Optimize python functions by marking certain promises about its behavior : Python

promise: bytecode optimisation using staticness assertions.

http://pypi.python.org/pypi/promise

This is a module for applying some simple optimizations to function bytecode. By promising that a function doesn't do certain things at run-time, it's possible to apply optimizations that are not legal in the general case.

Reddit Comments: Optimize python functions by marking certain promises about its behavior : Python: "
I gave a talk on this at our local users group a few weeks ago, the slides are online if anyone's interested:

http://wiki.python.org/moin/MelbournePUG?action=AttachFile&do=get&target=promise.odp

"

The above presentation is really nice - great, simple example of the power of this technique.

I was thinking along these lines. Having code where we specify two "speeds":

(1) flexibility/expressiveness/global-mutability/side-effects-happen-globally-immediately are important (at cost to throughput and low-latency)

i.e. dispatch on pattern matching ASTs on global mutable list of patterns, global mutable generic functions, global mutable generic methods

Allow global mutable objects in general. All mutable state is handled like distributed-revision-control & write-on-change, with all communication going through a key-hole (enforcing low latency by throttling large data transfers) as asynchronous messaging as transactions (and building transactions).

Side effects are GO!  Allow global state to change, allow outside communication, all happening ASAP, might pre-calculate while waiting for reply, but wait to the bitter end for reply none-the-less.

(2) throughput and low-latency (performance) are important (at cost to flexibility/expressiveness/global-mutability) - no side effects (all "side-effect messages" are stored, to be returned as a group when function returns, maybe each "side-effect message" is paired with a continuation)

Image via Wikipedia

i.e. dispatch on low level bytecode (think: LLVM), preferred data structures are immutable and have the "shape" of directed acyclic graph (possibly a much more restricted form of directed acyclic graph, where each node has an immutable index, and parent indexes are always less than child indexes), the limited explicitly mutable state is local to OS-thread or green-thread.  No traditional asynchronous side-effects allowed.

And there are speeds in the middle, by being explicit about what you are relaxing and what you are constraining. Typically, you expect to pay a "compiling" cost to get down to the low level bytecode - again, by being explicit, if you are doing a lot of this "compiling", you can improve by being explicit about what you are relaxing and what you are constraining.

Low-latency (ability to best react to asynchronous signals) will always be preferred over throughput. If you want to put throughput over low-latency, you have to explicitly say so, with the knowledge that you have very little influence over that isolated code (isolated so able to make no compromise in throughput)

[Edit]

Hey, cool.  People are trying "Promise" out on code in the wild.  Here is creator Ryan Kelly explaining how to make use of the bytecode improvements:

http://panela.blog-city.com/the_promise_of_faster_python_1.htm#1

1. Ryan Kelly left...

2010.01.05 Tue 4:25 pm :: http://www.rfk.id.au/

Matt, thanks for taking the time to put this together. The optimizations applied by promise are certainly not in the same league as something like psyco - they have to be quite well targeted to have any measurable effect.
Some clarifications: promising a function pure() doesn't optimise that function at all, but it can speed up things that call that function by inlining its bytecode at the call site. To get this to work, you have to use constant() to promise that references to the pure function won't change. Example:

..@promise.pure()
..def calculate(a,b):
......return a + 2*b
..@promise.constant(("calculate",))
..def aggregate(pairs):
......return

In this scenario, the bytecode for "calculate" will be inlined directly into the "aggregate" function and will save the overhead of many function calls.
By far the biggest speedup that can currently be obtained using promise is inlining pure functions that get called in a loop.

We are what our ancestors did or didn't eat.

Image via Wikipedia

I want to eat both the Paleolithic Diet AND the Modern Diet. Having both available on one giant plate is very appetizing to me.

DERIC BOWNDS' MINDBLOG: We are what our ancestors did or didn't eat.: "Ann Gibbons does a nice summary of our human ancestral diet and how it has changed to give us a modern array of diseases. Some slightly edited clips:

By the time hunter-gatherer modern humans swept into Europe about 40,000 years ago, they were adept at hunting large game and had also expanded their palates to dine regularly on small animals and freshwater fish....By studying the ratios of carbon and nitrogen isotopes from collagen in bones, ...the main sources of dietary protein of 27 early Europeans and Neandertals is known; fish eaters, for example, have more nitrogen-15 in their bones than meat eaters...the oldest known modern human in Europe—the 35,000-year-old jawbone from Pestera cu Oase cave in Romania—got much of his protein from fish. By 30,000 years ago, other modern humans got as much as 20% of their protein from fish.

The next big dietary shift came about 10,000 years ago, when humans began to domesticate plants and, later, animals. The move to agriculture introduced staples of the Western diet: cereal grains, sugars, and milk after weaning...The agricultural revolution favored people lucky enough to have gene variants that helped them digest milk, alcohol, and starch...when ethnic groups abandon traditional lifestyles and rapidly adopt Western diets, they often suffer. Researchers have known for more than a decade that the Pima of the southwestern United States have 'thrifty phenotypes': sluggish metabolisms that store fat efficiently and boost survival on low-calorie diets. That's probably because their ancestors in Mexico underwent frequent famine. When they eat the calorie-rich Western diet, the Pima develop high rates of obesity, diabetes, and high cholesterol, although their blood pressure stays relatively low...the Evenki reindeer herders and other indigenous peoples of Siberia have very high metabolisms, an adaptation to the cold that allows them to convert fat into energy efficiently. When the Soviet Union collapsed in the 1990s, many Siberians abandoned traditional lifestyles and diets. They too became obese and developed heart disease but in a different way from the Pima: The Evenki retained low levels of cholesterol and diabetes but developed high blood pressure.

Image via Wikipedia

Although we are what our ancestors ate, we are also what they didn't eat. In India, for example, more than 66% of the population in some regions experienced famine during British colonialism a century ago. Women who survived tended to have low-birth-weight babies, whose bodies were small and efficient at storing fat. It's as though these babies took cues during fetal and early development about their mothers' lifelong nutritional experience and adjusted their growth and body and organ size accordingly. Human stature often tracks the nutritional status of mothers, and it can take generations for descendants to recover. In India, average height in males dropped at a rate of almost 2 centimeters per century in the decades following colonialism...When these small babies gain weight in childhood, though, it stresses their smaller organs, such as the pancreas and heart, making them more susceptible to obesity, diabetes, and heart disease. This is the case in south India today, where many people have thrifty phenotypes with less muscle and more fat per body size. Yet they are shifting rapidly to a high-fat, high-sugar diet. As a result, India risks becoming the diabetes capital of the world.

"

Doris Kearns Goodwin: “What would have happened right after September 11th if President Bush had called for a Manhattan Project for independence from Middle Eastern oil?”

Image via Wikipedia

Wow. From a long range historical perspective, we have never really had a nation-wide war on terror. We only sent some expendable troops, and started some deficit spending to help military contractors and keep the money out of the hand of the lower classes.

Doris Kearns Goodwin: “What would have happened right after September 11th if President Bush had called for a Manhattan Project for independence from Middle Eastern oil?”: "
Yesterday, Meet the Press focused on the failed effort to blow up an airline on Christmas Day. One exchange was especially illuminating (transcript here)

MR. GREGORY: Doris, you’re familiar with writing long and wonderful volumes of history. And if the war on terror, if chapter one was written by President Bush, now it’s chapter two and beyond; and it’s still very, very complicated, an entire decade really defined by, by terrorist acts at the front end and at the back end, an attempted act at the back end. So much different than the wars we have fought in our past.

Historian Doris Kearns Goodwin, as always, provided a crucial historical perspective:

Image via Wikipedia

MS. GOODWIN: True. But I think there are certain lessons, even though the war on terror is a war about individuals, loose organizations, it’s not countries, there aren’t going to be treaties. We’ve learned things from other wars that I still think are valid here. Number one, you have to have allies on your side, and I think that’s what the Obama administration has begun to do. I mean, after we made the announcement about the Afghan escalation, NATO put in 7,000 troops. That showed that some work had been done at that point. I also keep thinking that somehow what we really missed in the beginning of this decade on the war on terror, what would have happened right after September 11th if President Bush had called for independent–a Manhattan Project for independence from Middle Eastern oil? What if he’d called for a lot more people to join the Army? We wouldn’t have had these same soldiers going back three and four times. What if we’d had a tax increase, as we’ve done in every other war, to fight a war? We wouldn’t be facing the deficits right now. So I think even though it’s a different war, the need to mobilize the spirit and the energy of the American people, so it’s not just our soldiers fighting those wars alone over there, is still relevant in history’s terms.

"

Procrastination and "present-bias" in being time-inconsistent

Image via Wikipedia

The relation between procrastination and "present-bias" in being time-inconsistent - economics of procrastination of Ted O'Donoghue and Matthew Rabin.
Was thinking along the line of a round-robin circular-list policy of open projects.  Even if only apply 5 minutes of work to each, at least avoiding outright procrastination.  Also, the start of a more realistic assessment to true commitment to each project - the most can bring self to work on in the daily hustle and bustle.

Economics focus: New-year irresolution | The Economist: "In a 1999 paper on the economics of procrastination, Ted O’Donoghue and Matthew Rabin pointed out that people are often unrealistically optimistic about their own future likelihood of doing things—such as exercise or saving—that involve costs at the time they are done, but whose benefits lie even further ahead. Mr O’Donoghue and Mr Rabin showed that this sort of behaviour can be explained if people are time-inconsistent. “Present-biased” preferences mean that people will always tend to put off unpleasant things until tomorrow, even if the immediate cost involved is tiny. As long as they are unsure of the precise extent of this bias, they believe (incorrectly) that they will in fact “do it tomorrow”. But since they feel this way at each point in time, tomorrow never quite comes. Such a model can therefore explain endless procrastination.

Image via Wikipedia

"Doing it Now or Later", by Ted O'Donoghue and Matthew Rabin, American Economic Review, March 1999
"Nudging Farmers to Use Fertilizer: Theory and Experimental Evidence from Kenya", by Esther Duflo, Michael Kremer and Jonathan Robinson, NBER Working Paper 15131, July 2009

"Put Your Money Where Your Butt Is: A Commitment Contract for Smoking Cessation", by Xavier Gine, Dean Karlan and Jonathan Zinman, World Bank Policy Research Working Paper 4985, July 2009
"

Boing-Boing: Crotch-bomber psyche laid bare in messageboard archives

Image by algo via Flickr

This is pretty intense. It is a reminder to myself to love love love mean people, especially people who are mean to me when I open up. The alternative is to descend slowly into a sick world where suiciding myself along with killing innocents seems like a good idea.

God, help me to love mean people better. God bless those who don't demand everyone be nice to them.

Why don't these guys kill themselves quietly? Oh well, at least the most damage he did was burning his own crotch. Although there are rumblings to bomb Yemen.

http://www.upi.com/Top_News/Special/2009/12/28/After-bomb-plot-Yemen-in-US-cross-hairs/UPI-17621262025750/

Christ, way to snatch defeat from the jaws of victory. A guy burns up his own crotch, and my tax dollars go to bomb 1000 Yemenis to create 100000 motivated terrorists.

From the usually worthless Boing-Boing:

Boing-Boing: Pantsbomber psyche laid bare in messageboard archives (spoiler: he used sad-face emoticons): "Now that we've gazed inside what is purported to be Umar Farouk Abdulmutallab's explosives-laden underwear, let's look inside his mind, by way of an archive of postings he made to the Islamic discussion website gawaher.com.
Would the prophet Muhammad have played soccer, were he alive today? Is it okay to eat meals with my parents, even though they're unclean? All this and more he asked, and we know this because Wired Danger Room dug up a slew of links.
The following post credited to 'Farouk1986' is to me the most chilling of all, because it is the most human. As was the response that followed, from another guy on the forum who'd been to the same emotional place [insulting insipid banter removed]

Basically, the problem I'm having is that I've been having extreme loneliness...for many years. I don't really know what to do because I'm not the type who likes to go out much, and I'm just shy and quiet. Even on the internet, I don't feel comfortable posting much because it exposes myself. Sometimes people are so mean.
So I'm trying to figure out what to do. I just wish I had someone to give me attention and stuff. I wish I had someone who would be there to listen to me, and always be nice to me. It really hurts to have someone neglect me or be mean. Unfortunately, a weakness of mine is that I'm sensitive, but I think I became more sensitive after something bad happened some years ago.

Image by ukhti27 via Flickr

I wish I had at least one nice person to talk to, maybe over e-mail or Messenger. Of course, if I could find someone to marry, then Insha'Allah I would have someone in real life to give me all the attention and affection I wanted. So far, the families we've met aren't interested in me, though.

Loneliness (Farouk1986, on gawaher.com)

"

Andrew Gelman on over-use of Economics Utility Model to explain all of psychological behavior

Image via Wikipedia

I was thinking about this recently. Many times, we can model people's behavior as a boot-strap process: people use a personal, informal, emotional process to decide whether to engage in rational (or semi-rational) utility analysis, or not.

[personal/informal/emotional process] ⇒ {{{decision point}}} ⇒ [begin rational utility analysis]

If they "drop out" at the decision point, nothing worth calling a rational utility analysis even gets started.

Many people are so overwhelmed by grappling with the critical issues of life, that they distract themselves into a silly stupor that makes a rational utility analysis impossible.

Andrew Gelman: Taxation curves and poverty traps - Statistical Modeling, Causal Inference, and Social Science: "

I think the concept of utility is extremely useful, and I've used it in my own applied work (see my papers on the utility of voting and on radon mitigation or the chapter on decision analysis in BDA). Utility is a model, and it's great.
My problem is when people think that the utility model can/should explain everything.
For example, as I've discussed on the blog, I don't think the utility model is particularly useful for explaining uncertainty aversion, seeing as the essence of the 'uncertainty aversion' phenomenon is that preferences can depend on how they are framed and how they are set up in terms of probabilities--two things that violate the classical von Neumann axioms in which preferences should only depend on the ultimate outcomes and their total probabilities, not on where these probabilities come from.
I think it's just sad that utility functions have become a default way of explaining all sorts of psychological processes that don't fit the model so well (requiring the sort of epicyclic adjustments that can make the model more trouble than it's worth). I can respect the general endeavor to take a model and push it as far as you can--to see what tweaks can be done to make it work further than it was originally intended--but, at some point, I think it makes sense to recognize the practical limitations of any mathematical model.
So, yes, I don't think utilities (or, for that matter, preferences) 'exist' in some Platonic sense. But I still think utility theory is great. I think the normal distribution is great, too, even though it can be misused in all sorts of ways!

"

Image via Wikipedia

In a follow-up comment by Gelman:

Nathan (and Dan): I think prospect theory is great. I just don't like trying to explain uncertainty aversion using a nonlinear utility function of money (which, as I and others have shown repeatedly, makes no sense at all when you try to look at it quantitatively), and I really really don't like having to explain this to people over and over again, people whose technical ability is such that they could've realized in the first place the impossibility of explaining uncertainty-aversion-at-any-scale using a curving utility function. And I also don't like the term "risk aversion" casually used in a way that blurs three different phenomena: aversion to risk, aversion to loss, and aversion to uncertainty.

Aggressive-Competence in Software Development : Titus Brown

Wow, this post is great stuff!

Image via Wikipedia

Aggressive-Competence in Software Development : Titus Brown: "
At the end of the day, there are things you can control, and things you can't control. You can't control what other people think of you, and you can't control how other people (including project leaders and professors) evaluate you. But you can visibly work hard, and defend yourself based upon that evidence.
I call the general approach of throwing energy at a project 'aggressive competence', and I think it's a necessary component of effective team software development. Everyone has days, or weeks, or even months where they look incompetent or ineffective; often that's because outsiders don't understand or appreciate the work that you've done. Tough on you, but I don't think it's reasonable to expect your boss, or colleagues, to look hard at your work to find reasons to praise you. Fundamentally, it's your responsibility to 'manage up' and communicate your progress to others effectively.
...

Image via Wikipedia

This is where I think there were mismatched expectations. The students expected that they were going to be managed, helped, and given clear expectations. They weren't. So they got bad evaluations.
What do I plan to do? Well, assuming that UCOSP + MSU goes forward next term, I will be communicating my expectations quite clearly to the students. And I will be asking for regular progress reports, sent to me and CCed to the project leaders. And I'll be sending them this blog post. And I'll be failing the ones that don't listen.
I'll end with a paraphrase of one of my favorite sci-fi authors: 'every new developer has problems on a new project. The extent of our sympathy for those problems, however, will be dictated by the efforts made to overcome them.'
"

The extent of our sympathy for those problems, however, will be dictated by the efforts made to overcome them. - David Weber, The Short Victorious War

Never give advice to a fool

Image via Wikipedia

Ugh, reminds me of what a fool I have been. Of course I have ignored this sound advice, because I, myself, am a complete fool.

Subhashita Manjari (A Collection of Sanskrit epigrams): Never give advice to a fool: "

पयःपानं भुजङ्गानाम्
केवलं विषवर्धनम् ।
उपदेशो हि मूर्खाणाम्
प्रकोपाय न शान्तये ॥

- हितोपदेशः

The milk fed to a snake only increases its venom. Similarly, the advice given to a fool leads to aggravation and not peace.

Image via Wikipedia

Advice is least heeded when most needed.Every evil has its remedy, except folly.
Talk sense to a fool and he calls you foolish.

"

Talking about Climate Change: Publishing your Probability Density and Reasoning

Image by melancholic optimist via Flickr

"Phil" from "Statistical Modeling, Causal Inference, and Social Science" blog, I believe Phillip Price [ http://eetd.lbl.gov/ie/apt/staff/PricePN.html http://www.creekcats.com/pnprice/Job.html ].

As I say in my comments "Thank you for publishing your probability density for climate sensitivity, and the precise reasoning behind it. This is, practically, the _only_ way to communicate one's considered beliefs for this subject - if one is truly interested in communicating and not just indulging in motivated obscurantism or the art of controversy."

How helpful to have someone state their probability density and reasons behind it. It communicates so much.

Phillip Price: Say a little prior for me: more on climate change:

"Four out of the last 15 posts on this blog have been related to climate change, which is probably a higher ratio than Andrew would like. But lots of people keep responding to them, so the principle 'give the people what they want' suggests that another one won't hurt too much. So, here it is. If you haven't read the other posts, take a look at Andrew's thoughts about forming scientific attitudes, and my thoughts on Climategate and my suggestions for characterizing beliefs. And definitely read the comments on those, too, many of which are excellent.

I want to get a graphic 'above the fold', so here's the plot I'll be talking about.
...

Finally, we get to the graphic. Each of these probability distributions is supposed to summarize the belief of a different person. In blue, we have an 'anthropogenic climate change denier.' This is someone who just doesn't believe that doubling of atmospheric CO2 could have any substantial impact on the global mean temperature. I don't know if any such people think the effect could be negative, but maybe they do; if they don't, then just move all of that negative probability into the low positive range somewhere. At any rate, these people are convinced that there is just the right amount of negative feedback to cancel out the known effect of CO2 and the expected effect of water vapor.
...
But I think the hypothetical 'skeptic' curve puts way too much probability on very low values --- not as bad as the 'denier', but still, this is someone who is unjustifiably convinced that negative feedbacks will come close to counteracting the effects of CO2.

(By the way, none of the lines are supposed to go below zero, or even go to zero, at 6C, but the drawing software I used has done some funny stuff there and it doesn't seem worth fixing. Oh, and each of the curves is intended to have the same integral -- unity -- but since this is just a by-hand sketch, they probably don't).

Above, I've opened my soul, as it were, to discuss why I believe what I believe. Part of my belief, actually a substantial part, is informed by a very simple physical model that I believe is useful in spite of its simplicity, that shifts my prior well away from 0 as a reasonable estimate of climate sensitivity. What if you don't have the physics background to evaluate such a model for yourself? Then, you're more or less forced to choose who you care to believe: deniers, skeptics, 'experts,' journalists, bloggers, friends...

In a comment on Andrew's entry about forming attitudes on scientific issues I said this:

When it comes to anthropogenic climate change, if someone wants to allocate some probability to the chance that the skeptics have it right, I think that's a very reasonable thing to do. Make it 90% mainstream, 10% skeptics, or even 75% mainstream, 25% skeptics if you are are heavily inclined towards the skeptical camp. But there are people out there who are 90-10 the other way! If you are an expert climate modeler and you think your colleagues have the science wrong, that's one thing. If you're just some schmoe who only knows what he reads in the papers, and you choose to assign a 90% or 95% probability to the conclusions of the small band of skeptics...where does that come from? Do you really think the experts in a field get it wrong 90% or 95% of the time?

I think I'll leave it there.

"

My comments:

Thank you for publishing your probability density for climate sensitivity, and the precise reasoning behind it. This is, practically, the _only_ way to communicate one's considered beliefs for this subject - if one is truly interested in communicating and not just indulging in motivated obscurantism or the art of controversy.

I wish I had the chops to draw one myself. I don't, so I rely on the experts currently publishing articles. I find it hysterical to shriek about the corrupting influence of funding - I am not holding my breath waiting for the appearance of researchers comprised solely of incorruptible energy, freed of the need for money because they draw sustenance from the empty ether. Oh, please. All humans have their self-serving motivations, and reasonable people deal it with accordingly and with due measure. And, scientists never claimed to not be human.

Image via Wikipedia

I wish I had the chops to draw the _other_ curve with regards to carbon dioxide - loss of tonnage of fished protein/nutrition due to ocean acidification. Global temperature and ocean acidification both have a large expected impact on human civilization.

> The Kyoto Treaty was a large scale effort...

Surely the intervention can only earn the description of "large" if *some* authority judges it probable to effect the desired change. Otherwise, I could call the Dubai Towers a "large scale effort" to bridge the span to the surface of the moon.

It can be honestly asserted that no large scale efforts have been made to reduce greenhouse gas emissions.

Putting too much faith in employment "Recalculation"

Image via Wikipedia

I have great sympathy for Arnold Kling's "Recalculation Story" of high unemployment. But he shoe-horns in some foolishness into his otherwise sound theory to try to head Keynesian interventions off at the pass. The main shoe-horned foolishness is that, magically, businesses will hire back everyone laid off, and hire also new workers entering the job market. I ask, Why? It does not follow at all - Businesses have never had so many techniques for keeping their staffs low, and all the techniques are improving in quality, as well.
The interventions to "cure" high unemployment _will_ be Keynesian. Because every other race of economists pretends that high unemployment is not a problem. I am exaggerating, but only a little bit.

I would not mind a "better" form of Keynesian, but those who are in the best position to provide those improvements are too busy pretending that high unemployment is not a problem. Or that businesses have techniques to turn marginal employees into highly productive employees, and they will use those techniques instead of the tried and true methods of raising productivity and shedding employees at the same time. Because of the time wasted on these fallacies, we will get the old Keynesian cure from the old Keynesians, because the Keynesians are immune to these particular fallacies (being too busy with their own fallacies).

Image via Wikipedia

The interventions to "cure" high unemployment _will_ be Keynesian, whether I like it or not, whether anyone likes it or not. Prepare accordingly - labor legislation in the United States _will_ greater resemble Continental Europe - because the current and expected social costs of high unemployment are very real (including lost taxation and lost consumer demand, besides criminality induced by idleness).

The Job Assignment Problem, Arnold Kling | EconLog | Library of Economics and Liberty: "
> I agree that high productivity growth is very good news and that it portends increases in hiring.
Running a business, if I have productivity growth, and, for whatever reason, I wanted productivity to go back down, the fastest way to do it would be to hire more employees.
I am not supposed to admit this, but I feel compelled to be honest. Higher unemployment and higher productivity growth have no demonstrated inverse relationship.
When the social costs of unemployment come home to roost, we will see legislation to enforce lowered productivity. I can complain about it, but it would be foolish to not see it coming. Somehow, I will be compelled to hire more employees, and productivity will go down. Because employees bring with them the burden of mandated entitlements and the burden of their own inner sense of entitlement. Like it or not, it is foolish for me to pretend otherwise.
> What people are calling a "jobless recovery" is what I would call the market taking a long time to solve the job assignment problem.
The market has solved the job assignment problem by not hiring. And productivity has increased. During the 'good times', the market was more tolerant of low productivity - now less so. Businesses act accordingly. Because, at this time, businesses are not compelled to bare a share of the current and expected social costs of high unemployment (including lost taxation and lost consumer demand, besides criminality induced by idleness).
"

A depressing follow-up to a reply to my comment:

> I disagree with manuelg' first comment regarding legislation to compel more utilization of human capital - that's been attempted by, among others, the Luddites, the Wobblies, and the Communists.

I didn't say it would work as advertised! ;-) It will "work" to reward a particular voting block, as one would expect. High unemployment provides the "cover".

> ... the workforce will either develop or rediscover markets that require human capital (R&D, personal service) or the supply will re-train to enter other labor markets a la buggy whip makers and farmers

Or neither. New positions never opening up at an adequate rate is a distinct possibility. Take R&D - why not use 40 Indian PhD's to do the work of 60 US PhD's? Take personal service - a goodly portion of the work of a personal assistant can be done remotely - again employing someone without the same expectation of quality of life and compensation in a place with a much lower cost of living.

The majority of the work will be of poor quality, but the cost savings will be compelling. Like the craftsmanship of a pocket-watch compared to the time display on a plastic cell phone. People spend a smaller fraction of their income on telling the time (pennies out of the cost of the phone and service), and lose daily contact with items to true quality, and that prior level of quality becomes a boutique luxury item, and employs far fewer people.

> ...re-train to enter other labor markets...

It may in fact be easier to "retrain" a US worker to become a ethnic Chinese worker (chopstick skills, walking stooped to pass through lowered doorways, reading/writing Chinese letter-forms) than it would be to retrain a US worker to have the skills and sense of responsibility and lack of sense of entitlement that the new employment positions opening up require. Look at the malaise of Japan's labor economy as the positions disappeared for lifetime employment by a single employer, never to return. Even after years of a desperate situation, it is not 100% certain that a human will change their attitudes to match a new reality.

Not to put too fine a point on it.