jogundas

This is about a robot that moves around in a curved space by merely changing its shape. Very cool.

What a cliffhanger -- what did you end up doing? Not sweeping the streets presumably?

A nice example of overfitting!

However, it is hard to imagine an actual application of the process. If I understand it correctly, the author suggests using a set of micro-models for annotating a dataset which is then used to train another model. The latter model can actually detect Batman in a general environment, ie, can generalize. However, enriching a training dataset by adding adjacent frames depicting Batman from the same movie will likely have limited usefulness when training an actual Batman detection (non-micro!) model. Or do I get the final application wrong?

TL;DR mutations happen.

The key paragraph is the following: "We are seeing some mutation coming up in some samples that could possibly evade immune responses," said Shahid Jameel, chair of the scientific advisory group of INSACOG and a top Indian virologist. He did not say if the mutations have been seen in the Indian variant or any other strain.

Both @makomk and @enkid seem to be right [0].

[0] https://www.cochrane.org/CD001269/ARI_vaccines-prevent-influ...

Your point [1] argues that the USSR collapsed due to internal reasons only. I believe that it is an oversimplification.

Cooperation with totalitarian states often prolong their existence. An example that comes to mind is the US subsidizing grain being sold to the USSR and thus preventing starvation, which would have arguably led to a collapse of the communist regime [a].

However, it is also possible to encourage the "feedback loop" that you mention by merely demonstrating the alternatives that are out there. Radio from the other side of the Iron Curtain [b] gave hope to many people in the USSR.

To sum up, thoughtful action from the outside can help bring down totalitarian regimes faster.

[a] https://en.wikipedia.org/wiki/Great_Grain_Robbery [b] https://en.wikipedia.org/wiki/Voice_of_America#Cold_War

Sounds very interesting! Could you direct the interested to your favorite review article or another starting point?

Thanks for sharing! Is there a reason why borg is excluded from the comparison?

You might want to be aware that Yandex is a Russian company before moving all your photos there.

This discovery is definitively cool, but not close to the phosphine-on-Venus cool. The present paper is similar to the former finding in the sense that what looked like a natural explanation (a p-wave order parameter) seems to have been ruled out (just like chemical origin of phosphine seems to have been ruled out). Hence, more exotic findings might lurk in this material.

Wiki says that "Hydroxychloroquine is being studied to prevent and treat coronavirus disease 2019 (COVID‑19), but all clinical trials conducted during 2020 found it is ineffective and may cause dangerous side effects."

https://en.wikipedia.org/wiki/Hydroxychloroquine

Is there something better?

Your comment reads like a quote from Rick&Morty!

The recent JAMA only says that Vitamin D does not help with depression. On the other hand, it seems to be useful in preventing respiratory infections, see the following for a review of randomized studies https://doi.org/10.1136/bmj.i6583 .

Tried it out on a face in Google Photos. Google Photos IDed the "fawked" face without a problem.

My favorite longshot in that direction is lifting carefully-shaped slobs of stone [0] up and down.

[0] https://omegataupodcast.net/299-gravity-storage/

I fully agree that treatment should be prescribed by a human doctor who can explain and answer questions. However, I would not agree that each of the data inputs to the treatment decision should be generated manually. That is already not the case.

Excellent point. That is what the lengthy prospective study phase as well as periodic auditing after deployment are for.

Please also note that there are several important differences as compared to the automotive industry. First, one could argue that the task at hand is trivial as compared to the self driving car. We are operating in a heavily constrained setting with much better understood data inputs and a hundred-year history of medical professionals trying to classify and systematize them. Moreover, our task is not time-critical. It sometimes takes more than a week for such an image to be reported on.

Full disclosure: I am a cofounder at a startup automating chest X-ray reporting.

It is true that ML algorithms are almost always trained on radiologist labels on the same modality, and thus take in the reader biases. I also agree that some radiologists are better than others as you imply.

As a patient, one does not know who will read their film. IMHO we as an industry should aim not at beating 99.999% of radiologists. We should merely make products which consistently perform not worse than an average radiologist at a particular institution. It is always thrilling to outperform humans with your software, but at the end patient outcomes are what matters. Those are about consistent performance over a long period of time.

Demonstrating this consistent performance is the challenging part, but it is possible to prove it through sufficiently careful and lengthy prospective trials. That’s what we are focusing on, and I would love to see the other players in the industry do the same.

From what I understand, NIST has been running a live "competition" on faces of 9 million people for some time already [0]. They are trying to evaluate racial biases in particular [1]. Was there a reason why you decided to not rely on what NIST was putting out?

[0] https://www.nist.gov/programs-projects/face-recognition-vend... [1] https://www.nist.gov/news-events/news/2019/12/nist-study-eva...

For these B2B applications, it is important to have your potential customers in your network. One can argue that achieving such a network is easier in the US, where most of the potential customers presumably are.

It seems that the mosquito effort was not a success after all [0].

[0] https://m.dw.com/en/genetically-modified-mosquitoes-breed-in...

It seems that the market is big enough to allow unicorns making "plugins" like https://en.m.wikipedia.org/wiki/Robotic_process_automation to emerge.

On the subject of Mongols, I highly recommend the popular lectures by Dan Carlin https://www.dancarlin.com/product/hardcore-history-wrath-of-... . Very entertaining, but also historically accurate and balanced.

It seems that a court has ordered Amazon to deliver only essential products [0], and Amazon's reaction is to suspend all deliveries.

They write in their other tweets [1] that the decision is based on (i) the fine of €1M per infraction imposed by the court, and (ii) the risk of contravening the decision being high due to complex logistics.

[0] https://www.bbc.com/news/world-europe-52285301 [1] https://twitter.com/AmazonNewsFR/status/1250481151313154048

Since the other commentor has already tried to address the benefits, I would like to point out that the costs are far from infinite: eg, NASA budget is 0.5% of total US budget (not GDP!). That is arguably a good investment, even if the only return is increased motivation to study STEM.

The article quotes him as saying "I did Google reverse image search for their profile photos". I am in Europe and Google reverse image search for faces of people certainly does not work. Is that a recent change? Is that local to Europe?

Is that a SOCKS proxy, or just a HTTP proxy? The github readme does not make that clear.

Surprised to see so few innovations for the sweaty back problem. Are the "trampoline" and plastic with holes really the only solutions in this space?

The most extensive writeup with references that I have seen so far is at https://fastlifehacks.com/n95-vs-ffp/ . Not sure if this is an original summary, or ripped off from some other site though.