6 ms·

David Brooks on Complexity

19 points by klochner 16y ago

klochner 16y ago
Paul Kedrosky picks him apart here: http://paul.kedrosky.com/archives/2010/05/david_brooks_go.html http://paul.kedrosky.com/archives/2010/05/david_brooks_go.ht...
- mindcrime 16y ago
  That's hardly a "picking apart." He basically says "Yeah, your one example, which was basically a metaphor anyway, well, it's actually true, but well, it's not as neat as you think it is." Ok, and?
  - klochner 16y ago
    Fair enough - I was being a little sensationalist in expressing "relevant article", but . . . Brooks uses it as an example, but I'm not seeing the metaphor. He uses the crosswalk as an example of how people put too much faith in safety systems. The example is flawed because crosswalks are not (well-designed) safety systems given that drivers can't see them. "The world is a crosswalk . . . ?"
- hga 16y ago
  Not that I don't mind giving Brooks a good lashing, for I too "Now and then I don’t mind David Brooks’ columns at the NYT" (heck, he wrote one in the last month that was better than OK :-), but isn't this sort of the reverse pattern of posting, in that the essay on how people who don't have a clue about complexity ritually invoke it as a "thumb-sucking, multisyllabic way of saying something is hard" as Kedrosky puts it? I'd go further, pointing out e.g.: The success of D-Day/Operation Overlord and a lot of other complex stuff that no one person could understand. We have proven techniques for dealing with these sorts of things. Would that our ruling class of social engineers that includes him be a fraction as good at dealing with their complex problems. Tree Mile Island is a poor example as he uses it: "In the first place, people have trouble imagining how small failings can combine to lead to catastrophic disasters. At the Three Mile Island nuclear facility, a series of small systems happened to fail at the same time." Well, I'm sure this is true, but the people who count in this, the designers of the reactor, imagined it and e.g. included a conic spreader at the bottom of the reactor vessel to allow for a serious core melt-down. Which is the worst case that indeed happened and that the system they designed contained, as they planned. No China Syndrome in the real world (in the US, at least). I continued to analyze this essay, but it's, say, 80% bogus and not worth the trouble of debunking in detail. Let's just say that when he closes by saying we live in an "imponderably complex technical society", he's mostly revealing that Sons of Mary like him can't ponder it. Sons of Martha like us can make a lot of sense of it and we will continue to do so.
- tbrownaw 16y ago
  How does "the reason that more people die in crosswalks is really that the cars can't see the crosswalks" count as "picking apart" an argument that familiarity and perceived safety lead to complacence? If anything it would seem to corroborate it (the cars see no difference, so it must all be due to pedestrian carelessness).
  - hga 16y ago
    Because that wasn't the context in which Brooks was using the example. He started that paragraph by noting the "people have a tendency to place elaborate faith in backup systems and safety devices." In response Kedrosky pointed out that "This isn’t like a functioning safety system in an oil rig, which is what Brooks is arguing, but more like a dummy dial unconnected to anything." It does support Brooks' overall theme (most especially his previous points about Feynman's observations on NASA's safety culture (or rather lack thereof)), just not as he used it.
    - tbrownaw 16y ago
      It doesn't matter at all if the safety system has any effect. The point is entirely that safety systems (functional or not) apparently make people more reckless. Examples of this are crosswalks (not effective), and the oil rig safety thing (which in this case was also not effective). <<That’s because they have a false sense of security in crosswalks and are less likely to look both ways. >> <<Transocean official apparently tried to close off a safety debate by reminding everybody the blowout preventer would save them if something went wrong. The illusion of the safety system encouraged the crew to behave in more reckless ways. >>
      - hga 16y ago
        Ah; I think I discounted the last item you quote since as far as I know it's wrong: instead, that was an cynical aside made by a Transocean "official" after the meeting in which the BP "official" in command made a/the bad decision. As quoted by another Transocean guy, "Well, I guess that's what we have those pinchers for", apparently referring to the BOP.
whyenot 16y ago
In ecology and the environmental sciences we often talk about ecosystem managment, how to manage large complex systems that we know we don't fully understand, for instance Chesapeake Bay, or the Florida Everglades. The best approach people have come up with is to 1) do a tremendous amount of monitoring and 2) when introducing change into the system, do it in small increments and check after each introduction to make sure it is having the desired effect. The disadvantage is that to do it correctly you have to move very slowly and cautiously.
- hga 16y ago
  You're left out the single most important thing: Being willing, heck, even interested in the question of "Have you made a mistake?" and then being willing and able to reverse course. As far as I know we don't see a lot of this in ecosystem management. The best example, at least in terms of the millions it kills every year, is the jihad against DDT extending to preventing twice yearly spraying of the interiors of dwellings to limit malaria infections.
  - whyenot 16y ago
    As far as I know we don't see a lot of this in ecosystem management. The best example, at least in terms of the millions it kills every year, is the jihad against DDT extending to preventing twice yearly spraying of the interiors of dwellings to limit malaria infections. What a bizarre statement. First, that is more an issue of public health and toxicology than an issue of ecosystem level management. Second, the people involved did admit they made a mistake, and the WHO has endorsed DDT spraying in areas with high malaria incidence since 2006. Its use in such areas has increased every year since. There is continued concern that increased spraying may breed resistance and that overuse in the home may lead to its own health problems, so the current strategy is to try other options first where possible.
    - hga 16y ago
      The emphasis here is on "the jihad extending". As in while there are ecosystem level management issues with its use, these have nothing to do with vector control inside the home. Which was prevented anyway in the general hysteria, which also extends to trying to stamp out all production of DDT; from what I've read, in the free world one plant in India makes it and there are fierce efforts to shut that one down. It's very nice, I'm sure, that the WHO and company admitted their mistake ... 3 decades later and how many 10s of millions needlessly dead in what is easily one of the 20th century's biggest genocides? Color me unimpressed. This is not an example of the paradigm you introduced in your top level posting, unless you also meant "very slowly and cautiously" fixing one's mistakes. ADDED: And good luck getting anyone in the First World to donate money for its use. As I read this section in Wikipedia, that's just not happened outside of one unit of the US government starting in 2006/7 ... and does anyone want to make a bet that that's still happening?
bm98 16y ago
Firewalls are a good computing analogue to the blowout preventers and crosswalks in Brooks' article. When behind a firewall, people tend to be less careful about security. The acclimation to risk that Brooks talks about is seen in computing all the time too. The longer a security hole goes without being breached, or a software bug goes without consequence, the more likely people are going to underweigh the risk.
diego_moita 16y ago
That's typical of most journalists and pundits: this guy never designed or managed a complex system but still believes he can give important opinions about engineering and management. Engineering has done a lot to manage complexity and minimize risks (e.g.: NASA's software, GE's 6-sigma). Most accidents are really caused by irresponsible behavior, incompetence, lax regulation or a prize big enough to make you forget the risks. The problem is that this is just too simple to be newsworthy.
- jmtulloss 16y ago
  There has been a lot done, and Brooks doesn't refute that. What he's saying is that human fallibility can trump all our high-tech precautions. It's our responsibility as engineers to understand this and design systems that assume human error and negligence.
chrismealy 16y ago
Matt Zeitlin points out that Brooks is paraphrasing an old Gladwell article: http://yglesias.thinkprogress.org/archives/2010/05/gladwell-vs-brooks.php http://yglesias.thinkprogress.org/archives/2010/05/gladwell-...
- hga 16y ago
  Paraphrasing without direct attribution. Very very close to plagiarism; in addition to the Gladwell quotes Brooks could have/should have said something like "As Gladwell has noted:"

19 points by klochner 16y ago

Paul Kedrosky picks him apart here: http://paul.kedrosky.com/archives/2010/05/david_brooks_go.html http://paul.kedrosky.com/archives/2010/05/david_brooks_go.ht...

mindcrime 16y ago
That's hardly a "picking apart." He basically says "Yeah, your one example, which was basically a metaphor anyway, well, it's actually true, but well, it's not as neat as you think it is." Ok, and?
- klochner 16y ago
  Fair enough - I was being a little sensationalist in expressing "relevant article", but . . . Brooks uses it as an example, but I'm not seeing the metaphor. He uses the crosswalk as an example of how people put too much faith in safety systems. The example is flawed because crosswalks are not (well-designed) safety systems given that drivers can't see them. "The world is a crosswalk . . . ?"
hga 16y ago
Not that I don't mind giving Brooks a good lashing, for I too "Now and then I don’t mind David Brooks’ columns at the NYT" (heck, he wrote one in the last month that was better than OK :-), but isn't this sort of the reverse pattern of posting, in that the essay on how people who don't have a clue about complexity ritually invoke it as a "thumb-sucking, multisyllabic way of saying something is hard" as Kedrosky puts it? I'd go further, pointing out e.g.: The success of D-Day/Operation Overlord and a lot of other complex stuff that no one person could understand. We have proven techniques for dealing with these sorts of things. Would that our ruling class of social engineers that includes him be a fraction as good at dealing with their complex problems. Tree Mile Island is a poor example as he uses it: "In the first place, people have trouble imagining how small failings can combine to lead to catastrophic disasters. At the Three Mile Island nuclear facility, a series of small systems happened to fail at the same time." Well, I'm sure this is true, but the people who count in this, the designers of the reactor, imagined it and e.g. included a conic spreader at the bottom of the reactor vessel to allow for a serious core melt-down. Which is the worst case that indeed happened and that the system they designed contained, as they planned. No China Syndrome in the real world (in the US, at least). I continued to analyze this essay, but it's, say, 80% bogus and not worth the trouble of debunking in detail. Let's just say that when he closes by saying we live in an "imponderably complex technical society", he's mostly revealing that Sons of Mary like him can't ponder it. Sons of Martha like us can make a lot of sense of it and we will continue to do so.
tbrownaw 16y ago
How does "the reason that more people die in crosswalks is really that the cars can't see the crosswalks" count as "picking apart" an argument that familiarity and perceived safety lead to complacence? If anything it would seem to corroborate it (the cars see no difference, so it must all be due to pedestrian carelessness).
- hga 16y ago
  Because that wasn't the context in which Brooks was using the example. He started that paragraph by noting the "people have a tendency to place elaborate faith in backup systems and safety devices." In response Kedrosky pointed out that "This isn’t like a functioning safety system in an oil rig, which is what Brooks is arguing, but more like a dummy dial unconnected to anything." It does support Brooks' overall theme (most especially his previous points about Feynman's observations on NASA's safety culture (or rather lack thereof)), just not as he used it.
  - tbrownaw 16y ago
    It doesn't matter at all if the safety system has any effect. The point is entirely that safety systems (functional or not) apparently make people more reckless. Examples of this are crosswalks (not effective), and the oil rig safety thing (which in this case was also not effective). <<That’s because they have a false sense of security in crosswalks and are less likely to look both ways. >> <<Transocean official apparently tried to close off a safety debate by reminding everybody the blowout preventer would save them if something went wrong. The illusion of the safety system encouraged the crew to behave in more reckless ways. >>
    - hga 16y ago
      Ah; I think I discounted the last item you quote since as far as I know it's wrong: instead, that was an cynical aside made by a Transocean "official" after the meeting in which the BP "official" in command made a/the bad decision. As quoted by another Transocean guy, "Well, I guess that's what we have those pinchers for", apparently referring to the BOP.

whyenot 16y ago

In ecology and the environmental sciences we often talk about ecosystem managment, how to manage large complex systems that we know we don't fully understand, for instance Chesapeake Bay, or the Florida Everglades. The best approach people have come up with is to 1) do a tremendous amount of monitoring and 2) when introducing change into the system, do it in small increments and check after each introduction to make sure it is having the desired effect. The disadvantage is that to do it correctly you have to move very slowly and cautiously.

hga 16y ago
You're left out the single most important thing: Being willing, heck, even interested in the question of "Have you made a mistake?" and then being willing and able to reverse course. As far as I know we don't see a lot of this in ecosystem management. The best example, at least in terms of the millions it kills every year, is the jihad against DDT extending to preventing twice yearly spraying of the interiors of dwellings to limit malaria infections.
- whyenot 16y ago
  As far as I know we don't see a lot of this in ecosystem management. The best example, at least in terms of the millions it kills every year, is the jihad against DDT extending to preventing twice yearly spraying of the interiors of dwellings to limit malaria infections. What a bizarre statement. First, that is more an issue of public health and toxicology than an issue of ecosystem level management. Second, the people involved did admit they made a mistake, and the WHO has endorsed DDT spraying in areas with high malaria incidence since 2006. Its use in such areas has increased every year since. There is continued concern that increased spraying may breed resistance and that overuse in the home may lead to its own health problems, so the current strategy is to try other options first where possible.
  - hga 16y ago
    The emphasis here is on "the jihad extending". As in while there are ecosystem level management issues with its use, these have nothing to do with vector control inside the home. Which was prevented anyway in the general hysteria, which also extends to trying to stamp out all production of DDT; from what I've read, in the free world one plant in India makes it and there are fierce efforts to shut that one down. It's very nice, I'm sure, that the WHO and company admitted their mistake ... 3 decades later and how many 10s of millions needlessly dead in what is easily one of the 20th century's biggest genocides? Color me unimpressed. This is not an example of the paradigm you introduced in your top level posting, unless you also meant "very slowly and cautiously" fixing one's mistakes. ADDED: And good luck getting anyone in the First World to donate money for its use. As I read this section in Wikipedia, that's just not happened outside of one unit of the US government starting in 2006/7 ... and does anyone want to make a bet that that's still happening?

bm98 16y ago

Firewalls are a good computing analogue to the blowout preventers and crosswalks in Brooks' article. When behind a firewall, people tend to be less careful about security. The acclimation to risk that Brooks talks about is seen in computing all the time too. The longer a security hole goes without being breached, or a software bug goes without consequence, the more likely people are going to underweigh the risk.

diego_moita 16y ago

That's typical of most journalists and pundits: this guy never designed or managed a complex system but still believes he can give important opinions about engineering and management. Engineering has done a lot to manage complexity and minimize risks (e.g.: NASA's software, GE's 6-sigma). Most accidents are really caused by irresponsible behavior, incompetence, lax regulation or a prize big enough to make you forget the risks. The problem is that this is just too simple to be newsworthy.

jmtulloss 16y ago
There has been a lot done, and Brooks doesn't refute that. What he's saying is that human fallibility can trump all our high-tech precautions. It's our responsibility as engineers to understand this and design systems that assume human error and negligence.

chrismealy 16y ago

Matt Zeitlin points out that Brooks is paraphrasing an old Gladwell article: http://yglesias.thinkprogress.org/archives/2010/05/gladwell-vs-brooks.php http://yglesias.thinkprogress.org/archives/2010/05/gladwell-...

hga 16y ago
Paraphrasing without direct attribution. Very very close to plagiarism; in addition to the Gladwell quotes Brooks could have/should have said something like "As Gladwell has noted:"