Friday, March 28, 2014

The Pattern on the Stone

Daniel Hillis has done a lot of cool AI stuff, so I picked up his book about how computers work used off Amazon.  Definitely worth $0.01 plus shipping, but not exactly a revelation if you've done any prior reading or thinking about the topic.

The book does a pretty good and pretty substantial job of explaining how a modern computer works from the ground up.  I say "substantial" because Hillis really lingers on some of the very basic building blocks of the physical implementation of Boolean algebra and the construction of a finite state machine, so that by the time you get to him saying that a computer is a finite state machine hooked up to a memory and coupled to some inputs and outputs, these words will all  end up meaning something concrete to you if they did not at the beginning.  

If those words all mean something to you already, you might consider skipping the book.   Though I can't say I learned a ton, I still think I cemented a few concepts like machine language that were a little nebulous before, and the book is well-written and a quick read, hence enjoyable.  




Tuesday, March 18, 2014

A message from our sponsor

A Will for the Woods:



Is reminiscent of Kurosawa's Ikiru in terms of emotional impact.



Spoiler alert: it's not a happy ending (or is it?):


Monday, March 17, 2014

Turing's Cathedral

So overall, I'd say that George Dyson's Turing Cathedral gets maybe 3 out of 5 stars.

As a history of some of the first digital computers ever built it's interesting, if somewhat exhaustively detailed.  Contrary to what the title might imply, the book is mostly about John Von Neumann's work at the Princeton Institute for Advanced study and out in Los Alamos.  Turing's invention of computational cryptography at Bletchley Park is mentioned in passing.  Many a tale of mathematical prowess and long hours worked during the camaraderie of the war is told.  One thing made clear is the way early computation was so entwined with the military.  I guess this isn't a huge surprise, nor has it really changed, but it's still interesting to me to see the direction science and mathematics was pushed in by external forces.  Without the urgent need to calculate shockwaves from a bomb blast, the computer might never have existed (or was it inevitable?).  

The book has another, more philosophical side though.  Dyson wants to construe the digital world as another great kingdom of life, complete with evolving and self-reproducing creatures.  The construction of the digital computer is then like the Big Bang of this new universe.  I am totally sympathetic to this idea, and one thing the book does make clear with its ridiculously detailed combing of their notes and archives is how completely Von Neumann, Ulam, Turing, et. al. were on the bus.  Hell, they built the fucking bus, and they had pretty advanced ideas about where it was headed.  Unfortunately, if you are not on the bus this book will not get you there, and if you were already on the bus, there's not sufficient detail (understanding?) to put you in the driver's seat.  There's quite a lot of suggestive analogy thrown around.  And complicated mathematical concepts are brushed over quickly; even if you are somewhat familiar with Godel, Church, and self-reproducing cellular automata you may find that you're not sure you understood the significance of that last paragraph.  You may read it again and still think it's kinda fast and loose, or maybe you just don't know enough about Turing's oracle machines?  

Which sums up the big problem with the book -- who was this written for?  Historians?  Fine, but then why the sci-fi wanderings that start halfway through and derail the story?  Mathematicians?  Hardly.  There's not enough substance and explanation of what are very difficult concepts.  Ray Kurzweil perhaps?  We'd just have to ask his exo-skeleton I guess.  Having too many audiences means that no one is going to go home happy.

The book did mention one curiosity I'd like to find out about.  Hans Alfvén won the 1970 Nobel prize in physics for his work on magnetohdrodynamics (who knew, right?). Turns out he also wrote a sci-fi novel under the pseudonym Olof Johannesson entitled The Tale of the Big Computer: A Vision.  And it sounds like a pretty good one, if Dyson description and excerpts are anything to go by.  Doesn't appear to be in print though?  


Wednesday, March 5, 2014

Schumpeter Saved My Blog

The vagaries of intellectual history are curious. 

Deleuze's book on cinema makes great use of the theory of signs of Charles Sanders Pierce.  Pierce I know vaguely as some marginal American philosopher.  Not sure what exactly he had to say.  But reading the secondary sources Deleuze refers to has proven to be a reliable source of useful new material for a good 15 years now, so I'm curious.  Five minutes into the Wikipedia entry I find myself wondering what justice this world holds if this guy remains marginal; beyond inventing transfinite arithmetic before Cantor, laying out the first minimal logical foundations of the modern computer, axiomatizing the natural numbers and set theory ahead of Dedekind, Peano, and Zermelo, and influencing everyone from Russell and Whitehead to Popper and Dewey, what, I ask, have you done for me lately?  

Joseph Schumpeter is another of these marginal figures.  P.Krug even uses him as a whipping boy.  I mean, there's not even an epic rap battle for this guy.  About all the has to show for himself is having a column in The Economist named in his honor, a dubious distinction at best.  Yet I just read his Capitalism, Socialism, and Democracy, and found it one of the most sophisticated pieces on political economics I've ever come across.  Not to be too incendiary, but the sweep and insight of his thought makes Keynes look like a green-eye-shaded bureaucrat by comparison.  

So I wanted to describe the book a bit and clarify exactly why I found it, as Mancur Olson said, "profound but not prescient".  

The first marvelous thing about Schumpeter is the way was able to write a book that essentially predicts the inevitable downfall of capitalism and rise of socialism while remaining so even-handed in its treatment of both systems.  In many ways you can't really tell where his own personal sentiments lie in this debate, though given some of his other writing I think we can infer he looked on socialism with some foreboding.  However, Schumpeter would be the first to suggest that his opinion doesn't much matter; he is simply trying to do the best job he can of dispassionately analyzing the functioning of an economic machine -- a machine not terribly concerned with the opinions of Herr Doktor, thank you very much.  

This lack of prejudice is incredibly valuable when it comes to really understanding the long term trajectory of our society.  Admittedly, bracketing your preconceptions and just looking at some fresh data is useful for making just about any decision in life, but it seems especially essential when the forces you are analyzing work over the course of generations, not minutes.  When you are concerned with their implications in 50 years, things are not necessarily what they seem at first glance.  And this is the time scale that Schumpeter is interested in.  He wants to know how we will be making decisions about what and how much to produce far after he himself is gone.   To try and reach that length of vision, you have to back away from so much short term noise that you forfeit anything resembling conventional political opinion or advocacy.  What you get in exchange is an almost geologic or evolutionary vision of the underground grinding that animates a social machine.  This is the real value of Schumpeter; he thinks about capitalism itself as an evolutionary system, which immediately gets you started thinking about what it might evolve into.

Now, you might object that people interested in time scales measure in half century increments are not really economists.  And it's true; unfortunately, this has not been considered economics for the past 50 years.  So I suppose we should resurrect the term "political economist" for Schumpeter, or simply call him a sociologist or a historian and get on with it.  It's a shame, though, that his thinking about the evolution of the rules by which the economic machine functions needs to be relegated to a different discipline than economics.  This is roughly similar to insisting that molecular biologists not talk about how the animals they studied evolved because this is a completely different discipline.  It's the kind of specialization that starts innocently enough, but ends frozen in absurdities.  I feel like something even worse has happened in economics. The more I read, the more I become convinced that economics was much more interesting and useful before they turned it into a mathematical "science".  Meanwhile, broader questions about the history and evolution of our productive system have been relegated not just to an entirely different discipline, but an inferior and "unscientific" one.

There is a long side note here about how this transition seems to have started when they mathematized Keynes, and how this is interesting because when we look back in 100 years we will (hopefully) see that Keynes was actually a very short-term thinker who suggested short-term remedies to what he saw as short-term problems in what was overall a stable system he didn't want to call into question.  Keynes, in other words, is a deeply conservative figure.  This contrasts strangely with the current political axes of Keynesians versus "Austerians", and maybe makes you wonder whether there isn't a time scale dimension to that debate which I've never heard much discussion of.  Even a notoriously "liberal" economist like P.Krug seems to be basically defending the status quo capitalism of the 50's and 60's.   But I digress.  

Let me just summarize the 5 main sections of the book.

Part 1)  The best, most generous and yet fair-minded discussion of Marx I think I have ever seen.

Unlike most any other economist, Schumpeter tries to understand why Marx has had such enduring influence despite the fact that many of his purely economic ideas seem to fall somewhere between mis-guided and just plain wrong.  Again, there's that obsession with mechanism in place of a preference for opining on the merits of something.   Looking beyond the obvious appeal of Marx the Prophet to the religion of socialism, Schumpeter attributes this enduring influence to two other factors in Marx's theory.  

First, the theory provides at least interesting explanations based on fundamental principles for big historical trends (eg. imperialism) that no one else was really even attempting to explain. Even if the principles seem wrong and the explanations don't always hold up, they are thought provoking, and at least they take up the challenge of explaining big stuff in the real world.  Never underestimate the persuasive power of a big theory encapsulated in a small slogan. 

Second, as both sociologist and economist Marx was the first to try and understand our world as the mechanistic outcome of a determined historical process.  This is another one of these ways of looking at the world that seems so completely obvious now that it would be unremarkable if you didn't consider how new this was at the time.  If we call Marx the father of the economic or materialist interpretation of history, and leave this theory at its bare bones -- asserting that the way we produce stuff influences the structure of our society and the way our society is structured influences the long-term changes in the way we produce stuff  -- we are all Marxists now (or should be).  To set history in motion like that, to give it an materialist engine and make the technology of production the carburetor, is a deep and abiding shift in the way we try to explain our world.

Part 2)  A completely different theory of the way capitalism actually works in practice that is emphatically not based on the idea of equilibrium.  

Schumpeter's creative destruction theory of capitalism has significantly more depth than the popularized dust jacket version which is usually associated with it, and places an emphasis on the actions of real-world entrepreneurs that I have never seen in an economist.  The crucial thing to realize is that he believes there is no way to understand capitalism as a static system; it is inherently dynamic.  Capitalism, taken as primarily a social organization, is a system for evolving the means of production.  The changes in technology that lead from one industrial revolution to another are not, in this theory, external events that happen to capitalism the way that meteor strikes happen to dinosaurs.  These are internal shifts within a capitalist system that is in fact set up precisely in order to produce these sorts of revolutions.  Creative destruction is the bread and butter of capitalism.

It's really hard to misunderestimate how different the view of our economy is when you adopt this vantage point.  It cuts through so many important questions in ways completely orthogonal to current political axes.  Schumpeter goes on to outline a few of these as he sums up the failure of classical economics to adequately describe real world capitalism.  Typical, current, common sense views about markets and the invisible hand of free competition, about the role of oligopolies and monopolies, about industrial consolidation, and about the relationship between equilibrium and efficiency all need to go by the wayside once you adopt a dynamic view of economics.  Suddenly, questions that had ready-made answers (free trade benefits everyone! monopolies are evil!) seem a lot more complex and case dependent.   

In a way, this shift in perspective is deeply troubling, mainly because you know that the academic and policy ship is still steered by people who learned that the earth is flat.  In exchange, though, for the certainty that all is for the best in the best of all possible perfectly competitive capitalist worlds, you get to see the economy a lot more like I see it professionally, from the ground up, like an ant roaming through the jungle.  Schumpeter is not the first but the only macroeconomist I have come across who seems to be looking at the same world I recognize from investing, that is, the real world of on-the-ground capitalism as it happens every day.  In this world the market is clearly not efficient and competition is almost never perfect, and yet it is still pretty fucking hard to come up with a way to just sit back and watch the money roll in.  And it is this is the same inefficient world that has been responsible for the incredible increase in our standard of living over the past 200 years, all despite its failure to resemble anything in an economics textbook over any part of this period.  

In other words, Schumpeter solves a great conundrum.  When you simply observe it in action every day, you realize that capitalism doesn't look anything like a free market.  Yet the outcomes it has produced over the long term seem to be basically what they would have been if it were.  From the usual perspective, this coincidence looks like a fortunate miracle.  Schumpeter doesn't take this miracle for granted but goes about deducing it.  His theory has the same kind of beauty as General Relativity -- it is a radically different theory of why things fall, but it still tells you that terrestrial apples drop on heads at roughly 32 ft/sec/sec.

Armed with his theory of the way a capitalist economy evolves, Schumpeter can then go on to make some predictions about how it will continue to evolve.  Basically, he concludes that there isn't much reason to expect any of the economic forces that have driven past growth to falter.  This may sound conservative to the point of vapid now, but in 1942 it definitely was not.  Remember, he is writing this during WWII, which followed hot on the heels of the Great Depression.  Sanguine assurance that in the long term the economy would return to its pre-1929 trend was not in vogue.  

You might expect, at this point, that Schumpeter will go on to predict the end of history and the last man, the metaphysical triumph of capitalism.  In fact, he goes on to hypothesize the gradual collapse of capitalism for political and social reasons despite all its purely economic success.  Indeed, he believes capitalism's victories are precisely what will tip it into decline; the success, rather than the failure of capitalism (as Marx would have it) will be what causes the system to evolve into socialism.  I'll quickly summarize what he sees as the specific forces that will erode capitalism from within, but first a quick quote that gives you the sense of Schumpeter's overall method.

Things economic and social move by their own momentum and the ensuing situations compel individuals and groups to behave in certain ways whatever they may wish to do -- not indeed by destroying their freedom of choice but by shaping the choosing mentalities and by narrowing the list of possibilities from which to choose.  If this is the quintessence of Marxism then we all of us have got to be Marxists.

Capitalism shapes its society to its own detriment in several ways.

1) Capitalism bureaucratizes progress.  Private property and individual/family entrepreneurship create an incentive system that leads to incessant revolutions in production.  These revolutions are the engine of the progressive increase in our standard of living.  Over time, society begins to adapt to this very pattern of successive revolution and the search for the next innovation becomes the province of a well oiled machine run by large corporations.  In other words, capitalism tends to automate the idea of progress, which means it doesn't need capitalists anymore.  Basically, the singularity already happened.  Welcome to the future!  Now that we're here though, we should quit paying Mark Zuckerberg for it.  

2) Capitalists makes lousy politicians.  The bourgeois grew up in the shadow of feudal rulers and succeeded by sticking to its knitting and letting someone else deal with the problems of ruling; these two classes are symbiotic.  Eventually, capitalism undermines the authority of these rulers, and yet doesn't provide the material to replace them.  

3) Capitalism evolves towards big business capitalism.  The economic logic of consolidation and economy of scale appears again and again despite the state occasionally pushing it back.  Large business are nothing if not giant bureaucracies.  They are typically run by professional managers who seem themselves as employees, not founders or entrepreneurs.  And they are owned by thousands of poorly informed shareholders, most all of whom have very little at stake in any particular company.  If concentrated individual property created the incentive to go out and innovate and build an empire, this dispersion of ownership and control takes all the risk and reward out of property.  The system of incentives completely changes in a way that is sure to favor a status quo gradualism; revolutions are strangled before their birth.  I particularly love this point because I was personally prescient enough to come up with a version of it no more than 60 years after Schumpeter.

4) As capitalism raises standards of living it creates the possibility for universal education, and with it a class of over-educated under-employed intellectuals who become critical of the system (this is similar to Nozick's thesis on why intellectual's oppose capitalism).  These folks are often in a position to form public opinion and they create an attitude of hostility towards capitalism in general.

5) Capitalism breeds habits of rational individualist calculation in each bourgeois mind.  Eventually, people start to get comfortable and wonder why they are working so hard.  They start to think that there's no point in founding a business that will create a family dynasty, and they start to wonder whether they would like to have a large family to begin with.  

As you can see, all these explanations of decline involve extrapolating what happens when capitalism is working well (or at least succeeding in the same way it has to date).  Which means that for Schumpeter, the failure of capitalism has nothing to do with revolution.  It will happens as part of the glacial shift towards new modes of production (and more importantly a shift in the manner of production of new modes) towards new social organizations, and towards new mindsets.  When the ground is prepared, the transition to socialism will be complete in all but name -- as well will see in the next section, Schumpeter's definition of socialism means that we may only recognize the revolution after the fact.  In which case, I guess it might well be televised -- as a PBS documentary.

Part 3)  A discussion of what socialism is or at least could be that takes it out of the realm of ideology and utopianism and makes it amenable to matters of degree.  

In fact, given his definitions, his prediction of the inevitable advent of socialism has proven true in many senses -- we are all socialists now (already).  While you might disagree with using the word this way, Schumpeter's definition of socialism is straightforward and common-sensical.  He distinguishes a socialist from a commercial society simply by considering how production is determined:

Commercial society is defined by an institutional pattern of which we nee only mention two elements: private property in the means of production and regulation of the productive process by private contract (or management and initiative). 
 
By socialist society we shall designate an institutional pattern in which the control over means of production and over production itself is vested with a central authority -- or, as we may say, in which, as a matter of principle, the economic affairs of society belong to the public and not to the private sphere.

Notably, that's it.  There's nothing about freedom and democracy and equal incomes and etc ... etc ... Once again, there's no ideology here, but a practical question of mechanism, a question which is amenable to a matter of degree.  With this definition in hand, you can in principle say things like, "10 years ago we were 42% socialist and now we are 50% socialist".  Just introducing the concept of measurement to this debate contributes enormously and goes to reinforce the notion that the process of economic change is evolutionary, and evolution is always a matter of degree.  The other virtue of this definition is that it only talks about the core element of socialism, and leaves it open what the rest of a socialist society might look like.  The precise opposite of a prophet, Schumpeter rightly sees the limits of this definition as a virtue; there are a lot of very different societies that would fall under the category, and we may end up with any of them.  

The rest of the section on socialism consists in Schumpeter briefly sketching out and then evaluating an economic blueprint of one possible type of socialism, in order to show that the concept of centralized public control of production is perfectly coherent and perhaps even superior in some ways to a capitalist system.  Perhaps not surprisingly, the thought experiment he sets up is based on the similarity (in economic terms) of a socialist society and a perfectly competitive capitalist society.  Imagine a capitalist economy that is so competitive that no firm is able to turn a profit and they all manage to just cover their costs and maintenance expenses so that the whole system continues to barely function and reproduce itself.  In terms of pure economics, that could be identical to socialism.  In the case of socialism, decisions about production would be made at the level of industrial boards using techniques like equally distributed vouchers to gauge demand.  In this hypothetical perfectly competitive capitalism they would come from the choices of individual producers in a market.  But it's pretty easy to see that the end result could look extremely similar.  Schumpeter goes on to add some bells and whistles to this thought experiment that are meant to show how socialists might deal with other problems that come up in the logic of economic production (eg. savings, technological change).  In the end, he does decent job of convincing you that socialism could duplicate a lot of the features of modern capitalism from a completely different starting point, a fact which he attributes to a general logic to the rational allocation of economic resources in any system.

In fact, Schumpeter goes one step further and proposes that intelligent socialists will certainly make use of all kinds of markets and price signals as tools to achieve their ends, and that in actual practice socialism will look remarkably like (very) big business capitalism.  Just substitute GE for the ministry of industrial production, GOOG for the ministry of technology, and JPM for the ministry of finance.  Not so far from our current economy ¿no?.  Of course, Schumpeter imagined that the public would gradually take over these companies rather than vice versa, but we'll come back to that question later.  This observation that big business capitalism naturally dovetails with socialism starts to be particularly interesting when you think about the system the Chinese seem to have built.  It has significant government involvement in the economy (though I wouldn't venture to say what % socialist they are) and yet still uses lots of tools like price signals and markets to help make economic decisions.  At this point, it's obviously pretty tough to claim that this sort of economic system won't ever work because, it's ... you know ... socialist! damn it!  Schumpeter evens anticipates the love the modern CEO professes for the efficiency of the Chinese system of making decisions by observing that a well run socialist system might solve a lot of problems that plague capitalism and slow down its progress.  The modern monopolist will feel right at home in socialism.

Unless of course he loses his job.  Which brings up the really tricky question about socialism as far as Schumpeter is concerned.  When a socialist system is well run, it can be at least as productive as any idealized free market capitalism or even its natural successor, big business capitalism (socialism itself is really just the end point of that sequence).  But how can you be sure it will be well run?  Capitalism has an advantage here because it solves this problem at the same time as it solves the problem of allocating production -- the people who repeatedly succeed in giving everyone what they want at the lowest cost are the ones who end up with the profits and hence guide the future allocation of capital.  Deciding what to produce, deciding how the fruits of this production should be divided amongst the population, and deciding who should decide this are one and the same algorithm in the capitalist system.  The idea is that the system self-selects its own leaders in the same way evolution does.  Of course, Schumpeter's point is that the alignment of these decisions breaks down naturally as capital evolves into larger and larger businesses with professional bureaucratic management and widely dispersed ownership.  But this only makes the problem stand out all the more when we come to socialism.  The giant bureaucratic system for making decisions about production becomes divorced from the system for making decisions about how we split up the production as well as about who decides.  In short, how do you appoint the bureaucrats who are supposed to run this show?  Under socialism, these can no longer be purely economic decisions.

Schumpeter adds some interesting thoughts at the end of this chapter about some of the extra-economic forces that already go into these decisions today.  He tries to soften the problem a bit by pointing out the way that capitalism has always existed within a larger social structure that had something to do with these decisions, even if this was just the minimal bourgeois state (ie. just the remnants of feudalism).  Socialism too should be able to borrow some behavioral elements that are not explicitly economic so that competent people rise to the top while the milk stays docile below.  But these explanations only go part of the way to solving the problem because the real answer to the questions of how the bureaucratic hierarchy is determined is: politics.  Which I think is why Schumpeter felt the need to devote the next chapter of the book to a theory of democracy.  In the end, however much a materialist you might be, the evolution of economics is inseparable from politics.  

Part 4)  A theory of democracy as nothing more than free competitive elections.  

If capitalism is the system where anybody can go into business making what they want in a free and open market, democracy is the political analog -- anybody is free to compete for office.  In other words, Schumpeter constructs a theory of democracy as a political method, not a goal to be aspired to in and of itself.  Fundamentally, we want democracy as a political system because of what it can do for us.  We want the freedom, the stability, and the flexibility we associate with democracy.  Despite the quasi-religious ring the term has taken on in the US, nobody supports the idea of democracy when the "the will of the people" is doing something like making black people slaves or exterminating jews.  Schumpeter doesn't see this as some shameful flip-flopping on our part; we have to evaluate democracy as a mechanism and judge it by its outcomes in exactly the same way we judge capitalism.  

In my opinion this blasts 99% of all discussions of democracy out of the water as utterly confused before they even begin.  It makes certain questions (eg. anything involving the "will of the people") look like they were asked by someone who still believes in Santa Claus, and at the same time opens up a whole new interesting set of questions about how democratic government actually does and should work in practice.  The overall effect here is similar to his theory of capitalism -- it is a much better empirical description of what is going on every day, but one that makes it surprising that the system works anything like as well as it does (in fact, surprising that it works at all).  To see the success of democracy in the past 200 years as fortunate accident, or at least as a special case amongst a more general set of possible outcomes, is deeply profound.  And scary as shit.  But this is how the system we have actually works.  Politics is about power, about getting re-elected, and only secondarily and incidentally about anything to do with the welfare of the nation.  We ignore the mechanics of this system, we allow ourselves to replace the machine with mysticism, at our peril.  



With this theory of democracy Schumpeter set up the final big question of his book.  On the one hand we have a capitalist mechanism for production based on competition for profits.  This mechanism tends to evolve through a logic of its own towards big businesses and from there into socialist bureaucracy.  On the other hand we have a mechanism for selecting people to run this state bureaucracy based on a competition for votes.  The two mechanism our now intertwined in a way they were not during the era of the minimalist state (which for Schumpeter is actually just the state leftover from feudalism).  How do you know the voting mechanism will produce people capable of running the economy when it has nothing to do with economic questions per se?  How do you know if production will make economic sense if its most proximate goal becomes a means to keep someone in office?  In short, how do you find a way within this political machine to limit the state and leave (what is left of) the capitalist machine some room to function now that production is inherently political?  Socialism will be a bit like safely riding a horse -- you have to tame the animal, but you also have to teach the rider how keep it on the trail.

In the end, while Schumpeter argues that socialism will gradually but inevitably dawn on us, he doesn't argue that it will necessarily be a good idea.  The question he poses in this final section is actually left open.  It seems he is neither an optimist nor a pessimist on the question.  That non-answer, it turns out, saves him from what I see as the largest objection to his theory.  

I began with the quote from Mancur Olson's Power and Prosperity where he called Schumpeter "profound but not prescient".  Hopefully, by now, dear god please, I have gotten across why I agree with the first part of that assessment.  The second part I expect is obvious.  Sixty years later, we still consider ourselves capitalists, so something must have gone wrong with Schumpeter's theory.  Though I feel like one might argue with jumping to this conclusion -- after all Schumpeter is talking about tendencies and matters of degree and very long time horizons, and I think he clearly got this trend towards bureaucratization and the watering down of the meaning of ownership correct, so he might not be wrong so much as slightly early -- I think there is some real change in trajectory that Schumpeter doesn't account for.  And it seems to me that this is exactly the trajectory that Olson took up.  

Olson gave a lot of thought to the same general questions as Schumpeter.  How does politics interact with economics?  How does each mechanism evolve independently and in tandem?  Perhaps just because he was writing later, he saw a different tendency at work -- he argues that economic interests tend to capture the political process and so pervert both the political system as well as the alignment of the capitalist system with long term consumer welfare.  In terms of its general classification, this is a theory about the evolution of capitalism as a socio-economic system that descends from Marx and Schumpeter.  In terms of its conclusions, it is distinct from both.  Capitalism won't evolve into socialism either because of its failure or because of its success.  Instead, it will co-evolve with the state in a way that confounds any judgement about whether it has triumphed or collapsed at all.  

This theory that free market capitalism (such as it was) will evolve into crony capitalism seems to me that most accurate understanding of our current predicament.  Does that mean that Olson somehow refutes Schumpeter though?   He clearly takes the theory in a direction that Schumpeter himself doesn't mention and doesn't seem to have thought of.  Instead of business evolving into a state-like bureaucracy that the state can then capture, Olson sees the state evolving into another tool that businesses can make use of in their competitive struggle.  Yet Schumepter left open question about whether the mechanism of democracy under socialism could find a way to limit itself so that the economic mechanism could remain sufficiently independent.  Is crony capitalism essentially the same thing as bad socialism?  There's a clear conceptual difference, but are they practically the same?  Maybe Schumpeter's well hidden foreboding with regards to what he imagine our socialist future would be like wasn't so far off base after all.

Part 5)  A history of actual socialist and communist movements from Marx to the end of the Second World War.  

A lot of this section is filled with forgettable detail (explain to me again who the Fabians were?), but the interesting overarching theme is that the world wars are like tectonic plates shifting under us, pushing up the new mountains of socialism.  Each earthquake brings us a little closer.  I agree with this strongly, though with a slightly different spin -- history teaches us that the state grows by constant emergency.  And once it gets new powers, it never gives them back.