Starting Over

I have a favorite thought exercise, especially when thinking about the sort of complex, dynamic systems that are interesting but difficult to write about: the health-care system, e.g., or education, politics, energy consumption, finance, cancer research, etc.

One natural way to approach such systems is to take note of what inputs and outputs already exist and then, isolating them, try to measure the success of each one. If you’re trying to assess U.S. public grade-school education, you can look at the great many metrics that represent the system — dollars per pupil, class size, incoming IQ’s, mode of instruction, length of day, etc. — and try to come up with ideas to make the entire system better by changing one or two or three inputs. I watched a lot of the Democratic debate last night, and saw many, many examples of such thinking.

But the thought exercise I’m talking about comes at things from a different angle. Complex systems are complex in part because of the way they evolve; that is just the nature of the beast. And so, while it’s important to understand why a particular system evolved as it did — to understand the financial, political, social, scientific, and psychological forces that shaped the way, for instance, that cancer is treated in this country — I find it useful to ask an entirely different question: if we were making this system up from scratch today, what would it look like?

I find this thought experiment particularly useful when interviewing people. Let’s say that someone knows an awful lot about medicine or education or energy. She probably has a huge storehouse of knowledge, and maybe even some strong opinions, but in the course of a typical day, she’s required to fiddle around the edges of the complex system, making very minor changes that will seldom have a big effect. But when you ask a person like this what she’d do if she could build the system from scratch — well, that tends to produce some interesting answers, and may shed light on systemic failings that may otherwise go unspoken.

Given my fondness for this approach, I was tickled to read this New York Times article by Jason Pontin about the biotech company Amyris. Using a technology called metabolic engineering, Amyris “has almost finished developing a cheap cure for malaria that could save the lives of millions,” Pontin writes. Next up: “new biofuels that may help save the planet.” Here is the passage that caught my eye:

Amyris chose to ask something more basic and more interesting: What would perfect fuels look like if they were designed from scratch? The start-up decided to concentrate on the second stage of creating a biofuel: fermenting sugars into fuel.

It is well worth reading the entire piece, for I haven’t done a good job here conveying the ideas brewing at Amyris. But my point is a simple one: just because systems evolve over time in a complicated, random, even contradictory fashion, there is no reason to think about solutions in that same way.


lermit

I find your piece as motivating as usual. The economics on this all get a bit confusing, now that it's summer.

Yet, it gets you close to a fundamental truth: garbage in, garbage out (GIGO). Thanks for sharing,

.lermit (your cousin? yo mamma!)

egretman

“new biofuels that may help save the planet.”

How can any carbon based fuel save the planet? It might save my SUV. Not sure how it's going to save the planet. You mean it might save our economies or our way of life?

Or will it lead to continued distruction of our planet? Why would we assume these guys have any clue what the planet needs? Does anyone else?

The headline could have been,

Amyris, new biofuels may help destroy the planet.

Andrew

I recognize the freshness that comes from the "starting over" approach when looking at entrenched systems... I find myself much more cautious, however, about disregarding the careful evolution of problems that are so complex... If starting from a clean slate did not have significant hurdles in terms of cost, acceptance by the public, distribution, etc. it would have been done already. You see this in the case of biofuels/energy infrastructure, healthcare systems, etc.

In short, looking at these systems with a fresh face may indeed clarify the goals we should be striving for, but it does not provide a short-cut to the end-all solution.

bbeam

-- replying to 3, Andrew

I think it is frequently the case that people working inside the prevailing paradigm get much easier buy-in (both literal and figurative) from their fields. Also, incremental change is relatively easy compared to scrap it and rebuild.

That said, I would prefer not to be the guinea pig for a brand new medical system, for instance. (However, I might be willing to do so for a new health INSURANCE system; I would probably prefer the devil I don't know ....)

tgalvin

Starting Over,

Here's an interesting thought experiment:

Q. What would the perfect Bubble look like?

While many processes are inefficient and ultimately damaging when viewed over a long term there is a reason why they evolved the way they did; it was probably the simplest solution that would satisfy ALL the important conditions.

Looking at a problem from a single perspective is easy, complex problems (with convoluted solutions that don't make sense from a single perspective) are a little bit harder. What is frustrating from an economist's perspective (speaking from my own personal experience) is that they have to deal with political considerations.

There are no fresh starts, this dismal science is about optimizing with constraints, not imagining them away.

tompaper

Stephen, Your post is spot on. Folks like Deming and the six-sigma gurus use the terms "systemic" change and "structural" change to indicate the difference between changes which are incremental and changes which are usually wholesale or epic. The automobile, the airplane, the personal computer were all structural changes. When you have structurally different systems, you can get significantly different results. The United States of America was a system developed by a group of people who fashioned government from scratch. Keeping states semi-independent was intended to facilitate experimentation, although I am not aware of a system which exists to tap into the innovation occuring within our fifty states. In education today, the charter schools are experimenting with structurally different systems and getting, not surprisingly, very different and usually better results than the current system. In publishing, I think we are witnessing structural changes and the traditional newspapers are seeing their traditional models destroyed. Craigslist.com is a great example of structural change dealing significant pain to traditional newspapers. Your question for publishing might be, "If we could build the news business over again today from scratch, what would it look like?" I think the answer includes people wanting news that better fits whatever they arelooking for...a la blogs, rss feeds and many other cool things coming out today. Very specialized. Along the lines of the 1 to 1 future, a book written a while back. Regards, Tom Paper www.data360.org

Read more...

Mike Lee

Excellent entry! It also made me think of Clay Shirky's entry "The (Bayesian) Advantage of Youth":
http://many.corante.com/archives/2007/05/19/the_bayesian_advantage_of_youth.php

The entry discusses a topic Fred Wilson (a VC in Silicon Valley) blogged about - that the prime age for entrepreneurs is in their 30s.

Shirky agrees. He posits that younger entrepreneurs' relative lack of knowledge is an advantage to older, more experienced entrepreneurs. He writes:

"I'm old enough to know a lot of things, just from life experience. I know that music comes from stores. I know that you have to try on pants before you buy them. I know that newspapers are where you get your political news and how you look for a job. ...

In the last 15 years or so, I've had to unlearn every one of those things and a million others. This makes me a not-bad analyst, because I have to explain new technology to myself first — I'm too old to understand it natively. But it makes me a lousy entrepreneur."

So essentially, younger entrepreneurs are able to start over and build a solution from scratch.

Read more...

Izeas GT

I liked this post a lot. It resembles to a great degree the question I keep thinking about, which could almost be called the ultimate generalization of yours: "What should be?"

Of course, it has to be tempered with "What can be?", but I still can't think of a much better question to start with for...well, anything.

ninesuns

I liked this post as well. Having worked in the Australian healthcare system, I have seen how we end up solving the wrong problems. Due to political and media pressures here, we are fixated on finding more doctors and hospital beds. When what we should be asking is why are we so sick?

see

Would have been nice if they'd named the chemical that "could be fermented from sugars derived from any biomass . . . have an energy content higher than ethanol[,] be usable in today's automobile, diesel and jet engines[, and] were insoluble in water, and hence could be transported through the same pipelines that now move oil."

I mean, if I had to guess, I'd say they're talking about butanol, which meets all those criteria. The organism that is usually used to produce butanol from biomass is poisoned by butanol rather faster than yeast is poisoned by ethanol -- something with more tolerance would really improve its viability.

mathking

Part of the point (I think) that Dubner was making is not that we necessarily can or should start from scratch when examining a complex syste. It is the idea that thinking about what you would do if you were starting from scratch can be a useful analytical technique.

Now on to quibble with a comment from another poster: "the charter schools are experimenting with structurally different systems and getting, not surprisingly, very different and usually better results than the current system." For full disclosure, I am a public school teacher. The data on charter schools is very preliminary, but in every state which has started to do comprehensive analysis of charter schools, there is no data to justify the "usually better" part of the above comments. There are certainly some charter schools that have had good results, but as a whole charter schools are not getting better (or even equal) educational results compared to public schools. Furthermore, in many states, they fail distressingly often due to fiscal mismanagement and high rates of staff turnover.

I am not saying that charter schools may not become an important, effective part of the educational system in the U.S. And I am not opposed to them. I think we should continue experimenting with them. But we need to recognize that they are largely experimental. The assertion that charter schools (and private schools) are better than public schools is not born out by the data.

To bring this back to Dubner's comments, I think that one of the primary benefits of charter schools is the freedom to experiment with wholesale redesign in education. But when you do experiments, you get a lot of failures.

Read more...

unacoder

in my business, "refactoring" is the systematic improvement of existing designs without affecting external behavior. your post puts me in mind of one of the most powerful techniques in IT for simplification of complex systems. the result is a removal of duplicated functions, now isolated to a single implementation. that implementation may now be optimized, moved, or stricken. to my mind, the same techniques can be applied to science, government, or any other area of human thought.

lermit

I find your piece as motivating as usual. The economics on this all get a bit confusing, now that it's summer.

Yet, it gets you close to a fundamental truth: garbage in, garbage out (GIGO). Thanks for sharing,

.lermit (your cousin? yo mamma!)

egretman

"new biofuels that may help save the planet."

How can any carbon based fuel save the planet? It might save my SUV. Not sure how it's going to save the planet. You mean it might save our economies or our way of life?

Or will it lead to continued distruction of our planet? Why would we assume these guys have any clue what the planet needs? Does anyone else?

The headline could have been,

Amyris, new biofuels may help destroy the planet.

Andrew

I recognize the freshness that comes from the "starting over" approach when looking at entrenched systems... I find myself much more cautious, however, about disregarding the careful evolution of problems that are so complex... If starting from a clean slate did not have significant hurdles in terms of cost, acceptance by the public, distribution, etc. it would have been done already. You see this in the case of biofuels/energy infrastructure, healthcare systems, etc.

In short, looking at these systems with a fresh face may indeed clarify the goals we should be striving for, but it does not provide a short-cut to the end-all solution.

bbeam

-- replying to 3, Andrew

I think it is frequently the case that people working inside the prevailing paradigm get much easier buy-in (both literal and figurative) from their fields. Also, incremental change is relatively easy compared to scrap it and rebuild.

That said, I would prefer not to be the guinea pig for a brand new medical system, for instance. (However, I might be willing to do so for a new health INSURANCE system; I would probably prefer the devil I don't know ....)

tgalvin

Starting Over,

Here's an interesting thought experiment:

Q. What would the perfect Bubble look like?

While many processes are inefficient and ultimately damaging when viewed over a long term there is a reason why they evolved the way they did; it was probably the simplest solution that would satisfy ALL the important conditions.

Looking at a problem from a single perspective is easy, complex problems (with convoluted solutions that don't make sense from a single perspective) are a little bit harder. What is frustrating from an economist's perspective (speaking from my own personal experience) is that they have to deal with political considerations.

There are no fresh starts, this dismal science is about optimizing with constraints, not imagining them away.

tompaper

Stephen, Your post is spot on. Folks like Deming and the six-sigma gurus use the terms "systemic" change and "structural" change to indicate the difference between changes which are incremental and changes which are usually wholesale or epic. The automobile, the airplane, the personal computer were all structural changes. When you have structurally different systems, you can get significantly different results. The United States of America was a system developed by a group of people who fashioned government from scratch. Keeping states semi-independent was intended to facilitate experimentation, although I am not aware of a system which exists to tap into the innovation occuring within our fifty states. In education today, the charter schools are experimenting with structurally different systems and getting, not surprisingly, very different and usually better results than the current system. In publishing, I think we are witnessing structural changes and the traditional newspapers are seeing their traditional models destroyed. Craigslist.com is a great example of structural change dealing significant pain to traditional newspapers. Your question for publishing might be, "If we could build the news business over again today from scratch, what would it look like?" I think the answer includes people wanting news that better fits whatever they arelooking for...a la blogs, rss feeds and many other cool things coming out today. Very specialized. Along the lines of the 1 to 1 future, a book written a while back. Regards, Tom Paper www.data360.org

Read more...

Mike Lee

Excellent entry! It also made me think of Clay Shirky's entry "The (Bayesian) Advantage of Youth":
http://many.corante.com/archives/2007/05/19/the_bayesian_advantage_of_youth.php

The entry discusses a topic Fred Wilson (a VC in Silicon Valley) blogged about - that the prime age for entrepreneurs is in their 30s.

Shirky agrees. He posits that younger entrepreneurs' relative lack of knowledge is an advantage to older, more experienced entrepreneurs. He writes:

"I'm old enough to know a lot of things, just from life experience. I know that music comes from stores. I know that you have to try on pants before you buy them. I know that newspapers are where you get your political news and how you look for a job. ...

In the last 15 years or so, I've had to unlearn every one of those things and a million others. This makes me a not-bad analyst, because I have to explain new technology to myself first - I'm too old to understand it natively. But it makes me a lousy entrepreneur."

So essentially, younger entrepreneurs are able to start over and build a solution from scratch.

Read more...

Izeas GT

I liked this post a lot. It resembles to a great degree the question I keep thinking about, which could almost be called the ultimate generalization of yours: "What should be?"

Of course, it has to be tempered with "What can be?", but I still can't think of a much better question to start with for...well, anything.