To NASA's credit there is nothing in the cited study that mentions anything about the value of the asteroid's metals, about mining at all, or even about this being a discovery. It's just an article about the asteroid's composition. It doesn't take any involvement from NASA at all for a "news" site to realize that they can turn valid interesting scientific observations into a sensationalist click bait headline and make some ad revenue.

NASA did some real science. Some parasitic news writers did some bad math and turned it into ad revenue.

Er... 10,000 x World GDP = $880,000 Trillion that is.

"$10,000 quadrillion ($10,000,000,000,000,000,000), or about 10,000 times the global economy as of 2019"

In additional to all the valid points about a) it's too expensive to bring to earth and b) even if you could bring it to earth cheaply, the resource price would plummet due to massive supply increase. The math in this sentence is also just wrong.

The world GDP in 2019 was around $88 Trillion. https://data.worldbank.org/ind...
$10 Quintillion = $10,000 Quadrillion = $10,000,000 Trillion.
10,000 x World GDP = $800,000 Trillion

"they were only able to subtract 0.2 billionth of a trillionth of a trillionth of a percent"

So why exactly is such a miniscule improvement actually interesting? They handwave over the very small improvement by saying "computer scientists hope this breakthrough will inspire rapid further progress" and "Yet this minuscule improvement breaks through both a theoretical logjam and a psychological one. Researchers hope that it will open the floodgates to further improvements."

But why do they hope that? It's cool that there is ongoing work on different approaches but it's hard to understand why such a small improvement is really cause for excitement/optimism. Is this just Computer Science Clickbait?

Digging through references it's still not at all clear to me what the boundaries of the sandbox are. Is it a technical space like a special network walled off from the open internet only accessible to approved participants? Or is it some kind of abstract "legal space" in which companies can choose to operate so that a different set of more flexible specialized regulators are assigned to them and liability is limited? Will the crypto activities that happen in the "sandbox" be real actual financial transactions or simulated ones?

Pan-European blockchain regulatory sandbox
A sandbox is a facility that brings together regulators, companies, and tech experts to test innovative solutions and identify obstacles that arise in deploying them. The European Blockchain Partnership is planning a pan-European regulatory sandbox in cooperation with the European Commission for use cases in the EBSI and outside of EBSI, including for data portability, B2B data spaces, smart contracts, and digital identity (Self-Sovereign Identity) in the health, environment, mobility, energy and other key sectors. The sandbox is expected to become operational in 2021/22.
https://ec.europa.eu/digital-s...

The Commission is also proposing today a pilot regime for market infrastructures that wish to try to trade and settle transactions in financial instruments in crypto-asset form. The pilot regime represents a so-called ‘sandbox' approach – or controlled environment – which allows temporary derogations from existing rules so that regulators can gain experience on the use of distributed ledger technology in market infrastructures, while ensuring that they can deal with risks to investor protection, market integrity and financial stability. The intention is to allow companies to test and learn more about how existing rules fare in practice.
https://ec.europa.eu/commissio...

Keep in mind that human eyes have a MUCH broader dynamic range than camera sensors and humans are very good at detecting movement. A human who was paying attention would have had a much higher chance of seeing the person much earlier than what you can see in the video. You can't use the video as an example of what a human would have seen.

This is exactly what Google Wave intended to be - a federated extensible decentralized instant messaging service with enough features to replace the features of every other existing messaging service. Their server/protocol implementation was deeply genius (built on top of XMPP) but unfortunately they were cut off way before the client was particularly practical and before federation could be realized. Had Google given Wave a decade to evolve I think it could have solved this problem - but there was no profit in solving this problem and early adopters struggled to figure out how the very rough (but innovative) client fit into their lives.

No company will ever build something as decentralized as SMTP or IRC because the business model of the web involves owning the eyes of a set of users. Allowing those users to jump ship to another service provider using the same protocol doesn't help the bottom line of any startup. On the other hand - open source efforts don't seem to be able to get enough developer attention, and I assume it's because developers know that not enough people will use these alternatives if the big players like Facebook/Twitter/Google with the massive user bases aren't going to cooperate. Wave and diaspora* were both innovative attempts to do decentralized versions of currently centralized communication tools but neither ever had sufficient traction among developers and early adopters.

I was born in 1976. I had logo in I think 4th grade which was the first time I'd ever touched a computer. That summer my grandfather got a TRS-80 color computer and I borrowed/kept my 4th grade math text book. That text book had a BASIC tutorial in the appendices which we had not looked at during class. I used that BASIC tutorial along with a manual that came with the TRS-80 along with some source code print outs from programs that my grandfather wrote (simple programs to help him choose lottery numbers) to learn to write some simple programs. One of my first programs told random jokes. That summer since I had limited access to my grandfather's computer I spent a lot of time writing BASIC programs on napkins at restaurants. A year or two later he gave me that computer and I started making more complex programs. My first ambitious program generated Basic D&D characters, rolling the stats and then suggesting an optimal race and class to go along with the stats. I continued to make that D&D program more and more complex throughout Junior High, adding AD&D rules and adding inventory and spell book parts.

I didn't learn structured programming (with procedures and data structures) until high school when I took a Pascal class at my high school (a math and science boarding school). I taught myself C from the K&R book right after I finished Pascal. I taught myself C++ while taking a C++ class in which the teacher didn't understand the language. The class used C++ in 21 days, but I used Stroustrup. Then I taught myself Perl from the Camel book in high school as well. Whenever I learned a new language I would first work on understanding and modifying example programs and then choose a program I wanted to write as how I would learn. I made programs to build mazes, a web app that was an early kind of forum/blog, and more and more complex AD&D character generators.

I also spent a lot of time in 7th and 8th writing programs for my TI-81 to implement various math formulas we were learning such as a program to do the quadratic formula and to compute other things about parabolas like the location of the focal point and a program to use determinates of matrices to solve systems of equations. The more math tools I learned the more fodder I had for writing programs to implement them.

I found in high school that one of the major obstacles to programming courses was that many kids understood how compiling and linking worked and many didn't. The kids who didn't had a really hard time even getting Hello World to work because the teachers taught the language but didn't teach the tools so the build process was just kind of like a magic incantation rather than a process they understood.

When I sit, my back and neck hurt because I slouch when I'm concentrating. When I stand, my back and legs hurt because I slouch when I'm concentrating.

On my treadmill desk, I never slouch, it's impossible to slouch while walking but it doesn't hurt concentration. So that's the ideal setting for me.

Instead of a sit-stand, I have an HDMI splitter and a wireless keyboard. Monitor at a sitting desk, monitor at the treadmill desk, they show the same thing, just move between them if I have to sit but I haven't used the sitting desk in months.

Scientific consensus means that the thing has been sufficiently studied and reproduced that the confidence is extremely high. This isn't just about "soft" sciences. This is true even in high energy physics. You gather some data and there's a statistical chance that it was all due to noise in the measurements or coincidence. Other people gather some more data and the chance that the conclusion is incorrect goes down. Lots of people gather more data, and one of them finds a counter example, but then more people gather more data and that counter example fits with the expected error bar. This is the consensus process. It isn't about feelings or opinions or subjective truths. It's about increasing confidence and reducing error to the point where the entire community of researchers is confident the findings are reliable and can be assumed true.

Scientific consensus isn't the same as truth. It's just the best proxy for truth we can have. Scientific consensus about Newton's Laws was wrong - but it was only wrong at then-unmeasurable scales and precision. The consensus was incredibly useful, even though it was slightly wrong, because the conclusions it gave were widely reproducible and produced predictions with very high confidence that other researchers and engineers could rely on.

Scientific consensus about climate change isn't "consensus" because some scientists "convinced" other scientists or because it's too hard to do repeatable experiments. It is consensus because repeated experiments and measurements and analyses have consistently increased the confidence and reduced the noise in the predictions.

Nothing is ever proven true. Things can be proven false. And things can be proven to be more and more and more likely to be true and less and less likely to be false (because we repeatedly fail to prove them false). At some point things are proven to be SO likely to be true that there is consensus that we might as well treat them as true until someone comes up with a paradigm shift (ala Newton -> Einstein).

The quote in this article assumes that there's never an error bar on scientific measurements. There always is.

http://xkcd.com/882/

I got a double-degree in Computer Science and Linguistics.

The great thing about doubling in Linguistics is that it is so interdisciplinary that you can use Linguistics courses for most of your general education requirements:

Behavioral Science = Psycholinguistics.
Social Science = Sociolinguistics.
History = Historical Linguistics.
Composition II = Syntax
Philosophy = Semantics
Elective Supporting Coursework for CS = Computational Linguistics, Cognitive Science Seminar, etc.
etc...

Then by the time you've finished your Gen Ed for CS, you've practically got your Linguistics degree.

And everything you learn in Linguistics is essentially about data structures and algorithms and rules and parsing and formal systems and symbol manipulation. The more advanced stuff gets into AI and natural language processing. It'll help your CS brain a lot if you learn Linguistics.

"Very little feedback"?

That depends on the school I suppose. I think that might be part of what makes a good school good is the amount of feedback you're able to get from your instructors. I got excellent feedback from my liberal arts/gen ed instructors.

I've upgraded my 3G to 4.2 from 4.1.1 and things are a bit better than they were before the upgrade but still not as good as before 4.x.

I originally opted for dealing with iPhones quirk (like no multitasking and the walled garden for apps) because I believed the party line that these restrictions helped keep the phone snappy, stable, usable. And the proof was right there: the phone was snappy, stable, and usable. It was a tactile joy to use iPhone 3.x.

Now my phone is so slow that when I push a button on screen it might be 5 seconds before the button changes color to indicate that I've pushed it, and another 5 seconds before the effect actually occurs. Apps crash all the time now. Hard resetting doesn't help. Turning off the search features doesn't help.

Apple's "upgrade" turned my perfectly useful phone, a phone that was a joy to use, into a nightmare of instability and UI freezes.

Upgrading to 4.2 has shown some improvement but not enough to stop me from jumping ship to the Droid X. I'm sure as hell not rewarding Apple by buying an iPhone 4.

Why not take a different approach. Assume that they'll find a way to bring in networked devices. Design a test for which that doesn't help them.

1) Include subtle variations on everyone's test so they can't just share answers directly. Everyone gets a unique test, but the problems are sufficiently similar that you can generate the answer keys for each algorithmically.
2) Make the test sufficiently complex that the only way to finish it on time is by working the problems directly, trying to communicate the problems out to an outside party, letting the outside party solve the problem, and waiting for the answer to come back takes a lot longer than just directly tackling the problem.
3) Have yourself (and a few grad students if available) walk around and pay attention to what students are doing. It should be fairly obvious if someone is using their iPad to chat with a friend rather than to access wikipedia or do some calculations. Just the manner in which the student is typing should indicate whether they're communicating with someone (lots of typing, short pauses to read) or looking something up (very little typing (queries), lots of reading).
4) Include graphical aspects to the problems that are harder communicate via text to an outside party.
5) Make problems where the student has to show their work, not just give an answer. The more words they have to write, the more obvious it'll be to detect patterns of cheating (lots of people with the same words).

Devices with cameras in them could pose a problem (all the "it takes time to type and typing is obvious" stuff goes out the window). But again, monitors walking around should have a pretty easy time noticing students positioning their devices to take pictures of the problems. Nothing beats monitors.

A good effective manager is going to select for talent which is not something that can be faked. The personality traits that are important to a team (and there are several types of person who are going to help the team, not just one) are going to the kinds of things that are innate to you, they're the way your brain works. Maybe for a certain job it's detail orientation, maybe for another it's extroversion...you can't fake or learn those things really...they're just part of who you are. They're the brightest pathways in your brain, well worn with repeated use, formed by your early experiences, that make you very efficient at a certain kind and style of thinking.

Don't try to figure out who you need to be to do this job. Instead figure out who you are and figure out the best job to exploit your tacit innate talents/personality traits. Luckily, on a software team there are roles for many different personality types (tester, coder, analyst, designer, project manager, tech writer, coach, etc).