simply because referrers are not sanitized.

This happens:

- Without user-consent

- More dangerously without the companies knowing it too.

On reporting, the companies do not want to fix these issues.

Shameless plug: You can find some of such cases, which I've been trying to highlight to the companies:

- https://medium.freecodecamp.org/how-airlines-dont-care-about...

- https://threatpost.com/def-con-2018-telltale-urls-leak-pii-t...

- https://cliqz.com/en/magazine/lufthansa-data-leak-what-a-sin...

- https://fosdem.org/2019/schedule/event/web_extensions_exposi...

Thank you for the work <3

HN thread about the Adblockers performance study mentioned in the article: https://news.ycombinator.com/item?id=19175003

Ghostery is open-source for almost an year now, You can check the code and repo: https://github.com/ghostery/ghostery-extension

Nice, is this the YouTube Red subscription?

Yes, I think there are certain domains specially spyware, malware that need to be blocked altogether and DNS blocking is the most optimal way for that, specially considering the wide adoption of IoT, where extensions cannot run.

afaik, in Chrome Web Assembly works fine when used inside Workers.