Public health is no longer an O-Ring production function

In the bad old days, health care in poor countries was just terrible. It wasn’t only the poverty, lack of hospitals and pharmaceuticals, and unsanitary conditions.  In addition, doctors gave very bad advice and they also didn’t work very hard, as outlined in this paper.  Citizens suffered accordingly.

Those conditions have improved somewhat, but actual health outcomes have improved a lot.  You still can’t trust the local medical advice in Tanzania, but guess what?  You have much better vaccines, greater access to antibiotics, more NGOs running health clinics, and better health care information, sometimes through the internet.  If your kid has diarrhea, let the kid drink water, even unclean water!  As for antibiotics (NYT):

Two doses a year of an antibiotic can sharply cut death rates among infants in poor countries, perhaps by as much as 25 percent among the very young, researchers reported on Wednesday.

In other words, the quality of the most important part of health care treatments bypassed the rest of the problems in poor economies and grew rapidly, even in countries with only so-so economic growth.  The rate of reduction in child mortality has tripled in many countries since the 1990s, and by no means are those locales major economic winners as say Singapore and South Korea were.

Therein lies one of the most important (and under-reported) global changes in the last twenty years.  It is now possible to have a decent public health system in a country with poor or mediocre political and economic institutions.

In other words, public health is no longer such an O-Ring service, an O-Ring service being one where everything has to go right for the service to be of decent quality.  And advances are much, much easier when the O-Ring structure no longer rules.

The O-Ring citation is to a famous Michael Kremer paper — a trip to the moon is definitely an O-Ring process, because if one step is off the whole mission probably is a failure.  But tasty fish curry is not — you can get a splendid version in some pretty dumpy countries, maybe even a better version in poorer places.

Electricity, however, it seems is still an O-Ring service, as evidenced by the recent power blackouts in South Africa.

What else is likely to become less of an O-Ring good or service in the next few decades to come?  And what can we do to hasten such progress?  Is there any chance of quality software production making that same kind of transition?  Or might some goods and services return to a greater connection with the O-Ring model?

For this post I am very much indebted to a conversation with Garett Jones.


That electricity is still an O-Ring service is also evidenced by the recent power blackouts in California.

And Boston, and Flagler County, in recent news,

Seriously, it's not so much "o-ring expertise" as the fact than an electric grid is one thing, one network. There are some degrees of redundancy, but no one on earth invests for 100% reliability. That would put you far, far, into declining returns on investment.

Detailed information here.

Maine had the most frequent interruptions. It's probably not the expertise. It's probably the ice storms.

(The MR narrative on this might be a bit whack.)

+1, Indeed.

In a country with too many children, too little opportunity and a culture that counters efforts to improve their opportunities does it make sense to offer vaccines and antibiotics? If they live in abject poverty because of too many people especially too many children why allow them to double the population?


So you think it is far better for a country with 100,000 dying every year to save every life and allow reproduction to skyrocket such that now 1,00,000 die every year???

You were scary back in 1940 when my granddaddy would have shot at you, but back then you had a gun. Now you've got a keyboard and you have all the disadvantages of being a moral monster without being an actual monster in reality.

Look up demographic transition. Try to understand what it actually means. Africa is the only continent where the bulk of the population hasn't gone through it. And we can see it happening there with fertility rates in South Africa falling from 6 in 1960 to under 2.5 today and we can see Equatorial Guinea at 6 at the start of this century bending down to 4.6 as they went from poverty stricken to well off for the region.

Improved medical care helps the demographic transition occur because parents can be confident that if they have only two or three children they will survive.

That is because you would rather posture than use your head. In a country that can support 5 million people and 30% die from childhood disease you would give them such great health care that they reach a population of 10 million where now 50% die from starvation. AND you would consider that as "humane". Some things arte better left alone.

Did you read my comment? Did you look up demographic transition? Did you see where I pointed out that improving health care reduces the fertility rate? You want to show me information from Europe, Asia, or South America that shows the opposite actually happens? That can't be done because that's not what happened there.

So they are emigrating why? By your statement all should be well in their home country. The facts don't support your theory.

The goggles! They do nothing!

Okay England's fertility rate peaked at around 1810 at just over an average of 6 children per woman. It is now 1.8. This is below the replacement rate of around 2.1. Every country goes through this transition as they economically develop. England was the first country to industrialize so it was a gradual process. It happens a lot faster theses days. Just 10 years good economic growth is enough to result in significant declines in birth rates in poverty stricken nations. If you want to limit suffering then:

1. Don't get in the way of economic development of these nations.
2. Specifically promote the education and equality of women.
3. Improve health so people can be confident they won't need to have many children so they can be sure some will survive.

I hope you are right.

Electricity has plenty of potential to become a non-O-ring product. Before subsidy or tax rooftop solar is around $1 US a watt here in Australia and electric vehicle manufacturers are buying battery cells for around $100 US per kilowatt-hour. Solar panels themselves are down to around 25 US cents per watt. So people in South Africa or where ever can set up solar power and battery storage, leading to distributed energy generation and grid ancillary services.

As for hastening the process, the US should get rid of its horrible tangle of permitting requirements for rooftop solar. I'd suggest one simple national system, but America is the land of the free so no permitting requirements might be more in keeping with the national character.

But many things about distributed solar in the US are bizarre. In Australia you can get a small solar system installed for what the average US installer pays to acquire a customer. That is advertising and other customer acquisition costs.

There's an (emotional?) appeal to every house being off-grid, or off-grid ready. But take something less complicated and less critical, like a swimming pool. It's fun most of the time, but a PITA when you need a pump replaced or filter overhaul. And that's something you can go without for a week or two while you sort it out,

So wonder green dreams include so many solar technicians?

There is an implied model in much of the solar commentary of the householder putting a few panels on the roof, a small battery in the garage (by the electric car), and viola, they are off the grid, and live happily, carbon free, and with no power or fuel bills ever after.

This model might not be a bad choice for a remote station in the outback, it doesn't scale well to urban areas (that 200 unit apartment block just doesn't have the roof space, among other issues), and is wholly infeasible for a hospital or a steel plant.

There is a reason most people don't have their own water systems, septic systems, and raise their own food.

It's not even feasible in most sub-urban neighbor hoods. There's not sufficient room on the ground, roof top panels are problematic and in the best case you are left with a system that might provide most of your power but still needs grid backup multiple times per year.

That’s the conclusion I’ve come to personally every time I’ve looked at it. Real grid independence is not realistic except in exceptional cases, and the numbers for household rooftop solar just aren’t compelling for me.

In Australian around one quarter of homes have rooftop solar but very few have any kind of battery system. This is because:

1. Batteries don't pay for themselves, and...
2. The grid (mostly) works.

But if batteries fell in price and/or the grid often didn't work, we'd see a lot more of them.

We have various schemes where companies are trialing putting batteries in people's homes and seeing if they can get it to pay. One sticks solar on their roof and a Tesla Powerwall 2 in people's homes for free and gives them a discount on their electricity bills, but they don't have control over the battery. Others want the household to contribute to the cost of the battery system. It's still in it's early days and they are still trying to work out how to most effectively extract money from households in the face of competition.

Maybe there's an emotional appeal to being off-grid, but I think a Johannesburg business is going to be more concerned about whether or not they can keep operating when the grid goes down.

As they can now pay around 17 US cents per kilowatt-hour rooftop solar, can definitely pay for itself at Australian prices. Rooftop solar is being installed for less than $1 US per watt Australia but I don't know what it costs in Jo-burg. A hybrid solar inverter plus a modest battery system helps avoid disruptions in production due to an unreliable grid. There is no economic benefit to attempting to use battery power to go off-grid at their current prices unless possibly the grid is really useless. And then it's a solar plus battery plus generator setup that will generally make sense.

Electricity in the developed world will remain reliant on huge electric grids. But in places that don’t already have those grids? Solar and batteries can provide those places a lot more electricity than they currently have access to.

Crop yields, hopefully. New genetic technology allows faster improvement than the old transgenic style crop modifications and can be economic to do for crops not named rice, soybeans, wheat, or corn. This could help make some of the government inefficiencies in roads, irrigation, and fertilizer less impactful.

The characteristics of what work are single purpose easy implementation solutions. Things that don't demand an infrastructure to make a difference, but benefit from infrastructure if it is available.

As mentioned above, solar panels along with the hardware to make it useful could make a huge difference. A working useful system with devices that do work using the solar generated power isn't trivial or cheap.

Something like trade guilds would probably be a good idea. They have their disadvantages, but to keep any system operating requires people who know how, and guilds provided these skills as well as a structure in which it could perpetuate. As societies mature they are no longer needed, but if you can't keep basic infrastructure running the basic services provided by a guild is a start. The power they accrue is not a negative in these circumstances; providing services to the many times corrupt and incompetent authorities by someone with the power to counteract the worst impulses can generate a productive friction that benefits both.

In Australia we can install commercial rooftop solar for under $1 US per watt before tax or subsidy. A lot of commercial enterprises also have generators to cope with grid failures. We seeing an increase in businesses that install a solar plus batteries plus generator set up. As battery prices fall we're obviously going to see a lot more batteries.

This is in a country with a functioning grid. in developing countries they are going to take technology developed for Australia (and the American Powerwall 2 is designed for Australia since it's seen as a better market than the US) and use it to cope with their shoddy grids. It's already happening. Chinese companies that cut their teeth making solar inverters for the Australian market are now selling thousands of small ones in South America and other locations. As batteries fall in price there will be more focus on that side of things.

The formula here is to take two buzzwords and combine them into a sentence to see how the ideas play. For this post it is public health (vaguely defined) and O-ring production function (a little more precise but very buzzy). A few posts back it was cooperation (very vague) and inequality (also vague).

What are the Malthusian implications of medical improvement without economic growth?

The developing world as a whole has averaged above 4% economic growth every year since 2000 except for the global recession years of 2001 and 2009. And this average is dragged down by very poor performance outliers in countries that are actively in civil wars or under US sanctions--remove those.

Improved public health can also improve economic growth going forward--Maoist China experienced massive gains in public health and life expectancy without economic growth, yet those gains translated into a demographic dividend that enabled fast economic growth in China once communism ended.

"What are the Malthusian implications of medical improvement without economic growth?"

Much the same as the Malthusian implications about everything else absent economic/technological growth. Basically horrible starvation. But that's not the world we live in. We've had economic and technological growth for centuries.

I think your question is really about population growth. The good news from the bad old Malthusian perspective is that birth rates fall very quickly after infant mortality improves.

Gapminder video

This is already what happened due to how badly the Russians handled treating tuberculosis, which compared to modern countries, remains common in Russia - possibly because the Russians, much like other 3rd world countries, cannot run an effective vaccination program..

The rest of the world can thank the Russians for creating multidrug-resistant tuberculosis, in large part because of how they followed the sort of public health model guaranteed to create that problem.

This is only partially true.
a) MDR is not that common in the shithole world for now
b) infection progresses to disease in only ~10% of the cases
c) mortality rate even then is low (20-25%)

But fingers crossed a more potent strain arises in south africa and starts taking out subhumans in large numbers.

You are a sad, strange, little man. You have my pity. Farewell.

Tuberculosis cannot be prevented by vaccination.
The BCG vaccine only decreases your risk of developing active disease by 30-40 percent, and not for a decade after vaccination.
But I guess deliberate ignorance about everything is strong with you racial liberasts.

Also, as I've mentioned before on this blog, I'm not russian (I belong to one of those other slav groups) but keep using that useless rhetoric.
All slavs are racist because we are not cucked like the racial liberasts in the west.

*not for a decade* = only for a decade

John, were you honestly taught such idiocy in a Slavic language school? - 'Tuberculosis cannot be prevented by vaccination.' Of course it can, as seen by the public health results in any modern country like the Czech Republic, Spain, or the USA. As a matter of fact, before the Russians created MDR TB strains through incompetence in using antibiotics effectively, TB was basically considered extinct in such countries.

'strong with you racial liberasts'

Are the tells part of the official guidelines to be followed?

'I belong to one of those other slav groups'

And yet, all those other Slav groups are proud of their heritage, most definitely including their long history of resisting Russian aggression, along with making fun of backward Russians. A Pole or a Czech or Croat would be insulted to be considered Russian, and have no problem pointing out they come from a modern country, compared to that shithole called Russia.

'All slavs are racist because we are not cucked like the racial liberasts in the west.'

See, this is what is so hilarious - Poles or Czechs or Croats most certainly consider themselves part of modern Europe. Which makes sense, considering how advanced they are compared to Russia. And of course, they are not as 'cucked' because they never mixed with the people in various -stan countries or Siberian natives or non-Slavs like Armenians or Georgians.

More liberast ignorance.
Most cases in western europe are not drug-resistant and most cases are actually among your pals - the subhuman non-white immigrants. They are not being imported from russia or anywhere else in eastern europe (well, actually the gypsy scum immigrating to germany from the Balkans probably have higher incidence than expected). But don't let that get in the way of your narrative.

Tuberculosis was not "prevented" by vaccination. Its frequency was decreased from its peak in the early 20th century via better living conditions, less crowded homes and workspaces with better ventillation (reducing transmission), decreasing the population of underweight people (yes, BMI is the strongest factor predicting the transition from dormant to active disease - which you would have known if your knowledge of tuberculosis was not gained solely by a 2-minute glimpse of the wiki page between writing comments in this thread). Vaccination helped but was not the most important factor and it alone cannot "prevent tuberculosis'.

"A Pole or a Czech or Croat would be insulted to be considered Russian"
Some would be insulted, some would not care (I don't).
There is absolutely no upside in me revealing my exact slav ethnicity. It's not relevant to any comments I have (so far) made in this blog. If I tell you I'm ethnicity X, you will just shift your pathetic trolling to something else (for example, trolling poles about being dominated by external powers for most of their history or trolling slovaks for not being a real ethnicity or whatever).
Better leave you squirming with the ridiculous russian troll thingy - that way, you are simultaneously embarrassing yourself and I'm having a laugh.

"they never mixed with the people in various -stan countries or Siberian natives or non-Slavs like Armenians or Georgians."
That is true, at least 15% of (self-identified) Russians are not white.
That's why, when we create the white ethnostate, citizenship will be conditional on DNA sequencing. Everyone from russia with more than 5% ancestry tied to the caucasus or central asia will not be allowed in.


'Most cases in western europe are not drug-resistant'

Absolutely - that is mainly a shithole Russian problem. It also helps, compared to what you think, that infant vaccination for TB is routine in Western Europe. And strange that you left out the antibiotics used to treat TB cases to prevent its spread post WWII - which also worked very well, until the Russians created MDR TB strains through their inability to use antibiotics effectively.

'Some would be insulted, some would not care (I don't).'

You don't know any Czechs or Poles or Croats, do you? Which is not a surprise, though who knows, you might be from Belorussia, the only other Slavic country idiotic enough to remain within Russia's political sphere. Even though they changed the name to Belarus, getting rid of any obvious connection to Russia.

'trolling poles about being dominated by external powers for most of their history'

Often Russians, who follow the wrong religion anyways. Besides, Poles are part of modern Europe, and know how to take as good as they get in such exchanges. Poles have never been punished for having a sense of humor, compared to Russian history.

'or trolling slovaks for not being a real ethnicity'

You do not know any Slovaks either, do you?

'Everyone from russia with more than 5% ancestry tied to the caucasus or central asia will not be allowed in.'

So sad - real believers in a true white ethnostate follow the one drop principle, but then that would mean no Russians allowed, wouldn't it?

I've lived in 3 different slavic countries (2 excluding my own) so I think it's safe to say I know at least 3 orders of magnitude more slavs than you do, if you know any at all.

We slavs definitely consider ourselves to be part (some would say a peripheral one) of western civilization as it was before the second world war and the sixties. Of the currrent cucked liberast version - not so much.
You can try to promote your liberast values considering non-white subhumans as equal to us at an independence march in Poland, at any Serbian or Ukrainian nationalist march or at Lukovmarsh in Bulgaria and see where that gets you. I'd wager it'll be in a hospital.

"real believers in a true white ethnostate follow the one drop principle, but then that would mean no Russians allowed, wouldn't it"
That would mean a significant (in size) minority of russians not being allowed in, yes. I don't have any problem with that.

No Pole or Czech would ever consider themselves a peripheral part of western civilization, so it would appear you have never lived in those two countries.

'at an independence march in Poland'

Like this one? 'Hours later, Borodacz stood beside the defaced monument to address a nationalist rally. “They told us that the Soviets are a brother nation,” he said to about 100 supporters. “And today, in the name of unity, they tell us to play with the butchers at a feast, to call our occupiers friends.” Waving Polish flags and banners of the far-right All-Polish Youth, the protestors marched through town, congregating before a stage on which a folk band was performing songs for the reunion event. “From trees, instead of leaves, communists will be hanging,” they chanted, drowning out the music.' The idea of ethnic Ukrainians considering themselves comrades in arms with Russians is even more hilarious than your usual writing.

'That would mean a significant (in size) minority of russians not being allowed in'

No, it would mean essentially all Russians being excluded. Poles or Czechs would vastly outnumber Russians in your dream state. But since you proudly proclaim not being Russian, you would probably be OK with that, just like any member of the All-Polish Youth.

Education. 1) Access to thousands of new apps, most of them free: Khanacademy, Duolingo, Codacademy, Quora, Wikipedia itself. 2) The return in rich countries (in poor country they never went away) of for-profit private providers specialized in concrete, marketable skills, like Flatiron or Fullstack for coding.

Money, with the maturity of electronic private currencies, backed by hard assets like gold or traditional currency.

Hopefully government, if the nascent centrifugal trend against the Nation-state keeps growing. More traditional States from secession, less power to supranational cartel like the UN and similar entities (they might have had a value after the war, but now they have been highjacked by the “Progressive” elites), hopefully the emergence of private quasi-States that could reintroduce the medieval concept of suzerainty.

Apologies, the last word in the second paragraph was meant to be “crypto currencies”

Some is changes in mortality per unit of GDP per capita.

Some is actual increases in GDP per capita, and that most improvements in mortality and public health happen at low bounds of GDP, now or in 1960 (see - - Fig 1).

I think you may be over-emphasizing the former, TC. Countries don't need to be "major economic winners as say Singapore and South Korea were", since being a major economic winner doesn't do too much to your national health, relative to going from $0 to $500 or $500 to $10000. It's quite enough to be a small economic winner, convergence not needed, a "broad global layer of what a Pakistani brickmaker would consider to be prosperity" is quite sufficient for the principle increases in life expectancy.

True, although I would draw a distinction to premature death due to things like malaria or easily preventable childhood diseases versus premature death due to things like obesity or drug use. In the former two cases (more prevalent in poor countries), the premature death was completely involuntary. In the latter cases (more prevalent in rich countries), people have voluntarily accepted a shorter life expectancy in exchange for the utility they get from consuming unhealthy foods and drugs that people in poor countries cannot afford.

So if two countries had the same life expectancy, but the major cause of premature mortality in one country was drug use, and in the other was infectious disease, I would say that the first country is healthier than the second because you could choose to be healthier in the first country by simply foregoing drug use, but you can't choose to avoid infectious disease in the second country. Thus, given that many (perhaps even most) premature deaths in rich countries are due primarily to lifestyle factors within people's control and that poor countries cannot afford, I would argue that the true difference in health between rich and poor countries is larger than the life expectancy statistics suggests.

That approach would not seem to have a clear pattern except at the lowest bounds. Many poor countries (5-10k) in the middle of the distribution are probably more dragged down by things like smoking than many rich countries, for'ex. Russia for ex. (Check out smoking death rate estimates - You'd probably be more likely to strengthen the change across the 0-10000 GDP range rather than strengthen an argument for further gains beyond that.

In general I don't think it's a useful framework though - as an individual you could choose to avoid many risks in poor nations as an individual, and many risks in rich nations. Do we extract HIV deaths from Africa, because they could choose to use more contraceptives, etc?

It would probably make more sense to look at pure measures of infectious disease deaths that can only be controlled by public health, if that was your methodology, as most other things will involve an element of "You could choose not to do that" in any country. (You could choose to give your child water, even unsafe water, in any country, to prevent diarrhea deaths, to use TC's example. It seems hard to argue that this is somehow more or less controlled by choice than an excessive diet or drugs, less so if anything.).

Zaua, to add to my above comment, consider Death Rate From Obesity statistics -

While we might think naively "Oh, richer countries tend to be more obese, thus they'd be a big winner from effectively removing death rates from obesity from consideration in life expectancy".

But this is not necessarily so - countries with the pure highest death rates from obesity are the 'stans and Russia, Egypt, Latin America, North Africa. (China has an estimated death rate from obesity similar to Germany, despite far lower prevalence of obesity!).

So if you remove death rates from obesity, you actually may most advantage a bunch of Middle-Income Trap countries and reduce the advantage of High-Income countries.

It's not necessarily the case that by removing obesity, etc. as a consideration you shift the life expectancy calculation onto infectious disease etc alone and so advantage rich countries.

Rich countries may be actually quite effective at handling lifestyle diseases (including drug addiction, STDs and so on) through money (and this is the point at which MR folks much more Right Wing than I would launch into a sales pitch about this is terrible and we ought to save the money letting these people die in order to *encourage* healthy behaviours among the remainder).

So if you eliminate lifestyle diseases, you actually shift the burden more onto things like cancer and dementia ("Hard Problem" diseases you might say) which rich countries aren't especially more effective at treating than middle income ones!

In the 1960s, the United States had an effective Phone Company that provided solid service: e.g., when you moved you'd get connected within a few days. Not many other cultures could manage that: e.g., Italy was notoriously bad at providing land lines.

Cell phones famously upended that: even Somalia had cell phone service.

Do it yourself solar power sounds promising: to power air conditioners you need a vast 0-ring infrastructure, but to just charge your phones might be something rural Africans could do from their own solar panels.

Solar power does not provide enough juice for almost any industrial or manufacturing activity. Solar panels may charge a phone but will not move an electric train.

That's not exactly correct. The primary constraint to solar power is intermittency, followed by costs and then with power density being at best a distant 3rd. Sure you would have to aggregate power and transmit it over distances to provide sufficient power for industrial use, but it can be done.

Sure, solar power is a relatively low density source, but it's not the lowest density source we use. Wood and bio-fuels are lower. Hydro is effectively lower.

However, to your point, solar (and wind) will primarily be used in a support role for electricity production for the next 30 years. It will be that long before they become the dominant source. Assuming no major changes to trend lines. (Yes, that's a big assumption, but true with almost any 30+ year projection.)

If power companies invest in vast solar farms, would that be a good thing or a bad thing? Would it preserve the O-ring production function for power generation?

The large positive health outcomes as the result of nominal health care services (i.e., the introduction of vaccines and antibiotics) is an example of the value of health care at the margin, when the starting point is zero. Availability of clean water likely produces a similarly large health outcome.

re: "vast solar farms", it depends.

If the utilities invest in industrial scale solar outside Dallas or Atlanta, with the goal of meeting the incremental peak power demands from AC during the afternoons, that's probably a reasonable to good engineering/economics idea. In moderation, it leaves the O-ring state of grid power untouched.

If the utility for Bangor Maine replaces their combined cycle natural gas power plant with solar and wind, they break the O-ring, and some still winter night they are going to have real problems.

Vast solar farms work similar to large power stations. But they do have the advantage of not needing any imports of fuel or even parts for extended periods of time, so they are more resilient than traditional power stations. So that is an advantage. In Australia the large majority of solar is on roofs as that is far more cost effective, but it's not hardened and won't work during a power failure because our grid, for the most part, works.

On software as an O-Ring service, Consider this analogy: As David Hays and I argued in The Evolution of Cognition, in the ancient world doing arithmetic calculations was a complex and uncertain business. Few could do it well, and mistakes were often made. The use of place notation and zero, the so-called Arabic numeral system, allowed for the compact and coherent statement of arithmetic procedures, algorithms, and arithmetic became something to be taught to school children. Can computer programming be recast in a similar way? If so then software production will no longer be an O-Ring service?

How might that happen. Perhaps AI in its machine-learning form (the current form that's being hyped to the skies) has something to offer. In a recent conversation with Kevin Kelley, Marc Andreessen said that his VC firm sees the future of AI as a platform or an architecture and not simply as a feature to be added to a product coded in the ordinary way (07:45). He doesn't give any technical reason but I can hazard a guess.

Machine learning works very well in a closed environment, such as chess and Go. It's not so effective in an open-ended problem space, such as natural language (I know, I know, still interesting stuff going on, but there are problems...). Properly conceived many application domains can be conceived as closed environment. So, let the machine learn that. I spell this out in a little more detail in a recent blog post.

'On software as an O-Ring service,"

Software is not an O-Ring service. People have a problem with time functions. An O-Ring service is something where a constant supply is a critical dependency path.

All primary (new) software development could halt tomorrow and the existing computer infrastructure would remain largely intact. Granted, there are plenty of edge cases, security patches, bug patches etc, where functionality can be improved, but that's not primary development.

I hadn't read the article when I made my comment. I've now blitzed through it. I'm not sure you're right.

Tyler referred to "quality software production"; I ommitted "production" from my comment, but that's what I meant. The third section of the article suggests that the argument can be generalized to "production functions in which quantity cannot be substituted for quality." That's certainly true for (primary) software production. You can't substitute 10 or a 100 journeyman programmers for one expert, and real experts are in short supply.

The problematic nature of primary development prompted Fred Brooks to publish the Mythical Man Month in 1975. It's been almost two decades since I worked in the industry but as far as I know primary development is still a risky and uncertain business. So Tyler's question still stands.

+1, good points

Just a footnote, a sharp cut in death rates is not what we typically relate to a decent health system.

Medicine is not only about saving lives, it's about giving life to you years: freedom, mobility....or productive years in economic terms. Productive years are increased by treating injuries better, minimizing hospital stay periods and managing pain.

We're so used to used to breaking bones, stay at home a few weeks and return to work that we forgot this kind of injuries mean lifetime disability in a place without a decent health system.

Systems have 1:1, 1:n, and n:n relationships. They progress in redundancy in that order. Services, usually n:n, are fine. If one burger shack is closed by the health department, there's another. Critical utilities are 1:n and a failure affects more customers.

So, I guess overall I'm not convinced that expertise is the critical distinguishing factor here.

After all, maybe those "bad doctors" wanted to prescribe those "missing antibiotics" all along.

To extend this just a tad,

There are millions of cars going millions of places this morning in America. It is an n:n effort with a high number for n. When some cars break down, as they will, we don't say the system has failed. With high n, it's easy to see failures in context.

Compare that to the moon mission in the essay above. That's very much 1:1. One supplier, one trip, all eyes on it.

I don't know, maybe I'm just restating this "o-ring production function" in a slightly different way, but as I say it might be more about the 1 and n than expertise.

Or related to expertise in that old sense that "if you have all your eggs in one basket, you should watch that one basket very carefully."

Advanced civilization is the most basic O-ring. That is why things have improved in backwards or catchup countries while arguably having become much worse for the richer countries. Especially the long-standing citizen groups as opposed to the gains for the recent immigrants. There are a lot of people in the US and Europe who would rather have lived in the world of the 90s on most technological margins with more traditional social norms and a lot fewer highly dissimilar immigrants and norms that disfavor forcing assimilation onto the newcomers (i.e. the default position for most of the Western world prior to the last few decades).

If you broke your leg in Australia you'd end up with one leg shorter than the other as the traditional method was to set the bones next to each other to knit. This may have been a less risky option in a world without x-rays, but has an obvious drawback.

What's the opposite of an O-ring process? A bottlenecked process?

Professors Cowen and Jones, are you sure medicine (as measured by child mortality, life expectancy, etc.) ever was an O-ring process?,_from_1770_to_2018.svg

What else? Urbanization in antiquity was O-ring. High urbanization rates were rare, only during ideal imperial conditions such as the height of the Roman civilization or Song Dynasty. Now urbanization is simply bottlenecked by rural farming adopting high yield techniques and a nascent urban industrial sector.

I'm sure long ago, good fish curry used to be O-ring, as it requires a vibrant spice trade. And now in rich nations perhaps it is bottlenecked by health regulations.

Virality in news and culture used to be O-ring (free press, strong institutions, print technology, intellectual density), now it's just bottlenecked by smartphone ownership rates and domestic internet censorship.

Above-replacement fertility used to be a bottlenecked process, now it is O-ring, requiring nearly perfect economic conditions and labor regulations. Similarly, obtaining a life long partner used to be bottlenecked, now it's O-ring.

Access to foreign cultures and ideas, philosophy, modern art, etc. used to be O-ring, now just requires an internet connection and fluency in one major language.

Starting a new colony used to be bottlenecked by naval projection, now it's definitely O-ring (readers of MR might be lulled into believing charter cities and "sea steading" are strictly modern innovations).

Low interest rate bond issuances used to require countries have O-ring macroeconomic conditions, but now...

Its useful to keep in mind how increasingly dependent much of today's software is on a very complex infrastructure of electrical power, high bandwidth communications, and remote software services. Its intensely O-ring.

In fact, as software becomes more embedded in every day activities, this O-ring fragility is probably increasing. One can only look with wonder at the idea that people are giving up their house keys for an app on their phone.

Moving software to the cloud certainly makes it an O-Ring service, but there's plenty of stand alone software. And it's silly to say "Software" requires a complex infrastructure of electrical power. Hardware does, but it's trivial to provide backup power to servers. It's just pointless to make the backup power extend for a long period of time when the machinery they control is dependent on the missing power.

Software companies attempt to efficiently combine resources to create software that can perform a function that users will pay more than the software's risk adjusted costs to create.

I disagree that moving software to the cloud impacts whether it is an O-ring service or not. Moving software to the cloud can, sometimes significantly, reduce costs or risks and increase the software's value.

It's the remaining tasks of software firms that continue to be O-ring services. I'm hard pressed imagining software ceasing to be an O-ring service though technology is continually reducing the investments needed to create impactful software (this will sometimes enable for the first time large investment possibilities). So we may get more software investment and more software but it remains an O-ring service.

"I disagree that moving software to the cloud impacts whether it is an O-ring service or not."

Moving software to the cloud creates a delivery dependency that doesn't exist without it. For example, I can run my servers off of relatively simple software, developed from a talented single coder, pretty much independent of a massive software development industry. It could even be software that's given away for free. (Linux)

For the cloud approach, I'm always dependent on Software as a service, running nearly flawlessly 24/7. That requires a bevy of background engineers and technicians constantly providing support. It may well be the low cost solution, because 1,000 techs spread over a user base of 10 million is still cheaper than 1 IT guy for a 25 person firm. But it is a much more path dependent scenario.

As I threw out my hdtv antenna (I live too far and behind a hill for a steady picture) I thought, wait a minute do I have any way to get news with the internet down? Do I even have a radio in the house?

I remembered my car still has AM/FM, so good for now.

You don't have a phone?

No good old copper landline, and I'd put low confidence in the cell system working with a full regional internet down level disaster.

"Hurricane Harvey knocked out internet and telephones service to almost 200,000 homes, more than 360 cell towers and 16,911 call centers. A study from the Federal Communications Commission shows that about 1,000 cell towers were knocked out during Hurricane Katrina.

But reports that the service to the downed cell towers was restored to about half of those customers only a few days after the storm."

So for those few days you probably want an AM radio.

I read somewhere that if you still have a phone jack, you can plug in an old-style phone (if you still have that) and call 911 even if you don't pay for a landline anymore. Supposedly.

Ah, interesting. No old phone. But I suppose in all these scenarios it's what "me and the neighbors have." Somebody probably still has the landline.

Assuming the old landline works. The companies that own the buried landline phone cables just don't seem as interested in maintaining them as they once were and, as decades-old insulation fails, they don't work so well anymore.

Although it may be comforting to imagine that all that old infrastructure would be there if for some reason the newer tech goes down and stays down for some time, it was never designed to last forever.

As stores, factories and warehouses use automated systems to minimize inventory by ordering just-in-time, a widespread internet breakdown would cause a great many things to become scarce very quickly. Even if stores could figure out how to stay open with their POS systems down, which is unlikely.

Oh well. But if stringing up phone lines is ever on the apocalyptic to-do list, I've got a source for the glass insulators. Mother tried to corner the market back in the 70s.

The radio is a good idea. It’s easy enough to get a small AM/FM, and keep a pack of 10 year AA. A small battery (usually used as backup power for phone or iPad) and a small USB LED have proven useful. For $100 one can buy a useful amount of emergency capability.

I bought a solar-rechargeable camping lantern at REI mainly because it seemed like the traditional battery-powered ones didn't last very well, or used up the alkaline batteries really quickly. It has a port for charging devices.

I love it, not because I give any thought to emergencies - I don't - but it provides the most beautiful light. GoalZero CrushLight. Come on, MR, my one and only product placement, I promise.

In earthquake country on the west coast of the US, residents are instructed to have a radio and spare batteries -- or a radio powered by solar or by a hand-crank.

There's a lot of junk emergency radios out there, some of which I've used in real life power outages only to have them stop working after a couple of hours. But I finally found what appears to be a good radio (which even has a light included, but I've also found a good hand-cranked flashlight too).

Those hand-cranked devices are not my only or even my main emergency devices. But I want them as back-ups in case I run out of spare batteries. (Many government websites still recommend having enough emergency supplies to last for three days. But Oregon, being either more realistic or more pessimistic about its emergency services recommend two weeks of supplies. So that's 14 gallons of water per person in your household. And I don't know how many batteries but even with a drawerful of spare batteries I'm not counting on them to last 14 days. Hence the hand-cranked devices.)

The thing about cell phones is that there are not very good, you know, phones. i-phones are the worst. I can always tell when someone is calling on an i-phone because they sound like they are talking through a tunnel. Of course, the phone is just one of the features of a smart phone, a feature that is rarely used by millennials so producers don't put much effort into the, you know, phone, instead focusing on the camera. Why not just buy a damn camera. Smart phones are really dumb, as they try to be all things to all people and don't do any one thing very well.

Get off my lawn, you darn kids! {shakes cane}

Bruce Booth touches on this point in a recent post on drug pricing. With pharmaceutical you don't need a competent surgeon, you need good manufacturing practices and distribution!

The rapid build-out of cellular service in sections of Africa has moved telephony from an O-Ring service.

In addition to the cell phone and solar energy examples that commenters are giving, I think we can add online banking in Africa, where reportedly many countries have skipped the credit card phase and jumped straight from a cash economy to an online transaction economy. Maybe it wasn't literally an O-ring system, but those old banks with their limited hours and long lines were real bottlenecks. (I stood in line for hours in Venezuela just to exchange some dollars for bolivars -- and this is was in the good old days, pre-Chavez and Maduro.
In Zambia we'd learned our lesson and minimized our need for cash at all and IIRC what little we needed we changed at the airport or hotel.) Now the bank is no longer the bottleneck or single point of failure.

Even though the article and the doctors address antibiotic resistance, the thought of routinely dosing tens or hundreds of millions of people with antibiotics makes me nervous. We thought DDT was the greatest thing since sliced bread and sprayed it everywhere until we realized the damage it was doing to the ecosystem. We have been administering antibiotics profligately and in the arms race between the bacteria and the inventors of new antibiotics it's not clear to me that the bacteria are losing.

Dose with antibiotics while improving health and sanitation until it is no longer necessary.

Most forms of antibiotic resistance in microorganisms can be managed and reduced. All it needs is a sensible low cost public health plan funded by taxation.

In other words, we're all going to die from drug resistant germs.

Adding technology, counter intuitively, creates more non O-ring cases.

Taxis using Uber are less dependent on "everything being perfect" locally

Online education removes the o-ring of a good education system, especially in more rural and agrarian societies.

As goods logistics become more localized, global O rings become less important. As 3D printing becomes more advanced, O rings become less important. As energy production becomes more decentralized, O rings become less important. Software needs fewer O rings because computers are easily replaced or tossed into the cloud.

Comments for this post are closed