Monday, February 28, 2022

Elasticity of Substitution

And now I've wrapped up the fourth book in my current series, tentatively titled "Completely Unrelated Books That I Happen To Be Reading Simultaneously For No Good Reason." This ends us on a strong nonfiction note with The Economics of Inequality, a relatively slim book on economic theory. It was written by Thomas Piketty in the 1990s, and was translated into English after the success of his later works.

Before reading this book, I saw some online reviews along the lines of "A good introduction to Piketty's ideas" and "More approachable than Capital in the Twenty-First Century." I have to disagree with those assessments. It's a good book and well-written, but also much more technical than his later books. It looks less intimidating thanks to being much shorter, but it's primarily written for other economists and doesn't have the easy flowing narration or friendly language of his doorstop books. Which is all to say that I'd recommend this for people who already like Piketty and are interested in the topic, not as a gateway to his other writing.

That said, I do recommend the book. I'd kind of expected it to be an introduction to his works on inequality or a rough early draft. Instead, it goes over different ground: while Cit21C and Capital & Ideology are primarily oriented towards explaining the origins of inequality and documenting how much it has risen, The Economics of Inequality is focused on finding the best tools for reducing inequality: what specific policies governments or society should take to make life more equitable for its citizens.

Reading this book was a little trippy at times: it was written in the 1990s, but pays a lot of attention to ideas that I personally have only more recently become aware of. There's a lot of focus on social justice, which is closely tied to but not synonymous with equality. Later in the book he writes at some length about universal basic income; even at the time the idea was several decades old, but at least here in the states it's only relatively recently that it's come to the forefront of political dialogue.

TEoI starts relatively slow with a chapter that mostly recaps (or, I guess, precaps) the idea of deciles and centiles from Cit21C. Piketty gives a pretty thorough explanation of what these segments mean and how they are calculated; having read over two thousand pages from Piketty this level of explanation felt a bit overlong, but it was probably very important to readers back then, and it may even be a better introduction now for new readers than his more recent books.

Things really start moving in Chapter 2. As we've been discussing since at least Marx, the main problem is that workers tend to receive relatively little from their labor while owners tend to receive large amounts, which results in wealthy people growing even wealthier over time while the working poor tend to remain poor. This is historically called the "labor/capital split", namely, how much of the revenue generated by economic activity should be paid as wages to labor versus how much should be collected as profit to owners. Historically the focus has been on "direct redistribution", where laws or labor action forces the owners to convert some additional profit back into wages. This works if labor and capital aren't very elastic: that is, if you always need X workers per Y machines in order to create Z output, then as long as the owners still earn some profit you can happily increase the workers' wage.

But Piketty finds that in practice labor and capital are elastic, especially as the economy evolves to be more service-oriented. There is a high "elasticity of substitution", which means that to some degree you can replace workers with machines or vice versa. (One periodic, mild point I tend to stumble over is the overloading of the world "capital" in these books. The classical definition of "capital" is "something that is used in production without being consumed", like a hammer or a factory. But its more prevalent modern form is basically a synonym for "wealth", like shares of stock or an interest-bearing loan. In this specific case Piketty is referring to the economic capital, while elsewhere he is referring to financial capital.) What is the implication of a high elasticity of substitution? If the price of labor rises, then owners will try to eliminate jobs and replace them with machines. We're seeing this happen now with self-check-out lanes in supermarkets; those machines are very expensive to install and somewhat expensive to operate, but they don't require health-care benefits or wages. This will reduce the amount of revenue flowing to wages, which automatically increases the share going to the owners.

Piketty argues that instead of direct redistribution tactics, like requiring owners to pay for benefits, we should use fiscal redistribution. This means taxing owners (well, everyone really, but owners pay the most) based on the profits they earn, and then use those revenues to transfer wealth to the lower-income people of the country, in the form of direct payments or social programs or education. This accomplishes the game goals as direct redistribution, as it is taking money from the rich and giving it to the poor, but importantly it decouples the tax burden from the use of labor. An owner who earns $1M a year will owe the exact same tax whether they employ 0 workers, 10 workers or 100 workers. So there is no added incentive for the owner to downsize.

Reading this was a little surprising, mostly because it's probably the clearest contemporary argument I've read for old-school capital-L Liberalism. The idea of collecting lots of taxes to accomplish social goals was dominant in the mid-century, but seems to be universally derided today, with the right obviously hostile towards any redistribution (and often believing that the poor deserve their lot due to inferior work ethics) and the left today much more energized by progressivism, which generally seeks a small but strong state that compels actors to behave well rather than a large state directly facilitating social transfers.

Piketty backs up his arguments with the same sort of comparative analysis he uses in his later books, particularly focusing on France, the United States, United Kingdom and Germany for the last 130 years or so. This allows us to see what aspects remain consistent over time, and how other elements trend one direction or another based on policies. He notes in the foreword and a footnote where some specific statements may not be accurate; in particular, in the book's original text he observed that the labor share is consistently about 2/3 of GDP, but further research since then has shown that this share has declined since the 1980s. That doesn't invalidate his point, but it adds more nuance (which, now that I think about it, is one of the big thrusts of Capital & Ideology.)

As with his other books, Piketty carefully distinguishes between income and wealth. At the time he was writing this book in the 90s, income was the biggest driver of inequality: after a long period where the richest 10% of people in a country would earn about 2.5x the average wage, it had climbed to about 3.8x the average (and of course would continue to rise in later decades). After looking at after-the-fact redistribution, he then focuses on how and why wages have diverged so much. An early topic is education, sometimes called "human capital", which he mentions in his other books but is more of a focus here.

It takes time and money to acquire skills that you hope will increase your income. In an ideal world, everyone would have meaningful access to acquire these skills so they can eventually perform the jobs that they are best suited for. In practice, though, the wealthiest people from high-status backgrounds will be able to complete their educations and get great jobs, continuing their privilege down to the next generation. People from poorer backgrounds may do even better at those jobs, but they don't get those opportunities: they can't afford a four-year degree, or fear going into debt, or are mindful of the risks they will bear without money or connections to backstop potential failure. There's an obvious social-justice component to this problem: it is unfair that people who are already advantaged increase their advantages. But there's also an argument to make on the basis of pure economic efficiency. If the entire labor force de-facto had access to education, then the best people would end up in the right roles, and the economy as a whole would be more productive.

There are some really interesting specific examples Piketty looks at. In Germany, many large employers fund apprenticeship programs and trade schools, free of charge to anyone who enrolls. And there's no requirement that someone completing the free program has to work for the company that paid for their education. Why do employers pay for these programs, then? It's because Germany's strong labor unions have negotiated strict wage schedules across the entire industry, so the same position pays the same wage at every company, so there's no incentive for a worker to take free training from one company and then switch to another for paid employment.

This sort of scheme almost certainly wouldn't work in the United States. Even by the 90s we had seen two decades of wage competition between companies, causing bidding wars that drove up the prices of managers, doctors, lawyers and other high-income people. So, for example, a hospital here wouldn't want to underwrite the cost of a doctor's medical training, because that doctor could and probably would leave to work for a higher-paying clinic.

France was outraged to learn that it was rated the "least egalitarian" country. France prides itself on giving all citizens access to a public education, but in practice the elite (children of managers, politicians, etc.) go to elite schools, and the state spends 10x as much per student in an elite school as it does on a non-elite school. There's a big gap between the ideals of an education system and the reality of who benefits and why. Germany doesn't make the same lofty claims as France, but in practice the German system is more egalitarian.

 As usual, Piketty is good at being precise about what he's talking about. Here he distinguishes between whether a course of action is good for social justice purposes, for economic efficiency, or both. Even though he shows in Chapter 2 that fiscal redistribution is preferable to direct redistribution, he notes that minimum wages are still important in many contexts, in particular in monopsony markets (where a single employer or group of employers artificially hold wages lower than they should be). Similarly, affirmative action policies help unlock the untapped potential of historically disadvantaged groups, which benefits the economy as a whole in addition to being morally just.

The US comes off pretty well in this book. It has its problems with rising inequality, but has done a better job at employing more workers at all levels than France. The Earned Income Tax Credit in particular receives praise for helping to fight the "U curve" of marginal rates (there's a "poverty trap" where someone who was previously unemployed and starts working finds their social benefits dramatically reduced, decreasing their incentive to take a low-paying job in the first place; the EITC compensates for this, and Piketty would like to see other changes like reducing or eliminating payroll taxes at low income levels). Of course, this book was written in the 90s before George W. Bush was "elected" and his tax cuts passed, which strangled the capacity for transfer payments, slashed the taxation of financial capital and places most of the burden on work instead of wealth. It's a bit disenheartening to think about where we might be at now if we hadn't taken that turn; still on a bad trend after Reagan, but the broad-based economic growth of the 90s might have continued longer instead of our rapid return to Gilded Age levels of inequality.

The book ends with this excellent conclusion: "Although it is essential to identify efficient redistribution wherever it exists, it is pointless to denounce every inequality as a sign of gross inefficiency that the right policy can eliminate. To do so is to delegitimize the taxes needed to finance fiscal transfers, which may not eliminate every imagined inequality but nevertheless help to attenuate very real inequalities in standards of living." I think this kind of sentiment is what I love and admire most about Piketty's writing. He isn't a cold realist or a starry-eyed idealist: he's a serious, practical and humble man who recognizes that perfection is impossible but we can certainly make things far better than they are now, and the tools to do so are absolutely within our grasp.

Thursday, February 24, 2022

Rising Sun

I've fallen out of the habit of reviewing entries in Charles Stross's "Merchant Princes" series on this blog. I enjoy them a lot, but between being relatively breezy reads and inevitably ending in cliffhangers they're a bit hard to review. That said, I think I've now completed the entire series: "Invisible Sun" just came out, and for once I'm basically caught up to the series.



This series started in the early 2000s, with the first book set in 2002 and published in 2004. The gap between the in-series year and the real year continued to grow with each subsequent entry, with the sixth book set in 2003 and published in 2010. The recent books, sometimes referred to as the Empire Games trilogy, realigns the timelines, skipping forward about 15 years to come closer to our own time.

But not our own timeline. One of the things I love most about this series is how it gets progressively more bonkers as it goes along. In many ways the 2020 vision of the "Timeline 2" version of the United States of America seems like a only-slightly-pointed critique of our own: a paranoid state, bullying towards the rest of the world, obsessed with spying on its own citizens, pathologically incapable of de-escalation. The specific reasons for that are slightly different, though: in the Merchant Princes state, the 9/11 attacks were shortly followed by the 7/16 attacks, where spies from an alternate universe detonated nuclear bombs inside Washington, D.C., killing George W. Bush and Dick Cheney and much of Congress and the Supreme Court. Donald Rumsfeld became the Commander in Chief, authorized a totally annihilating counterstrike against the "Timeline 1" Gruinmarkt, and doubled-down on the PATRIOT act and other forms of domestic surveillance.

One of my annoyances with the middle books of the series had to do with Miriam's ambitious plan of technological arbitrage, using discoveries from Timeline 2 to speed up innovation in Timeline 3 without falling into a development trap like happened in Timeline 1. I loved the idea of this and wanted to read entire books about Miriam kicking butt and schooling people. Instead, after getting us all excited about it, she got whisked away to an evil Baron's lair, impregnated against her will and spends much of the series as a passive victim.

All that cool stuff eventually happens, but it mostly happens off-scene, in the 17-year gap between books 6 and 7. And really, that's part of the point. Miriam and the surviving members of the Clan want to avoid the mistakes of the past and have the civilization grow up alongside the tech, instead of having it be dropped into their laps. That means, for example, building a generation of computers running on vacuum tubes that they know will be obsolescent; but those more mechanical computers are easier for people to understand, and with some prompting they can then go and invent silicon processors on their own, and actually understand how they work.

The final trilogy focuses on a mix of new and old characters. The old are very old, and often complain about the aches and bodily indignities that come with age. Some of the new characters are related to the old, most notably Rita, who was Miriam Beckstein's illegitimate daughter we heard about way back in Book 1 and is now a young adult. Others are brand new, including Rita's girlfriend Angie, Rita's adopted grandfather Kurt (who we eventually learn is a very-deep-cover sleeper agent from the now-defunct East Germany), and Elizabeth Hanover, the daughter and heir to "King" John Frederick.

The earlier books culminated in a successful democratic revolution in Time Line 3, led by Sir Adam and successfully toppling the British crown-in-exile. In this timeline, the Industrial Revolution didn't start until the 1900s, France is a superpower controlling all of Europe and Africa and most of Asia, and there has not been a large-scale democratic experiment since the ancient Greeks. One intriguing thing we learn early on is the particular structure of the New American Commonwealth. It's a democracy, but as the very concept of democracy is unfamiliar to people, it has unique institutions and processes to defend against a return to autocracy. That system, we eventually learn, is modeled after the Islamic Republic of Iran. Sir Adam becomes the "First Man", essentially equivalent to the Ayatollah: not involved in day-to-day administration or policy, but the ultimate authority who will step in if the democratic ideals are threatened. A Council of Guardians-type body reviews and approves candidates and parties, excluding monarchists and others who seek to end democracy. And then there's the popularly-elected parliament who set the law and the ministers who execute it, under the watchful and generally lenient eye of the First Man.

The main plot of the later books has to do with a series of schemes between France and Britain. Earlier in the series, New Britain seemed doomed, relatively safe between the two oceans but helpless to block a steadily rising France. After 17 years have passed, France still has superior territory and population, but the New American Commonwealth has vastly outstripped it economically and even militarily, thanks to the advances engineered by the Ministry for Temporal Intelligence and the Department of Para-historical Research. The exiled King of New Britain has been forced to claim sanctuary in the court of his erstwhile rival the King of France. The Dauphin is promised to the princess Elizabeth Hanover, setting the stage for the destruction of democracy: France will then use its mighty armies to invade the Americas and restore the British monarchy, albeit as a client state.

Catching wind of this plot, secret intelligence in the Commonwealth comes up with a counter-scheme of their own. Elizabeth does not want to marry the Dauphin, who is much older and openly cavorts with mistresses and, rumor has it, has the pox. World-walkers make contact with her while she's in a finishing school and they come to an agreement: she will defect to the Commonwealth, renounce any claim to the throne, gain citizenship and a large sum of money. At the end of Dark State, this goes slightly wrong: her rescuer is shot, and then American agents from Time Line 2 break up the operation, leaving her stranded in an unfamiliar version of Berlin.

It's a fun, energetic story. I felt like the Elizabeth part was the core, but there's a ton of other stuff going on too: lots of political maneuvering in the Commonwealth when Sir Adam dies of natural causes and they face their first succession crisis; ongoing attempts by black-ops groups from the US to gain leverage over or neutralize the Commonwealth; and, just to keep things exciting, the introduction of an interstellar race of insect/machine hybrids who have also learned how to world-walk and have annihilated all life on millions of alternate versions of Earth.

There's a lot of plot threads all running at the same time, which is really cool, but it also means that the people in each individual storyline don't get a whole lot to do. Elizabeth did most of her exciting stuff in the previous book; here, she learns how to use a phone, goes shopping, flies on a plane, and talks to a camera: that's about it. It's pretty common for us to spend a little time with someone like Adrian Holmes or the green recruits and then never get their POV again. That isn't unique to this book or author: whenever I read, say, GRRM, I'll feel like I'm just getting interested in what one character is doing when the author whisks me away to a different plot line.

I had a few minor annoyances. I'm reading the first edition, and there are a few typos; I didn't see any in the previous books, so I'm sure they'll be corrected in reprints. I was also a bit turned off by some repetitive framing: reading someone talk about what they're going to do, then reading about them doing it, then reading them describe what they did. That would be appropriate for something like a television series where people are dropping in and catching up midway through, but doesn't seem necessary for a novel that you can read in a few days.

But I think most of my (mild) complaints boil down to me wanting more stuff to happen, and that's very petty of me. There's a lot that goes down, and more importantly, it satisfyingly ties up all of the major plot threads of the series. That's no small feat for a long-running speculative fiction series! Wanting more is a strong sign of how much I enjoy it.

Besides the straight-up plot, I also enjoy Stross just nerding out on topic. Unlike his cyberpunk books that have deep and trenchant insight into technology, these books go hard on economic, historic and political topics. I love the attention to, say, the fashion industry: we see how the Commonwealth's garment business is organized, with everything made to order, and thus clothing is much more expensive, but also of significantly higher quality, and not reliant on sweatshop labor or other abuses.

There is some straight tech stuff that's fun, too. One of my favorites is JUGGERNAUT, the enormous space battleship that MITI constructs in secret. It's apparently based on a real project called Orion that uses nuclear bombs to create the propulsion that launches a vehicle into orbit. In our own universe that would be a terrible idea for very obvious reasons, but in this series, you can build that spaceship in an uninhabited version of Earth, nuke it, then switch timelines once you're in orbit. The constant nuclear explosions are terrible for electronic equipment, and so the computers on board JUGGERNAUT are mechanical instead, able to ignore EMP bursts.

The whole approach towards world-walking has evolved throughout the series, and Stross has the most fun with it in these final books, imagining how Newton's laws work. Another great late example is a near-stealth nuclear attack: in one timeline you fire off the ICBM, wait for its booster to eject, wait for the warheads to detach, and then switch it into another dimension, undetectable until immediately before impact. That's one way to make a bang! 


Based on the afterword, it sounds like Stross is definitely done with this series, and I'd say that he's more than earned it. I'm grateful that he wrote such a satisfying and enjoyable conclusion.

Wednesday, February 23, 2022

Corporate Overlords

Stellaris has lately occupied the place for me that Civilization used to hold in the 90s and 2000s: the strategy game I keep coming back to in between other gaming ventures, endlessly replayable, with a long personal list of variations I want to try and new victory conditions to achieve.


One of the things that kept Civilization fresh was periodic expansion packs, typically two or three per main entry. Stellaris has far surpassed that, with 18 pieces of DLC and counting, ranging from minor cosmetic additions to major overhauls of systems or new dimensions of play. Thanks to the largess of my brother Andrew, my latest game included two brand-new (to me) big expansions: Megacorp and Nemesis. I'm not quite finished with that game, but I've defeated the Crisis (which arrived in 2420, a new record for me!) so it should be smooth sailing from here on out. This post mostly focuses on the "new" things, with some coverage of said game.

In order to get the most out of the Megacorp expansion, I decided to play as a Megacorp. In future games I won't necessarily do that; I see now that the expansion adds a lot more stuff besides literal Megacorps, so I wouldn't have been missing out if I'd gone with something else. That said, Megacorps are really interesting to play as, particularly for diplomatically-oriented playstyles like mine. Megacorporation is a new government type, roughly equivalent to an Oligarchy but with a unique set of civics to choose from. There are a few minor changes to diplomacy; in particular, you can't vassalize other empires or make protectorates or tributaries, but you can make them Subsidiaries: you agree to defend them, and they pay 25% of their Energy income to you.

The defining feature of a Megacorp, though, is Branch Offices. You can open Branch Offices on planets settled by other civs that you have a commercial pact with. This costs an amount of Energy and Influence scaling with the distance from your borders. Once established, you earn Energy income based on the Trade value of the planet. You also gain the ability to build Megacorp Buildings in your branch office. These are usually symbiotic buildings that give a benefit to both the hosting empire and to yourself. My personal favorites were:

  • The Commercial Forum gives them a Merchant job and you 25% more income from the planet.
  • Corporate Embassy gives them a few Clerk jobs and boosts your Diplomatic Weight, increasing the Economy part by 5%.
  • Mercenary Liason Office gives them a Soldier job and their armies from this planet some extra starting XP, and giving you more Naval Capacity.
  • Temple of Prosperity gives Prosperity Preacher Jobs and increases Spiritualist Ethics Attraction (more on this below).
  • Xeno-Outreach Agency gives them a boost to Trade Value and increases your empire's Immigration Pull.

There are a lot of other building options, but most of them give you a flat amount of resources like some Minerals or Food, which I didn't find very useful; a boxed-in Megacorp that can't expand on its own might find them useful, though. You are limited to constructing a maximum of 4 buildings per planet, though, and less than that based on the level of the capital building. There are some pretty interesting factors to consider when planning the buildings. Some, like the Commercial Forum, provide more value based on the wealth of the planet. Others, though, provide a flat bonus regardless of the Trade. In my game I only opened Branch Offices on high-value planets, but I could definitely see an argument being made for always opening them, and just focusing on the right bonuses for each planet.

My empire was named the Hanar Mercantile Guild, but on further reflection we were basically space Scientologists: a ruthless for-profit religion that collected staggering sums of money from the entire galaxy, and poured the proceeds into real estate and further expansion of our evangelizing. Specifically we followed the "Gospel of the Masses" and "Free Traders" civics, later adding "Mastercraft Inc." as our third one. I took Fanatic Xenophile for the large 20% Trade bonus, but the extra Envoys also came in very useful when buttering up other civs to approve commercial pacts. I opted for Spiritualist mostly to get the Gospel of the Masses civic. This was my first game taking Spiritualist as an ethic, but all of my games so far have included large Spiritualist populations: due to my xenophile playstyle, I sign agreements with Spiritualist empires and inevitably get a crabby Spiritualist minority to placate. Gospel of the Masses gives you bonus Trade for every Spiritualist pop on your planet or on a Branch Office. Imagine my surprise when I discovered that I was the only empire in the entire galaxy to be Spiritualist! I had my work cut out for me. These dianetics weren't going to e-meter themselves.


I selected Hegemon as my Origin so I could bootstrap a Trade League. As with my other Federation or Hegemon starts, I was prepared to restart if I had a bad position or allies, but this one was pretty good. The Jesselniks had plenty of room to expand to the west, the Bosh'tun were soft-boxed-in by a xenophobic fallen empire to the south but could still grow a bit, and I had lots of room to grow east. I put all of my starting Envoys into supporting the Hegemony, until my Cohesion got up to 100, at which point I convinced the other two starting civs to reform into a Trade League. There is huge synergy between the Trade League and Megacorp: automatic free commercial pacts with all members, bonuses to trade value, trade route protection, and so on.


Another huge advantage of the Trade League is that it unlocks a unique Trade Policy that combines Consumer Benefits with Marketplace of Ideas. So each Trade gives you 0.5 Energy, 0.25 Unity and 0.25 Consumer Goods. I ran that for most of the game, which lets you cruise through Ascension Perks and avoids a common early-game resource bottleneck. Before long I was running a large and consistent surplus in Consumer Goods. Fortunately, Megacorps have discounts when trading on the Galactic Market (and even bigger discounts when they host the Market). In the past I've manually sold large batches of items when I hit my limit and built Resource Silos on starbases to increase my capacity. This time I was good about setting up automatic monthly trades, selling off excess items (consumer goods, food) and buying useful things (alloys). There's a little bit of micro involved there thanks to the fluctuating prices, but overall it let me run a leaner and more efficient economy without huge stockpiles of useless goods. Eventually, though, I was offloading so many consumer goods that their value had cratered to something like 0.1 energy per good, so later in the game I swapped back to pure Wealth Creation. I still had a surplus of Consumer Goods and my income skyrocketed.

Trade is huge for megacorps. In all of my other games of Stellaris, trade has been something you passively acquire as you build city districts: you hope to keep as much of it as you can, but it's a mildly beneficial side-effect. For megacorps, though, it's absolutely worth focusing on trade: you can get a lot of trade multipliers, so clerks will generate more energy than the equivalent technicians would. You also have the flexibility of adjusting your trade policy and directly converting trade into unity and consumer goods (assuming you're in a Trade League, which you should be). You also have easier access to Merchants, giving a further boost to trade generation. But all of this means insane Piracy levels thanks to the extremely lucrative trade routes back to your capital. Fortunately, there are several Federation perks and Galactic laws that can increase your base trade protection, but even so my navy was kept busy for a long time while I desperately hunted the Gateway technology.

The boosted Trade can feel borderline broken at times. One big example was dealing with Fallen Empires: they could have Overwhelming Technology and Fleet Power, and still be ranked as Pathetic relative to me due to the insane strength of my Economy. This let me do some stuff that I would otherwise never get away with, like colonize systems adjacent to the Fanatic Xenophobes.


One complaint I'd read online was that if you play as a Megacorp, you inevitably end in a galaxy with multiple other Megacorps, which isn't fun since they'll aggressively open their own Branch Offices and block you from expanding your corporation. I'm not sure if that's changed or if I was just lucky, but I didn't encounter any other megacorps in this game.


As I said above, the Megacorp expansion adds a lot of stuff besides the titular Megacorps. The Caravaneers are fun; somewhat like the Curators from Distant Stars, they're a unique Empire with a more story-based Diplomacy interface and some cost/reward choices to make. They remind me of Ferengi from Star Trek, eager traders who love commerce. You do need to keep on your toes, as some of the deals are good but many are objectively bad. There's also a bunch of related events that can trigger when they're around, some of which have very rewarding outcomes. I didn't luck out on any of the mystery boxes I purchased. Very late in the game I was bored and decided to destroy their home base, which ended up being really fun and funny: in addition to some instant loot, you also gain a permanent increase to your Opinion with all other empires, and unlike everything else in the game it does not decay. Furthermore, shortly after you destroy their base, they'll pop back up again in an adjacent system, where you can destroy them again, and collect more loot and a stacking opinion bonus!


Megacorp also adds a bunch more Megastructures. These don't require taking an Ascension Perk, which is a nice change from the ones Utopia added. I acquired basically every new structure other than the Mega Art Installation. I conquered a ruined Sentry Array and restored it, built a Strategic Coordination Center (which boosts your naval cap, starbase cap, defense platforms, and sublight speed) and an Interstellar Assembly (boosting diplomatic weight and opinion). I also made the Science Nexus and the Mega Shipyard as usual.


One minor gripe I have with the game is that so many big parts of the game only come at the very end of the tech tree. By the time you can start actually building your first megastructure, you're researching future techs. It feels weird to build a Science Nexus after you've already learned all of the unique technologies. And while the Interstellar Assembly is cool, by the time I could build it I had already passed the "I Am The Senate!" threshold of running roughshod over the entire galactic community. Instead of the megastructures requiring a lot of investment and then propelling you forward (as with Civilization Wonders), they're built by empires that are already ahead of everyone else and then ensure that nobody could possibly catch up. I'm not necessarily complaining, since I like being ahead, but I think they would be way more interesting if they were available earlier in the game.

While Megacorp made big obvious changes to my game from the very start, the Nemesis expansion didn't really become a factor until later in the game, but I liked that. The mid-late game tends to be the least exciting, so anything that shakes that experience up is good!

I'd previously played a game with the Espionage system. Somewhat like how Federation unlocks many more options related to Federations, Nemesis unlocks most of the functionality of espionage. I think I like it. It gives you some levers to interfere with other countries without actually going to war with them, which is a good match for the types of games I typically play. The base-game Gather Intelligence is borderline useless, but the Nemesis-only Operations can be pretty cool, especially options to steal technology or to poison relations between two allies. The main downside is that espionage eats up a valuable Envoy; but as a fanatic Xenophile I was swimming in envoys, and would often park them spying on my biggest rivals while I wasn't completing First Contacts or building up Cohesion. Broadly, espionage gives you something different and potentially useful to do.

The Traditions were a bit different this time around; thanks to the extra tree offered by Nemesis, I had to choose which ones to pursue. For the first time ever, I did not choose Discovery: science is incredibly useful, but I tend to take a commanding science lead relatively early in the game, and I (correctly) believed that I would do well even without taking this. I grabbed Expansion early to support my rapid growth out, finished Diplomacy to soften up other empires to join my Trade League or at least sign a Commercial Pact, took Mercantile to further supercharge my Trade, and filled out with all-round strong traditions like Prosperity, Domination and Harmony. 

Likewise, I did not take Technological Ascendancy like I normally do. I explored some brand-new ones I was curious about, like Consecrated Worlds to support my Spiritualist faction, Universal Transactions to supercharge my Megacorp orientation, and Xeno-Compatibility to benefit from the influx of immigrants I was receiving. For the ascension path, I took Psionic Ascension again. I've always enjoyed this one for the writing, but by the end of this game I realized that it would work much better for a xenophobic empire. Psionic is a very powerful trait, but only your founder species gets it, so you'll be much better off if 100% of your pops are Psionic than if only 5% are. Similarly, Psionic Leaders are amazing, but in my game I had to cycle several dozen new leaders before a Psionic one would pop. And unfortunately it looks like Xeno-Compatibility won't pass down the Psionic trait to hybrid species, so that particular combination wasn't nearly as useful as I had hoped. I have yet to try the Synthetic or Bio ascendancy paths, so I'll probably try one of those the next time I play a Xenophile game.



This is my third time opening the L-Cluster and my third time discovering the Drakes. That's a pretty good outcome, but according to the wiki it should only happen about 20% of the time, so that's a little weird. I'd like to experience something different in a future game. Likewise, I got the Unbidden again in this game, maintaining a 100% streak of encountering that particular Crisis. It isn't too surprising since I always ban robits and always discover jump drives. I might shoot for a full synth game in the future to try and trigger the Contingency, or maybe just use my Game Settings to invoke the Scourge. As with my last game, this one (on Captain) was pretty anticlimactic. I had a Federation Fleet, a Galactic Fleet, and a bunch of Battleship Fleets. They all flew to the Unbidden insertion point within a couple of weeks and wiped them all out by the time they'd taken three systems. It's a far cry from my first victory, where I desperately retrofitted all of my surviving ships to properly outfit them to counter the Unbidden's shield-and-laser build.



Since then I've mostly been amusing myself with Galactic Community shenanigans. I got myself appointed Galactic Custodian (even before the Crisis triggered) and have mostly gotten my way, but not always! A new wrinkle of the latest game version is that when you send an Envoy to Improve Relations and have the appropriate Diplomatic Tradition, there is a small chance each month that the Envoy will earn a Favor from that empire. Now, in the past I've bought up Favors and used them to achieve diplomatic agreements or persuade a federation vote. But they can also be used to add 10% of the other empire's Diplomatic Weight to your own vote in a Galactic Council vote. Which is a long way of saying that, if a total of 10 Favors are called in against you in a vote, then it will completely wipe out your own position: even if you have more Diplomatic Weight than everyone else combined, with enough Favors they can block your votes. Which feels a little frustrating, but also kinda exciting, as it's a rare example of a catch-up mechanic in Stellaris.


I can't really blame them for trying to block me, though: everyone was mostly cool, even when I was voting more powers to myself, but my big scheme was to push through the Divinity Of Life series of resolutions that boost Spiritualism and slow down and eventually ban robotics. Again: I was the only Spiritualist empire in the entire galaxy: more than half of the others were Fanatic Materialists, and quite a few had started Synthetic Ascension, so it really was a matter of life-and-death for them. I failed a couple of times on the vote, which always has a 20-year timeout before you can re-propose. I vigorously used my Council and Custodian powers to hustle votes open and closed as soon as I could, trying to burn off opposing favors faster than they could be re-acquired. This late in the game I actually had an Influence surplus, and once it reached the 1k max I re-proposed the Divinity proposal. I cashed in a bunch of opponent Favors, which were much smaller than mine but were able to counteract the 6 or so against me. I kept a close eye on the vote, and as soon as I could gavel the session closed I did so. There was much howling and gnashing of teeth, but that same proposal weakened the diplomacy of the Materialist opposition, so I needn't fear its repeal. Already the winds of change are blowing, as civs are finally joining me in embracing Spiritualism and paying for their audits.

I have a while left to go in this game and am not sure if I'll run out the clock, end it, or try for something big. I was tempted to try and become the Crisis in the 75-80 years I have left; I'd saved my final Ascension Perk just in case I decided to pursue that.  But you can't become the Crisis as a Xenophile, and I'm Fanatic. Doing the ethics shift would be pretty tedious, so I might save that for another time. Although, there are ethics shifts if I decide to become the Galactic Emperor, or if I get a "lucky" roll in the Shroud and get a Chosen One, so... we'll see.


I am glad that I was able to more-or-less wind up this game when I did. A new update to Stellaris is imminent, and it looks like it will be pretty disruptive. But potentially really good! The focus is reworking Unity, which sounds amazing to me. In every game that I've played I finish with a ludicrous surplus of Unity: all ascension perks claimed, all Unity Ambitions active, and still millions of Unity points piling up with nowhere to go. And that's without ever building any autochthon monuments or otherwise focusing on Unity: in this game, with the Spiritualist boost and Priests, it was even more ridiculous. Anyways, the change is wide-ranging, but it seems like the gist is that Unity will be used for intra-empire activities that have previously cost Influence: stuff like deciding elections, reforming the government, issuing edicts, managing factions and so on. Influence will continue to be used for extra-empire activities, like making claims, building Outposts, proposing Galactic Community resolutions, and so on. I'm sure it will take some getting used to, but in principle I think it's a fantastic change that makes sense and will keep both resources valuable.

Tuesday, February 22, 2022

Waiting To Exhale

I rarely read short stories, and even more rarely write about them, but I'm about to make an exception! I just finished "Exhalation", a collection of stories written by Ted Chiang. My friend Dan gave me this book last year in a clever and heartwarming bid to get both of us reading again. I had basically no expectations going in; all that I knew about the author was that he wrote the story that was later adapted into the (excellent) movie Arrival, so it was a lot of fun to get acquainted with a new (to me) writer and voice.




I think it's fair to describe this as a book of sci-fi stories, but it takes some time to make that assessment, mostly because the very first story reads like something out of Arabian Nights. On further reflection, though, even that story is basically sci-fi, even if set during the Caliphate. Someone constructs a device, that device has very well-defined effects, and then the story explores the effects of that device.

Ted seems to play with the same idea across multiple stories. One early concept is fate. That opening story, The Merchant and the Alchemist's Gate, features a very constrained form of time travel: you can travel to the future or the past, but you cannot truly change anything. If you do something in the past, then you have already done that thing in your past: all of history has already been written. A much shorter story called What's Expected Of Us? has a similarly rigid view of fate. Here, someone invents a device that consists of a button and an LED. The LED lights up a few seconds before you press the button. This seems like a toy or a gimmick, but has profound implications that basically disprove the existence of free will. Both stories seem to say that we cannot change anything, but the emotional impact of that revelation is opposite. In the latter story it causes existential dread and a desperate plea to act as though our actions matter even when we know they don't. In the former story it's mostly a sense of comfort and relief: everything is happening as God has planned, and we needn't agonize over what might have gone differently.

I'd be tempted to say that Ted is a firm believer in determinism or predestination, but he then explores the exact opposite perspective in the last story of the collection, Anxiety Is The Dizziness of Freedom. In this world, pressing a button on a device causes a quantum divergence that leads to two parallel universes emerging from that moment. They are essentially identical at first, but gradually vary as time goes on. In one universe you might get into an argument with your spouse while in another you make love; in one universe you might find a dropped wallet on the sidewalk and decide to return it, while in another you may never see the wallet, or might find one and decide to keep it. In this story and this universe, everything is changeable, not fixed. The diverging realities show that we're all capable of making good choices or making bad choices, and the story gently urges us to be good.

The story that most touched me was The Lifecycle of Software Objects, which sometimes felt like it was written for me. As a software developer, it resonates strongly from both a business and a programming standpoint. With the long scope of the story it captures the obsolescence and decay of digital platforms, programming languages and techniques. It also is very evocative of the online fan communities I've been a part of. I think of something like the Baldur's Gate mod scene and sites like Gibberlings 3 and Spellhold Studios and Sorcerer's Place. Baldur's Gate was a big game back in the day, but most people have moved on to newer and flashier games or left gaming entirely. But a hard-core committed group remains active, more than twenty years later: writing new content and swapping stories and experimenting with new campaigns. Sometimes a seismic shift occurs, like Mac OS dropping support for 32-bit programs, and a good slice of that community will just vanish. The people in that community have passion, but passion isn't enough, you need money to convince programmers to do the boring and tedious job of migrating 20-year-old software to something that will run on your new laptop. Beyond all the tech stuff, this was also probably the most emotionally affecting story: the digients are vulnerable and cute and tragic, while Ana and Derek have incredibly relatable desires and struggles. For a while I resisted reading more of the story because I dreaded where it might be going: not being too familiar with Chiang's writing, I could see it ending with a heartwarming resolution or a funny twist or a soul-crushing loss or a terrifying horror show. In the end it's melancholy and mature, leaving several threads tantalizingly untied but with a lot of thought about where things are headed.

Most of Ted's stories seem to basically take place in our world: he adds an invention or a discovery and then explores how it would change things. A handful of the stories take place in clearly different universes, though. The first is the titular "Exhalation", which was a delight to read mostly because I initially assumed that it did take place on Earth and was narrated by a human, and I had to keep readjusting my mental frame of reference every couple of paragraphs. Another striking story set in what ultimately turns out to be another universe is Omphalos. Unlike Exhalation, the world of Omphalos does seem very earth-like, with cities and lecture halls and a postal service. The world seems slightly askew, suggested early on with alternate spellings for places like Arisona, Yosemeti and so on. The narrator of this story is a "Church scientist", an archaeologist who finds evidence of God's creation by examining tree rings and other natural phenomena. This world consists entirely of believers, and for good reason: there is physical evidence that the world is six thousand years old, and the oldest human remains have no belly buttons, proving they were created and not born. There is a Church, which focuses directly on worship, but all other fields, especially the sciences, also seek to honor God by uncovering more signs of His work.


The trippy revelation eventually comes in Omphalos: in this universe, there is a God, and He did create the universe, and He does have a special plan: but that plan doesn't involve this world. Another planet is the center of the galaxy, and the one the narrator lives on is just background, a distant light in the sky for the chosen ones to look up at. It's another shattering revelation that upends peoples' sense of values and purpose, but the story ends with a determination to continue learning and living (which is a decision that the people in "What's Expected Of Us" weren't able to make).


I've been poking away at this collection for quite some time. For better and worse, it's easier to mix in short stories with other readings, and earlier this week I was alarmed to discover that I was in the middle of reading four separate books. That's far too many, but also an encouraging change from some points in the pandemic when I was having a hard time reading any books at all! I now have that number back down to a far more manageable two books. I am pretty happy with my leisurely journey through this collection, though, as it's allowed each story to linger a bit longer and allowed me to chew on them a bit more. There's a lot to like in here, from the graceful writing to the fascinating scenarios to the thoughtful consideration of moral action, and I'm looking forward to reading more in the future.

Monday, February 21, 2022


I've been a distant fan of Jason Schreier's work for some time. He's almost certainly the best journalist covering video games. There are lots of great columnists and critics, but the vast majority of "journalism" in the field is just reprinting corporate press releases. Jason is very different: he actually interviews front-line developers, gets scoops on how and why projects failed, and has deep insight into the causes of the many problems that repeatedly plague this field. I've eagerly devoured Twitter threads from him, a few long-form articles, and even a great podcast appearance on Bullseye. It was a particular treat to read Press Reset, his second nonfiction book.


This book combines several of my current obsessions: video games, labor and capital. It reinforces a lot of my prior beliefs, but adds a deeper layer of understanding and nuance. In lots of ways, the book feels tailor-made to my interests. It starts off strong by opening with a chapter on Warren Spector, one of my favorite game designers of all time. And it doesn't just start with Deus Ex, but goes back to Steve Jackson Games and Origin Systems. One of my frequent old-man gripes is how the Ultima franchise is completely missing from contemporary discourse about games. I'm very aware that there are MANY games out there, and I've only experienced a narrow sliver of them (mostly in the RPG/adventure/strategy genres), but somehow the book manages to keep playing all of my hits: Ultima, System Shock, Rock Band, BioShock, XCOM and more.

One reason for this is that, as I've previously heard but not fully understood, game development is a relatively small and collegial field where the same people keep bumping into each other: someone leaves, starts a new company, hires their old colleagues, those colleagues leave and start their own company, which then re-hires people from the first company, and so on. That sort of cross-pollination is infamous in the Silicon Valley tech world, and it's fun to see how that plays out on the national (or even international) video game market as well.

Location is one of the major themes of the book. Jason keeps coming back to just how hard it is to work in games: the sudden surprising layoffs that happen when a game does poorly, when a game does well, when a game doesn't do anything. Software developers in Silicon Valley startups can find other gigs within the Bay Area; people working in the film industry in Los Angeles can look for other projects when their previous one ends. Gaming shares the same market volatility but not the same geographic cohesion. If your Austin gig ends, you might need to relocate to Rhode Island for the next one, then Chicago for the one after that, then Montreal, then Melbourne. The companies almost all offer the salaries and 401(k)s that you would expect from a white-collar job, but the actual employment pattern is much closer to seasonal work. Near the end of the book, Jason focuses on this as one aspect of volatility that seems solvable:  the boom-and-bust cycle seems very unpredictable and doesn't have an easy solution, but the rise in remote working during the pandemic may (may) offer a path forward that's less disruptive to peoples' lives.

Heh, I feel like this blog post is unusually disjointed, which is strange because Jason's book is very smooth and coherent! I'll hit a few more random thoughts:

One small section that I particularly loved was a brief description of the development cycle for Ultima Forever. That's actually a game that I was tracking somewhat closely on this blog, from my initial confused excitement ("BioWare is rebooting the Ultima franchise?! That's awesome!) to my dawning skepticism ("Oh, wait, EA just branded another studio BioWare. And it's mobile-only. Hmmmm.") to my inevitable crushing disappointment. It's a thing that I followed closely, but now I realize just how little I understood about the project: why it happened, who was involved, what their goals were and what they took away at it. I really only saw the tip of the iceberg, and after reading this book I understand so much more about what was happening beneath the surface and why.

This is a relatively ground-level book, mostly focusing on the experiences of individual developers as opposed to the whole studio: what it feels like to get called in to a pizza party and told that you're being laid off without severance, or the camaraderie of sitting in an open office with seven other people trying to make combat in a game feel great. But there are a few parts where politics intersect the story, always in very interesting ways. One of the most bonkers plots has to do with Curt Schilling's 38 Studios, which takes a long and messy detour through Rhode Island GOP politics. Organized labor isn't much of a factor for the book, but in the very end Jason does broach union organizing as a nascent force in the industry and one that offers a potential solution to the workers' pains. Jason makes clear his own personal bias, writing in a footnote "I can speak from personal experience here - at Gawker Media, we organized [a union] in 2015, which helped protect us from potential problems when our company went bankrupt in 2016 thanks to a vengeful billionaire."

Jason doesn't come off as very polemical, but personally I found my mind often drawing connections between the specific harms detailed in this book and the universal problems created by our current flavor of capitalism. As one specific example, a lot of the problems with EA have to do with the fact that the executives only care about quarterly results, so their decision making is entirely driven by what will most please analysts in the coming months. As Jack Bogle would say, though, it's counterproductive short-term thinking: you can easily juice your numbers a bit for this quarter by closing down a studio or rushing a game to market, but that same action will also kill off a beloved franchise that would otherwise pay well for decades or more. The executives are focused on counting instead of focusing on what counts. But those execs are acting in their own rational self-interest thanks to their compensation structure: they'll make millions if they fail, and hundreds of millions if they succeed, so why wouldn't they go for it? While the Warren Specters of the world are happy making games that are merely profitable, Wall Street and the C-suite demand exponential growth, no matter the human toll. As with so much of America, changing this dynamic will probably require changing our ownership structure. What would it look like if workers had seats on the EA board in Redwood Shores? Or if executives knew that they would be at the same company twenty years from now, reaping what they sowed back in 2022?

I do find it hard to mentally categorize exactly where video game workers should fall in the labor hierarchy. On the one hand, they're treated very poorly in some very unique ways. Crunch, which Jason doesn't dwell on a whole lot, is one of the most egregious: video game companies will hire people for salaried positions, then use social pressure to make them work 100-hour weeks instead of the 40-hour they should. But, on the other hand, these are still white-collar workers who are earning six-figure salaries, have health care and comfortable work environments. The layoffs always sound terrible, but I was surprised by how many times workers describe the process as being "very kind" or "generous", with many months of severance and active assistance in locating new positions; lots of workers don't get that sort of support. (Interestingly, this is one area where the hated Electronic Arts seems to actually be really good, while cool independent studios seem far more likely to ask people to continue working without pay after running out of cash and then leave them insolvent.) A lot of it seems to come down to just how badly people want to work in this industry: people who are very passionate about games will make sacrifices and accept a lot of turmoil in order to do it. Which sounds positive - passionate people love working on games! - but seems to be cynically exploited by studios to treat workers less well. Late in the book we hear about several people who leave the video game industry entirely and take other jobs, like writing entertainment software: inevitably, their new jobs pay better and are infinitely less stressful than their previous ones. Of course, all workers should be treated well, and video game developers winning more rights shouldn't take anything away from lower-status jobs.

It's kind of astonishing just how many ways games can go wrong. Like a lot people who play games, I've had a relatively simplistic view of it: "Someone wants to make a game, everything is going great, and then the big mean publisher comes in and tells them that they need to ship it before it's ready, and it all turns out terrible!" Well, that is one thing that does happen. But, at least judging by Press Reset, the publisher is just as likely to save a game as destroy it. BioShock Infinite, for one, was careering towards disaster thanks to the mercurial management of Ken Levine, before 2K sent in someone to make the company focus on finishing the game. And yes, there are lots of times that the publishers are pushing hard for something that financially benefits them. Gamestop looms large in the second half of this book: thanks to a rise in reselling physical media of games, sales correspondingly dip, which makes EA and other publishers demand content that can't be resold: an online pass or multiplayer code or something, which harried developers then try to graft into the game they want to make. Or the publishers decide that traditional games are dead, and now everything has to be an MMORPG. Or a mobile game. Or a MOBA. Or whatever the flavor of the week is.

There are two particularly crushing and bizarre stories. One is The Bureau: XCOM Unclassified. Everyone who was working on this game hated it: it was started at a separate studio, then transferred over to 2K Marin. It wasn't their idea, and it wasn't a genre or style of game that they had experience in making. And, as gamers, they were baffled as to why it existed: XCOM was a turn-based strategy game, with a recent and well-loved update from Firaxis, so who the hell was the market for an XCOM third-person shooter? After reading a book about publishers killing off cool and promising new games, it's beyond strange to keep reading about how 2K would not kill off this game that was behind schedule, over budget, and that everybody (everybody) hated.

Similarly, Mythic Entertainment's mobile version of Dungeon Keeper is bizarre, but probably even more sad. The studio was trying to stay afloat after losing money for many years, and pushed by EA to make a mobile came to compete against Clash Of Clans. (I am not at all tuned into the mobile game market, and was stunned to learn that Clash Of Clans has earned over six billion dollars!) The OG Dungeon Keeper was a beloved 90s icon, a quirky and subversive strategy game from the legendary developer Bullfrog. At first this seemed like a good candidate for a mobile game: the mouse point-and-click interface of the original game translates well to tapping on a screen. But EA kept demanding more and more monetization, in the form of inserting timers that would slow down the game unless players paid real money to skip. The developers recounting the experience come off as glum and defeated: it didn't feel fun to play with all those timers, but, after all, EA is the one with all the market research and business insight, so they must know something that Mythic didn't. And the game came out, and Mythic was flayed alive for ruining a classic. Which seems so much more heartbreaking than making something you were proud of and enjoyed that didn't get to see the light of day.

There's a lot more good stuff in this book, but those were the stories that most grabbed my attention. The book does reinforce my preexisting relief at not working in the games industry: I'm sure that there are a lot of joys, but it does not seem to treat people well. And the increasing specialization of the field seems like a turn-off: thanks to the scope and technical fidelity of the latest games, someone can't just be, like, an artist in a game, they need to specialize on specifically lighting scenes, or just defining joints, or just animating movements or something. It's good to see smaller indie studios still carrying that torch, though. I'll be curious to see where the industry moves in the future. In some ways it seems like AAA games are already slowing down: we've gone nine years without a new Grand Theft Auto and eleven years without a new Elder Scrolls. I'm curious if we'll see an increasing bifurcation, with expensive and rare tent-pole games along with a steady tricky of $10-$30 small-studio efforts, or a recalibration of the industry that pays appropriately for appropriate labor and changes its product as a result. That's what the title "Press Reset" ultimately refers to: given that the current situation is so bad, could we start over and do things differently in the future?