Based on the article it sounds a lot like envelope tracking:

The new advance is essentially a blazingly fast electronic gearbox. It chooses among different voltages that can be sent across the transistor, and selects the one that minimizes power consumption, and it does this as many as 20 million times per second. The company calls the technology asymmetric multilevel outphasing.

But looking back at one of their papers: http://www-mtl.mit.edu/~jldawson/Dawson_digest2009.pdf it is not exactly envelope tracking. They are starting out with an outphasing amplifier (dividing the signal into two constant amplitude variable phase signals which can then be amplified with an efficient nonlinear amp and recombined into the original signal) and then adding a discretely variable drain bias (much like an envelope tracker) but the purpose of the variable drain voltage is actually to minimize power dissipation in the combiner's isolation resistor. Thus they compare their efficiency and linearity only to outphasing types of amplifiers. I would be curious to see how this stacks up to pure envelope tracking amplifiers.

I used a Graph Gear 1000 for a while. It has incredible style points, feels solidly built, and the retracting tip is a nice feature for when you keep the pencil in a bag or pocket. However after around 6 months of regular use with storage in a backpack the springiness in the clip was lost, and due to the unusual tip-retracting mechanism this mean that the tip would not stay retracted. Also I found that I was not a big fan of the narrow and hard metal grip, the heavy weight during long use, and the narrow eraser.

Nowadays my preferred pencil is the Pentel "Quicker Clicker" series, specifically the .7mm model, PD347. The wide body with a thin layer of rubber on the grip (enough to soften without being squishy) is very comfortable for me, and the large and easily accessible eraser is very handy. The non-retracting tip means I don't like to carry it around in my pocket, but since I only use it at my desk(s) and it's much cheaper than the Graph Gear I simply bought enough to cover all of the places where I'll need one.

Rather than simply modding you up I decided to try this out, and it works! Which is kind of creepy.

Since you dismiss the Newcastle University study only on the basis of Mark Bittman being a biased columnist (fair enough, he probably is), you may be interested to read the recent NY Times article which also explains some of the methodology differences between the studies which lead to different conclusions. In any case, I'm reserving judgement on both studies for now.

http://www.nytimes.com/2012/10/16/science/stanford-organic-food-study-and-vagaries-of-meta-analyses.html

True, a pick and place requires a large scale of operation before it becomes worth the capital and the reels. But it shouldn't take much effort to create a solderpaste mask - whenever you're defining a new pad just add the solderpaste to the pad definition and you're done, all you need to do is export the layer at the end. This is helpful even for relatively small board runs - you stencil on paste, place parts manually, and reflow in a lab toaster oven / hot plate, etc. There's very little capital and once you get above a few boards or so (depending on the board and the assembly guy) it can be faster than hand soldering.

I imagine you could even use the inverse of the soldermask layer as a paste layer, at least for non-RF boards where everything except pads is usually covered in solderpaste. It's not ideal because then a bit of misalignment puts some paste on top of the soldermask which risks bridging for fine pitches, but I know there's CAM software out there which could shrink the solderpaste layer to correct for that.

There's a huge difference! What you propose involves the actual loss of billions of dollars in property investment and a major impact on the sport in the US for some years due to loss of venues. What happened on WoW will require some overtime from the admin team to fix, resecure, and restore to pristine state from backups over a fairly short period of time.

So to better respond to the WoW hack: "Who cares, it's just an online game that they can restore from backups before you know it."

You missed some important points when you draw your conclusion:

This very important section of the article (emphasis mine) is conspicuously absent from your post:
Yet even within its narrow framework it appears the Stanford study was incorrect. Last year Kirsten Brandt, a researcher from Newcastle University, published a similar analysis of existing studies and wound up with the opposite result, concluding that organic foods are actually more nutritious. In combing through the Stanford study she’s not only noticed a critical error in properly identifying a class of nutrients, a spelling error indicative of biochemical incompetence (or at least an egregious oversight) that skewed one important result, but also that the researchers curiously excluded evaluating many nutrients that she found to be considerably higher in organic foods.

At this point that given that two research institutions have published metastudies with opposite conclusions, and that errors and oversights have been identified in the Stanford study, I'd have to say that the jury is out on this topic.

No, we haven't for a long time.

Taco Cowboy's link a few posts above addresses this and acknowledges the decreased energy density but made claims that large-scale supercaps would provide better economy in terms of $/(kWh*(number of cycles)). But of course $/kWh is also important because if they are highly capital intensive then they won't necessarily be attractive.

What is the $/kWh of industrial supercaps right now? The article you linked above had some projected numbers from startup companies, one of which seems to have gone under and the other of which is still in startup mode. I just helped put together an experimental off-grid PV system with 3kWh lead acid capacity at somewhere from $200-$250/kWh storage cost.

Is there anywhere a hobbyist or researcher could buy a few kWh of indsutrial supercaps?

Umm, this is how government funded R&D is supposed to work. The government spends money on projects that are too risky or too long-term for private companies to easily justify funding on internal dollars. Then when a higher level of technical maturity is reached private companies can take the body of knowledge, develop their own products for the government and private users, and start selling products to the government (and other customers) for less money than the government would have spent to build to develop and build it themselves (or pay a contractor to do for them). It's a win win win: the company wins, the government wins, and non-government customers win because they can now buy tech that up until then was controlled by the government.

It has happened in all sorts of fields including space. Government science and R&D funding has helped advance a huge number of fields.

But the government and many of the external researchers it funds are not in the business of making and selling products, so the plan for "technology transfer" and "transition" (to industry, whether real industry or the "prime" contractor industry) can be an important consideration in receiving govt dollars for a project depending on the work and technical readiness.

The world of manufacturing is faster and more flexible than ever because of cheap human labor in Chinese factory.

Need to have your factory assemble a new phone? You can train human employees in a day.

Maybe your supplier ran out of LCD screens, so you buy an order from another supplier but it comes loaded differently into a differently size box, or maybe it has a different connector. Time to call in your robot consultant to reprogram the purpose-built industrial bot so it can handle this new form factor. Oh wait, you have human labor? Just show him how to do it slightly differently and away you go.

If you had read the articles you would see that this robot is designed to be quickly configured to perform new tasks with a minimum of technical skills required. There will obviously still be a role for purpose-built bots that have better economy of scale, as that will be cheaper for extremely high quantity long production runs. A lot of manufacturing is meant to be fast and flexible with production or assembly lines being set up in days rather than weeks/months, maybe to handle the initial burst of demand as a new product is launched, then set up to produce a new product a month later. This robot is designed to enable US labor to compete with cheap Chinese labor in the world of fast and flexible manufacturing with less of a cost bump for the US product.

Even though oil does seep into the ocean naturally in some locations, human-caused oil spills are still bad news. Life will find a way, eventually, but in the meantime there is significant damage to the local ecosystem which was caused by humans, and resulting damage to human livelihoods. Even if you don't care about the ecological damage for its own sake do you not sympathize with the damage to human livelihoods?

To compare it to another scenario: floods sometimes happen naturally, but are sometimes caused by dams bursting, maybe due to human negligence. Does the fact that it can happen naturally make it okay if a burst dam destroys a town?

Is it actually common for TA's to deliver lectures? At my school professors are not allowed to require their TAs to lecture - any professor who made a habit of it would find themselves in hot water from the labor union. Maybe it's different at other schools.

Sometimes grad students can teach courses as a lecturer but it's difficult to get approval for this, and it is a higher pay grade than TA because it's more work and skill. Introductory language courses are often taught by grad students, usually with one professor creating the lesson plan, teaching one class, and supervising a number of grad student instructors who teach the remaining classes in a given semester.

Just because you personally and possibly your industry are overworked doesn't mean you should belittle how hard teachers work. Why are you putting in so many hours anyway? In a school the (good) teachers put in extra time out of the love of teaching and their desire to see their students succeed and the teachers are overworked because there's not enough money to hire extra teachers.

Is IT really an important enough job to be putting in 100 hours per week? That's 14 hours/day 7 days/week during crunch time!! Are you at least getting OT or ST for those extra hours? You might be doing it out of love for your job but your manager and company are doing it because they love you putting in ~1.5 people's worth of work for 1 person's pay.