leecarraher

That purpose isn't covert, it is funded by the National Nuclear Security Administration(NNSA) under the Stockpile Stewardship Program, which built the facility to mimic fusion conditions to assure the viable of the US nuclear weapon stockpile. From time to time other scientist get to use it for other fusion experimentation.

what about pumped storage hydroelectricity? It currently accounts for 95%-99% of potential energy storage capacity, is far cheaper and more efficient than current commercially available chemical battery storage methods.

I would like to see more vetting in dockerhub and security in general, yes it's based on cgroups which are specifically security minded, but not all docker users are aware of this. A novice user pulling a random docker image, and running default as root is bound to cause some issues.

Exactly, docker standardizes something that was an otherwise nightmare to maintain across platforms and organizations. Next we need to standardize the orchestrator, my money is on kubernetes, but I wouldn't say it has won over everyone just yet.

People fall into 3 categories is a pretty course classification. There are hot prepared meals, cut prepared meals you still need to cook, microwave meals, raw ingredients and everything in between. Meanwhile stores have been reducing waste. Someone is buying the stuff and it isn't accounted for by just the 10%. My guess is that the past decades economic windfall just means we are eating out more than the previous decade. Eventually economic downturn will make eating out prohibitively expensive for many, and they will rediscover their "love" or more likely somewhere on the spectrum of cooking appreciation.

They seem to not sell used baby gates at second hand baby stores either. id guess for a similar reason and that being that they don't want to claim liability in asserting the item was in original working order.

I think there is a bit of a catch-22 in regard to ridership and using already limited city roadway resources. I am privileged to have a protected bike lane for a about 50% of my commute to work, but can't deny the argument that it is vastly underutilized in comparison to roadways. Unfortunately this seems to further polarize cyclists and non-cyclists when I do need to share the roadway. It is a difficult problem that large scale public works cannot solve alone.

python handles big int automatically:

sh-4.2$ python -c "print(-805387388120759743+804357581458175153+126021232973356313)" 42

my concern is that people become conditioned to look for these lights, and draw the conclusion that light off = camera off. This is a bad precedent, as multiple built-in laptop webcam attacks to disable the led recording lights can attest to. If you need privacy, don't be in front of a camera or microphone. don't trust a clearly software controllable status light.

i bought conductive thread many years ago with the intention of sewing it into clothing for just this purpose, but then reassessed my general threat position, and realized i could probably put it to better use as lighting for a halloween costume.

i once solved a similar problem with set theoretic approach using matroid theory and the greedy method for constructing a matroid basis using a thesaurus as my independence oracle. This was over a decade ago, so I no longer recall the results, but it certainly seems similar in the goal of finding a primal set of words that can in some way define all others.

I recall this as a standard 'canned' interview question. The goal was to assess if the applicant had a general passion for software development as well as a drive to learn new things. While I had no particular issue with the goal of the question at the time, I always felt it a bit one dimensional. At a different job in a more corporate setting, I find the push to be just the opposite. Time-off is time to unwind, use it as such. I believe companies have found success in this cultural imperative as a means of preventing turnover and burnout.

By no means do the authors suggest that this is an anti-surveillance system. Perhaps the methodology could be utilized, as has been shown in a variety of adversarial attacks on DNNs. But the GAN approach is entirely target DNN specific. Any minimal change to the DNN would require an entirely different adversarial solution. More practical approaches have existed for some time to thwart surveillance (https://hackaday.com/2008/08/27/testing-ir-camera-blocking/)

A simple litmus test that saves you from most clickbait, si whether the actual paper title contains, more-or-less, the equation P=NP or P!=NP. I cite Andrew Wile's proof of Fermat's Last Theorem, titled "Modular Forms, Elliptic Curves and Galois Representations" . The title will be about the how, and not some pop article title.

i like the razor edge they put on the bottom of the computer, such that after using it for a while, people feel compelled to ask me about my day and offer kind affirmations.

wasn't owning a horse made quaint by the fact that owning a car was far more advantageous? so what is replacing the car? even the authors own anecdote/ethos plug isn't that they cut the landline cord because they could just share a landline phone with their friends, it was because a personal cell phone was a more advanced replacement. maybe i am looking too deeply into an opinion piece from an author who thinks that their ability to predict the rise of cell phones in 1998 makes them an authority prognosticator. well him, zac morris, my fellow arrl, and anyone whose had the pleasure of cruising the citizen's band.

I've always measured rich people sports by the cost spectrum of equipment available. Football, usually provided by some team or organization or school, soccer, is shoes and a ball, maybe a goal, but not necessarily. Then there is tennis, or cycling, where you can spend about as much as you want to buy some perceivable edge over others. Running is an outlier, but the article does suggest the sheer time it take to train is a barrier for many.

"ok step 1, let's start up Wireshark(or equivalent sniffer)..." JK. Just another I said something "I think" I never said before and now here's an ad for it-test.

they may not want to include intermittent sources of power.

An awesome looking tool, however I'm a little wary of building software for an expensive multi-gpu server that isn't fully vetted. Seems like a ripe target for cryptojacking. On that note, people should probably avoid pre-built binaries of this for that reason.

great write-up, well written and relatively easy to follow.

in light of this article (https://qz.com/455109/entrepreneurs-dont-have-a-special-gene...) featured on hackernews 3 years ago, it would appear that both 20yo and 40yo entrepreneurs where both part of the financial elite prior to founding a startups.

happy to see John Baez fans on here, i like his series on his favorite numbers https://www.youtube.com/watch?v=vzjbRhYjELo

For nn's in my experience out of memory, and preprocessing tends to cause an equal number issues as the nn optimization. Which tfrecords and streaming seem to solve. Are there similar object loading facilities in pytorch? Though I have not specified models in keras, since it is now part of tf i presume the formats are compatible.

I will be curious to see what competitors of Microsoft, who host projects on github (google,facebook, etc) do. Where previously Github was an unbiased host for open source projects, now being under the umbrella of MS, other entities might see it as clashing with their corporate policies.

Hopefully ICA susceptible De-anonymization techniques are no longer HIPAA best practice. Or perhaps this is a study to prove that newer additive and multiplicative techniques, are also susceptible to De-anonymization attacks.

It could be an epidemiological study on aggregate populations for some communicable disease. In all cases the hospital side would be bound by HIPAA to anonymized any data they provided. Google does similar prediction studies based on search, and it is very valuable to the CDC for allocating flu vaccine.

well they did right? Isn't that what they paid cellebrite $900K to do. (http://www.globes.co.il/en/article-fbi-pays-israeli-co-90000...)

"The tests ran on a desk(t)op with 16GB RAM" a single node setup doesn't fit the distributed, fault tolerant, use case that Spark was designed for, rending this comparison moot.

python, and more specifically, keras on tensorflow and sklearn. R has great packages, as does matlab, and many other languages as well. But if you want to do any deep learning, tensorflow is the way to go.