mW= 1/1000th of a watt.
mW=W = parts per 1000 efficiency.

480mW/W = 0.48W of light out for every watt of electricity in.
This is a deep blue LED.
It is very bad if you measure it in lumens per watt because the eye is quite insensitive to blue light.

Whatever the answer - 30%/44% (and you can't do it that way, you've got to integrate over the spectral response of the eye and see if you actually care about colour - green light at 600lm/W is not a functional white light) - is still vastly higher than 10%.

You cannot get the highest efficiency per watt LED bulbs, simply because they would require more LEDs than are absolutely required, and cost more, for no consumer visible benefit other than the watts.

Err - no.
Look at the URL, there is a clue why you're spectacularly wrong.

Current LEDs (blue ones, which white is based on) exceed 50% quantum efficiency.

http://www.digikey.com/product... - for example - does 48% electricity to light.

http://optics.org/news/6/2/6
http://www.nature.com/nmat/jou...

The writer of the original article should be shot, hung, shot, and then boiled.

It is riddled with so many inaccuracies that it's meaningless.
'10%' - yes - 10% is mentioned ' Our first devices already exhibit an extrinsic quantum efficiency of nearly 10% and the emission can be tuned over a wide range of frequencies by appropriately choosing and combining 2D semiconductors'
But going from that to LED efficiency is ridiculous.

It is comedically ridiculous to claim that it's going to result in products this year.

It's worth noting that the best existing 'warm white' LEDs bulbs can already produce about twice as much light per watt as compact florescent.
(if they are made with around double the normal number of LEDs and a more efficient power supply).

This perfectly answers the question of where to get emergency housing in a place easily accessible by trucks for people who have wifi to
entertain themselves and are in areas secure enough that going to the external toilet is not dangerous.

It's called an autobiography.

Is the mill in question over $50/lb?

Violating the manufacturers instructions runs the risk of damaging equipment - and - for example - having sharp bits of the broken equipment stab the patient.

You are not required to incriminate yourself.
This however does not mean you cannot be compelled to give physical items,or access to physical items (including fingerprints).
The cops have no right to demand you produce your passphrase.
They have a right to demand the bit of paper they know you wrote the passphrase on.

Licence has a qrcode or similar onto the DMV website.
(proper verification apps ensure that the URL is actually the DMV website and ignore any other URL)

Just be aware that that doesn't work everywhere.
You don't get emergency calls in the uk, for example.

Oh yes!

That is not hardware.
The hardware - the FPGA has remained constant.

Software runs on hardware - yes.
Software cannot increase the capabilities of hardware - well - not quite.
The most literal meaning of this - apart from limited things like overclocking is of course broadly true but may be hugely misleading.
If you've got a really advanced program on each of a network of computers, doing a given task - there are many ways in which it can seem to increase its capabilities, without really doing so.

Giving up the designated task and freeing resources.
Co-opting other systems into adding to its resource.
Optimising the way it performs the task so that it at least does it reasonably well, but much cheaper.
Sharing computations over multiple devices which were expected to be done on one.

There are many systems where 'dumb' algorithms are tens, or thousands of times less efficient than optimum ones.
Optimum algorithms are in many cases intractable for humans to find.

Optimising computational efficiency over time as machine learning is a really valuable thing to do.
Looked at from another angle, this can come quite close to 'evolution'.

However, at this stage, it is not required.
Simply as the threat is well over ten years out.
How much over - good question.
Is it too early to raise concerns and encourage people to go into fields where they may think seriously about this topic - no.

If the tax code was rational.
The problems are that multiple levels of tax code interact in complex ways that vary with the exact addresses involved in the claim.
So, you're not writing one codebase which does taxes, but in a very real sense, thousands.