Cynddl

This is just a dataset of postcodes, no demographic information about the population alas.

The article is available here, in open access: https://www.nature.com/articles/s41467-019-10933-3

Co-author here. We designed a statistical model, which is never 100% sure a re-identification is correct. There is, e.g., a non-null probability that two individuals in the US share 5, 10, or even 15 demographics attribute.

Co-author here, we will add the source code very soon. A little swamped with the press coverage, but the source code in Julia+Python is coming.

Yep, we used a dataset by the ONS, which includes only England and Wales. This is indeed not clear on the website, we'll fix that.

Wups, what is your postcode? Feel free to let me know at luc (at) rocher (dot) lc if you don't want to post it here.

We used the complete list of US/UK postcodes (edit: England and Wales only for UK) to implement the demo, so all of them should be there.

Not sure why you were downvoted. Our infrastructure worldwide relies a lot on location data, incl. GPS data.

There is actually an article from the New Yorker answering your question: much less traffic, all planes grounded, no train, etc. [1]

> The U.S. Department of Homeland Security classifies sixteen infrastructure sectors—including dams, agriculture, health care, emergency services, and information technology—as critical, and therefore particularly vulnerable to sabotage. All but three require G.P.S. for essential functions.

[1] https://www.newyorker.com/tech/annals-of-technology/what-wou...

> Audio snippets are not associated with user accounts as part of the review process.

This is a bold claim. The VRT journalist were able to re-identify few data subjects and confirmed with them that their data was being listened by Google contractors. "‘This is undeniably my own voice’, says one man, clearly surprised." wrote the journalist.

> The Google Assistant only sends audio to Google after your device detects that you’re interacting with the Assistant.

Well, this is not what the contractor said [1]. The contractor heard private conversations, sex scenes, and violence including "a woman who was in definite distress".

https://www.vrt.be/vrtnws/en/2019/07/10/google-employees-are...

I'm confused. Can you justify your claim? What is wrong with the French economy and how is it related to worker protection? France is, e.g., behind Germany in term of employer protection according to the OECD [1].

[1] https://stats.oecd.org/Index.aspx?DataSetCode=EPL_R

In France, such contracts do not imply that a company cannot fire employees. Contracts can be terminated for economical reason, which however needs to be justified.

If your company is working well, should you be able to downsize and fire employees without a clear motive? To a certain degree, France believes the answer is no.

This algorithm actually aims not to detect photoshopped faces, but faces warped by the Face-aware Liquify tool. I wonder how it compares to state-of-the-art commercial forensic solutions (see [1]), and more generally how DL algorithms compare with such closed source forensic tools.

[1] http://www.bbc.com/future/story/20170629-the-hidden-signs-th...

And 2 months ago: https://news.ycombinator.com/item?id=19590567

Thanks for the detailed post! I'm usually wary of using close-source programs for sensitive tasks such as connecting to all my remote servers with SSH.

I haven't found any details on the libraries you use, especially for cryptography, nor the steps you have taken to secure your software. Where can we find more info?

The title should be: jTools - The definitive common javascript web components.

I'm confused, did they legally entered the Chernobyl Exclusion Zone? I was told that biking into the Chernobyl Zone, e.g., is forbidden due to the risk of contamination when you exit.

Edit: Indeed it was not. From Instagram, “we spent the night tiptoeing around razor wire fences, coasting through sleeping security checkpoints, and riding frantically away from some surprisingly alert and vigilant guard dogs.” This validates what I know from Ukraine. Entering the exclusion zone is a very lucrative touristic business, with prices around $100-200 for a single day trip in a group.

OCaml can have nearly the same performance as C.

Naive question: why is it important, compared to the developed features and new architecture?

Better article (despite taking 45s to load!): https://www.redhat.com/en/about/brand/new-brand

> We formed a task force to redesign it, bringing together people from different disciplines at Red Hat: graphic design, UX design, creative strategy, video, brand, product marketing, and more. And we worked with Paula Scher, a partner at design firm Pentagram, to help us focus. […] After 5 months of research, explorations, and brainstorming, the team found an elegant solution that still reflected Red Hat and could grow with us.

More interesting questions:

1. What does not work in WP's model that you will address? Moderation? Transparency? Funding?

2. How do you plan to have 1000x the number of articles on WP? How do you plan to moderate and follow the accuracy of these 6,000,000,000 articles?

3. Is it an English-only project or do you value international contributions?

> The arbitrary threshold of what is notable and what is not doesn't cut it in the Knowledge Age. There are currently 5.8 million English language articles in Wikipedia, and Google had 1 billion objects (200x Wikipedia's size) in its Knowledge Graph when it launched in 2015. We estimate internally that there are 1000x the entities to cover than what Wikipedia has today. It’s an exciting challenge!

This is a very weird comparison. Why not helping Wikipedia and contributing to this huge open knowledge base? The rules set by the WP community for acceptability clearly do not imply that “actual technologies, projects, products, theoretical electrical components and academic ideas” might be removed if written there, as the author suggests.

That articles, or even drafts, for specific American VC-funded projects did not reach WP's acceptability threshold does not mean that there is something fundamentally wrong in Wikimedia's approach. I'd love to see a better comparison.

> The etc. could include hyper-local ads, making the system self funding.

Yikes. Why do public goods always need to have ads to be self-sufficient?

> This can be far cheaper than tactile pavement and so more widely deployed […].

The cost of deploying low power transmitters, developing and certifying new systems (not everyone carry a smartphone) or mobile apps to receive the signal, educating impaired people across the world to use them, then regularly maintaining the electronics across a city will hardly be cheaper than passive tactile paving.

> This questionnaire included a series of questions about whether respondents had a specific psychiatric condition, including depression or anxiety (categorical response options: yes, no, unsure), and whether their conditions had changed (categorical response options: better, stayed the same, worsened) after 5-MeO-DMT use.

This is a self-reported study. Asking whether a substance or drug improved someone condition will often result in positive answers, whether or not it was therapeutically useful.

In the sample, more than 90% of individuals had already used 5-MeO-DMT, 41% had depression, 48% with anxiety. 79% reported that “their anxiety was better following 5-MeO-DMT use”. This is indeed not a balanced sample, and self-reports provide very weak results compared to a proper double-blind study. While doing a double-blind study on such substances might be challenging, there are definitely tools and environments that can help correctly measure the true effect of a substance on depression on anxiety.

I don't see how this study proves anything.

The original title is "Fast-Acting Depression Drug, Newly Approved, Could Help Millions". Not everyone reading HN lives in the US nor knows what the FDA stands for.

The article reports the experiment, the results, and not only FDA approval.

From Jon Kleinberg also, I highly watching his recommend his presentation [1] on the inherent trade-offs when trying to achieve algorithmic fairness.

It's a little more technical than this paper, and presents very well why removing discrimination in algorithmic decision making is a complex task.

[1] https://www.youtube.com/watch?v=4X3Z7FPwkA8

> Due to the non-random sample, as well as the lack of placebo control, it could be argued that sub-acute enhancements are due to uncontrolled factors such as psychological expectations, or the environment in which the drug is taken.

Indeed. The setting of the retreat (“After ingestion, participants were instructed to stay on the premises, and were able to do what they wanted, as long as they did not disturb other participants. Facilitators provided music, tools to draw and/or write, and food. In the evening, all participants and facilitators came back together as a group.”) might have had a significant effect on the results. These should be, at least, compared to a similar retreat where no psilocybin is ingested.

Otherwise, the effects could come solely from the environment, e.g., a calm and relaxing retreat far from a busy life.

> The Raspberry Pi is a little underpowered for serious computation, so I need to offload processing to my local Kafka cluster.

Couldn't this be done on the Raspberry Pi with a O(n) search in the OpenFlights data? Probably less sexy, but a simple grep with location and time range should be enough.

This website does collect personal data. Google's FAQ on GA simply states that the first party should obtain consent before transferring data to a third party (and transfer of consent might not be GDPR compliant, but that's another issue).

Here, the first party (analytics-bypassing-adblockers.netlify.com) has to obtain consent before collecting personal data. And IP addresses are not the only personal data that GA can collect.

> Hello from Google Tag Manager. This text is being added by a tag running from GTM.

One should note that this inclusion, without an opt-in consent banner for instance, is not GDPR compliant. The URL https://analytics-bypassing-adblockers.netlify.com/proxy/htt.... sends personal data to a third party (Google) without my explicit consent. See Article 7 and Recital 32 of the GDPR:

> Consent should be given by a clear affirmative act establishing a freely given, specific, informed and unambiguous indication of the data subject’s agreement to the processing of personal data relating to him or her, such as by a written statement, including by electronic means, or an oral statement.

> this isn't cancer research that informs treatment regimes nor social science that informs public policy. it's code and theorems

This is a dangerous line. Being a reviewer in CS, I don't think peer review is easy. The work in this field is complex and require a lot of time to correctly evaluate the claims.

First, CS/maths is not just code and theorems. There is indeed a lot of research which requires data and empirical validation, or propose new models.

Second, it's not easy to verify claims, even for code or theorem. This is why peer review requires many reviewers, often from 2 to 6, to evaluate one paper for a conference or journal. No one can evaluate in 5min whether a 40 pages paper on ArXiv makes sense or not. When I read a 70 pages complex crypto paper, I prefer a peer reviewed, community-backed work than a PDF not verified by other researchers. When you're a reviewer, you realize that many papers are often butchered, or contain (partially) false statement, not always easy to spot.

I see more and more posts on HN that rely on ArXiv for scientific articles. While many researchers in CS publish their research on preprint servers, these are (a) only one part of the scientific publications out there and (b) articles often not (yet) peer reviewed.

I get that ArXiv is a convenient way to download scientific papers, especially without access to universities or libraries' access, but one should always be careful with non-peer reviewed research.