She's probably in the minority, but this reviewer winced a little on seeing that the most prominent accolade for Tim Harford on the cover of "Adapt" was "Britain's Malcolm Gladwell". Because her impression of the latter writer is that of a genius, unimaginatively retrofitting the same thesis and analysis format onto different (and successively more pedestrian) ideas. Actually, Harford's two previous books (which she owns) are both pop-economics which attempt to make the dismal science more relevant in everyday social spheres and the second was probably an effort to squeeze more juice out of the first. It did, too. This reviewer was happy for more.
But Adapt is a change of direction, (perhaps mimicking its own content?). Nominally a road-map for success in solving problems or innovating, the author sets out thesis elements chapter by chapter. With this format, he also measures out the size of the ground that the book's thesis is intended to cover and apply to. And it is ambitious: inventions (many of which--nuclear fusion, HIV vaccine--are late), business success, military strategy, development economics, and finally you the reader. But Harford's economic/journalism background prints-through such that the emphasis is noticeably on, well, making money.
The first two thesis elements are pulled directly from biological evolution: random variation and unfettered selection/rejection. Two more are woven in later: complexity (or rather, its mitigation) and "decoupling". These are more to do with risk control, or preventing catastrophic failure where it can't be blithely waved away. But in front of this is an emphasis on preparing for, welcoming, and navigating multiple errors. This is what Darwinian evolution does, harshly and dispassionately. It is also what the software business does, because "failure is for free" (experimentation costs close to nothing). But many breakthroughs (fusion again) have very high hurdles to develop. And that's before the investment of creators', CEOs or totalitarian dictators' egos are counted. The result is a pathological failure to experiment ("randomly mutate") and a drastically deformed decision basis (in other words, unnatural selection).
So, the eventual crash of the USSR is diagnosed as due to lack of either of these attributes, rather than having much to do with private profit motive (though they are presumably linked). The unfortunate fate of Peter Palchinsky--appointed by the Tsars to report on inefficiency in the (dreadful) mines, then later executed after too much correct diagnosis and recommendation--attests to this while at the same time corroborating the message.
In the business world, barriers to random variation seem similarly high and mostly due to the hierarchical replication of central planning which is not only still widely believed in as a success formula (astoundingly), but also hard to dismantle because of the cognitive dissonance--to almost everyone--that this would cause. Harford helpfully points out case studies (Timpson and WL Gore; both decades older than Google) that are many strides ahead in having a business model for experimentation and randomised peer selection. One of the most important desiderata for this is the creation and protection of safe-havens in business (beautifully likened to the Galapagos Islands as incubators of speciation/evolution). Such safe havens are comfort zones where ideas can proliferate and failures blessed as steps forward not back--they rule out blind alleys after all. This echoes but does not mention Susan Cain's "Quiet" in some respects.
The difficulties for development strategies for poor countries are similar. Ideas may abound, but tend to be assessed from afar, which results in sticking with losers and accordingly not looking hard enough for winners (the "play-pump" example is illuminating). Also, there is a moral opposition to controlled (randomised and blinded) testing (of the things that are dreamed up to help the poor), but as the author points out, uncontrolled experiments are surely ethically inferior, teaching us little or nothing. The advent of "randomistas" to the development sphere--such as Esther Duflo and Sendil Mullainathan--seems encouraging.
The author devotes an apparently paradoxical chapter to the financial crisis of last decade, in which the two ingredients of variation and selection look to have been the root cause of the world's near meltdown--far from being the keys to survival and economic health. Why is this? Firstly because of complexity, which substantially masks the ability to select (the same is true with climate change--this reviewer found chapter five very salutory). And secondly because of close coupling--the same phenomenon by which a world record attempt at toppling dominoes really should take care not to stand all the dominoes up without safety-breaks in place that would, say, prevent a butterfly landing on one from prematurely decking all of them. The proposals for de-coupling seemed a little unoriginal and lukewarm to this reviewer, however.
And finally, she wasn't that enamoured with the final chapter either--about how to adapt oneself and not fall into the many retardants of sunk cost, cognitive dissonance, and ego investment. Since this is the shortest of eight, Harford may have been struggling here too. And most of it tells of Twyla Tharp, a choreographer who brought a ballet-cum-musical to the pre-Broadway stage, got ruthlessly panned, but had the necessaries to comprehensively fix up the performance in response and eventually produce a smash hit. It was hard to see what could be easily copied here.
This doesn't matter. Overall the book is a fascination, and this review leaves much out. Next one soon please.
But Adapt is a change of direction, (perhaps mimicking its own content?). Nominally a road-map for success in solving problems or innovating, the author sets out thesis elements chapter by chapter. With this format, he also measures out the size of the ground that the book's thesis is intended to cover and apply to. And it is ambitious: inventions (many of which--nuclear fusion, HIV vaccine--are late), business success, military strategy, development economics, and finally you the reader. But Harford's economic/journalism background prints-through such that the emphasis is noticeably on, well, making money.
The first two thesis elements are pulled directly from biological evolution: random variation and unfettered selection/rejection. Two more are woven in later: complexity (or rather, its mitigation) and "decoupling". These are more to do with risk control, or preventing catastrophic failure where it can't be blithely waved away. But in front of this is an emphasis on preparing for, welcoming, and navigating multiple errors. This is what Darwinian evolution does, harshly and dispassionately. It is also what the software business does, because "failure is for free" (experimentation costs close to nothing). But many breakthroughs (fusion again) have very high hurdles to develop. And that's before the investment of creators', CEOs or totalitarian dictators' egos are counted. The result is a pathological failure to experiment ("randomly mutate") and a drastically deformed decision basis (in other words, unnatural selection).
So, the eventual crash of the USSR is diagnosed as due to lack of either of these attributes, rather than having much to do with private profit motive (though they are presumably linked). The unfortunate fate of Peter Palchinsky--appointed by the Tsars to report on inefficiency in the (dreadful) mines, then later executed after too much correct diagnosis and recommendation--attests to this while at the same time corroborating the message.
In the business world, barriers to random variation seem similarly high and mostly due to the hierarchical replication of central planning which is not only still widely believed in as a success formula (astoundingly), but also hard to dismantle because of the cognitive dissonance--to almost everyone--that this would cause. Harford helpfully points out case studies (Timpson and WL Gore; both decades older than Google) that are many strides ahead in having a business model for experimentation and randomised peer selection. One of the most important desiderata for this is the creation and protection of safe-havens in business (beautifully likened to the Galapagos Islands as incubators of speciation/evolution). Such safe havens are comfort zones where ideas can proliferate and failures blessed as steps forward not back--they rule out blind alleys after all. This echoes but does not mention Susan Cain's "Quiet" in some respects.
The difficulties for development strategies for poor countries are similar. Ideas may abound, but tend to be assessed from afar, which results in sticking with losers and accordingly not looking hard enough for winners (the "play-pump" example is illuminating). Also, there is a moral opposition to controlled (randomised and blinded) testing (of the things that are dreamed up to help the poor), but as the author points out, uncontrolled experiments are surely ethically inferior, teaching us little or nothing. The advent of "randomistas" to the development sphere--such as Esther Duflo and Sendil Mullainathan--seems encouraging.
The author devotes an apparently paradoxical chapter to the financial crisis of last decade, in which the two ingredients of variation and selection look to have been the root cause of the world's near meltdown--far from being the keys to survival and economic health. Why is this? Firstly because of complexity, which substantially masks the ability to select (the same is true with climate change--this reviewer found chapter five very salutory). And secondly because of close coupling--the same phenomenon by which a world record attempt at toppling dominoes really should take care not to stand all the dominoes up without safety-breaks in place that would, say, prevent a butterfly landing on one from prematurely decking all of them. The proposals for de-coupling seemed a little unoriginal and lukewarm to this reviewer, however.
And finally, she wasn't that enamoured with the final chapter either--about how to adapt oneself and not fall into the many retardants of sunk cost, cognitive dissonance, and ego investment. Since this is the shortest of eight, Harford may have been struggling here too. And most of it tells of Twyla Tharp, a choreographer who brought a ballet-cum-musical to the pre-Broadway stage, got ruthlessly panned, but had the necessaries to comprehensively fix up the performance in response and eventually produce a smash hit. It was hard to see what could be easily copied here.
This doesn't matter. Overall the book is a fascination, and this review leaves much out. Next one soon please.
via JREF Forum http://ift.tt/1jRSVHV
Aucun commentaire:
Enregistrer un commentaire