Shared posts

09 Oct 06:04

Habt ihr eure geschäftskritischen Daten in einer Cloud ...

Habt ihr eure geschäftskritischen Daten in einer Cloud in den USA gespeichert? Die hier auch! Ganz großartige Idee, wie sich völlig überraschend rausstellt.

Sind schon voll vorteilhaft, diese Abo-Lizenzmodelle, wo einem die Werkzeuge gar nicht mehr gehören, die man so benutzt.

18 Oct 19:36

Heute morgen saß ich in einem Kubernetes+Docker-Vortrag. ...

Heute morgen saß ich in einem Kubernetes+Docker-Vortrag. Ich beobachte ja immer gerne, wie sich Felder entwickeln. Und Docker und Kubernetes hat sich irgendwie nicht so viel bewegt, seit ich das letzte Mal geguckt habe. Wobei, das stimmt nicht, es gibt jetzt noch mehr Layer und Produkte, die empfohlen werden, um Probleme zu lösen, die man vorher (also vor Ausrollen von Docker und Kubernetes) noch gar nicht hatte.

Mir ist ja nach wie vor unklar, wieso Container als Methode zum Downloaden und sicheren Ausführen von Fremdsoftware aus dem Internet verkauft wird. Wenn man nicht gerade ein Cloud-Provider ist: Wie viele Szenarien beinhalten denn bitte, dass man unvertrauenswürdige Software ausführt? Wenn die Software nicht vertrauenwürdig ist, dann führt man sie nicht aus. Oder man lebt halt damit, dass die Software Mist machen kann, und lässt sie nicht an die Datenbank mit den Payroll-Daten oder das Corpnet ran.

Ich finde, Container muss man eher als Deployment-Technologie betrachten. Für Software, die man im Moment per # ./install deployed, oder per RPM oder DEB. Da schiebt man jetzt halt nicht ein RPM mit Dependencies herum, sondern aus dem Buildsystem fällt ein Image mit allen Abhängigkeiten, und die schiebt man auf das Zielsystem. Wenn das Ärger macht, dann fährt man das vorherige Image nochmal hoch und hat nur ein paar Sekunden Downtime. Das ist ein Versprechen, das Container grundsätzlich einlösen können, und da sind sie auch nicht schlechter als irgendein gammeliges manuell hochgepatches (wenn überhaupt) Redhat mit handinstallierter Anwendung — sondern haben sogar echte Vorteile.

Die ganze Idee hinter Containern ist dann, dass man beim Deployment agiler wird, und Fehler nicht so schlimm sind, weil man das alte Image noch hat. Dann muss man aber natürlich auch agil sein und seine Images häufiger neu bauen, insbesondere auch mit den aktualisierten Abhängigkeiten frisch. Warum das keiner macht, ist aber eine Frage für einen anderen Blogrant :-)

30 Jan 09:36

Staying Classy

by Scott Alexander

Siderea writes an essay on class in America. You should read it. In case you don’t, here’s the summary:

1. People tend to confuse social class with economic class, eg how much money you make. But social class is a more complicated idea involving how respectable you seem, how educated you are, and what kind of family you come from. An assembly line supervisor might make the same amount of money as a schoolteacher, but the schoolteacher would probably seem more refined and be able to access better social circles.

2. Classes are cultures. People in a certain class have their own way of dressing, speaking, decorating, and behaving. They have distinctive ideas and values. This is why a lower-class person cannot simply claim to be upper-class and so gain all the benefits of upper-class-hood; it would be as hard as trying to pass for Japanese. Lower-class people can learn their way around upper-class culture, but it’s a difficult and lifelong project done most easily if you already have upper-class resources.

3. Talking about class is taboo because we like to believe we’re a classless society. We talk about income instead and pretend it’s class. Class breaks through in a couple of phrases like “rednecks” or “white trash” or “white collar” or “coastal elites”, but people use the phrases without usually having a broader idea that it’s class they’re talking about.

4. Class prejudice is complicated. It combines the practical superiority of being upper-class to being lower-class (because you have more money and opportunity) with the very dubious value judgment that upper-class culture is superior to lower-class culture, or that lower-class culture is just people trying to do upper-class culture but failing. But lower-class people like lower-class culture and generally do not want to adopt upper-class culture, except insofar as it’s necessary to advance. Analogies to race and assimilation are obvious.

5. People mostly understand their own class, and the class one step above or below them, but have only vague stereotypes of classes further than that. This limits social mobility; you can’t join what you can’t understand.

6. College is a finishing school for the upper classes. They send their children there to learn the proper upper class values and behaviors. Even if community college does a great job teaching whatever trades it teaches, it will not teach you how to be a part of the upper class, and this will seriously limit your opportunities.

7. Politically, the left pretends class doesn’t exist; the right talks about it, but only to yell at the underclass and say that their culture is wrong. Race is really complicated and will be left out of this analysis.

I notice Siderea is a psychotherapist, which doesn’t surprise me. We in mental health get a pretty good cross-sectional exposure of everybody and get to hear about their lives, and with enough data points the structure comes into sharper relief.

Just to give an example: suppose a lady comes in with really over-permed dyed curly hair wearing several rings, bracelets, and necklaces. Her name is Sherri and she calls you “darling”; she’s also carrying her lunch, which is KFC plus a Big Gulp. Without knowing anything else about her, you can peg her as working class. Maybe she won the lottery ten years ago and is now the richest person in your state. It doesn’t matter. She’s still working class.

Or suppose a thin 25-year-old man comes in wearing glasses, a small close-cropped beard, and a Led Zeppelin t-shirt. His name is Alex and he apologizes for being three minutes late. This guy is probably middle-to-upper-middle-class and college educated, maybe not a great college but still college-educated. And maybe he’s fallen on hard times and doesn’t have a dollar to his name. It still doesn’t matter. He’s still middle-to-upper-middle class.

And you start to learn you can predict things about these people, the concerns they’re going to have, the kind of things that happen to them. Who their friends are. How they relate to their friends: Sherri will expound upon the flaws of every single one of her ungrateful coworkers; Alex will reluctantly say he went through a tough breakup a year or two ago. What kind of drugs they abuse, if they abuse drugs (maybe Sherri has smoking and drinking problems; Alex has probably tried marijuana and LSD but is embarrassed to say so).

But this kind of innate stereotyping is different than a formal taxonomy. Siderea links to Michael Church’s attempt to explain what the classes actually are. This is another piece you should read, but again in case you don’t:

1. 10% of people are in an underclass consisting of “generationally poor” people who may never have held jobs and who come from similarly poor families.

2. 65% of people are in the labor class. They work jobs where labor is seen as a commodity, ie there’s not as much sense of career capital or reputation. They base virtue and success around Hard Work. Its lower levels are minimum wage McJobs, its middle levels are assembly line work, and its higher levels are things like pilots, plumbers, and small business owners. The stratospheric semi-divine level is “celebrities” like reality TV stars who become fabulously rich and famous while sticking to their labor class roots.

3. 23.5% of people are in the gentry class. They fetishize education and career capital. They engage in all sorts of signaling games around “fair trade” and “organic” and what museums they go to. At the lower level they’re schoolteachers and starving artists, at the mid level they’re “professions” like engineering and law, and at the highest level they’re professors and scientists and entrepreneurs. The stratospheric semi-divine level is “cultural influencers” like Jon Stewart or Steven Pinker who become famous and (maybe) rich while sticking to their gentry class roots.

4. 1.5% of people are in the elite class. Although you can be borderline-elite by getting a job in finance and making a few million, the real elite are born into money and don’t work unless they want to. Occasionally they’ll sit on a board or found a philanthropic association or something. They don’t believe in “professional achievement” because working is lower-class; they might compete in complicated status games around who throws the best parties or has the best horses or whatever.

5. The highest class (E1) are insane psychopaths who burn the global commons for shits and giggles. They tend to be drug lords, arms dealers, and morally insane billionaires. Most famous politicians and businesspeople are not of this class and most people in this class are not famous.

6. The three main classes (labor, gentry, and elite) are three different ‘infrastructures’. To be in labor you need skills, to be in gentry you need education, and to be in elite you need connections. There’s no strict hierarchy (eg not all gentry are above all labor), but you can picture them as offset ladders, with the lower gentry being at the same rung as the higher labor and so on.

7. The Elite control everything; the constant threat is that Gentry and Labor will unite against them, which might very well work. The Elite neutralize this threat by making Labor hate Gentry as “effeminate” or “pretentious”; they also convince Labor that the Gentry are probably secretly in cahoots with the underclass against Labor. Elites also convince Labor that Elites don’t exist and it’s Gentry all the way up, which means that “anti-1%” sentiment, which should properly get Labor and Gentry to cooperate against the Elites, instead makes Gentry hate the Elites but Labor hate Gentry. Politics boils down to Gentry being good people trying to improve things, and Elite conning Labor into hating Gentry to prevent things from being improved.

8. While all classes can have good and bad people (except E1, which is wholly bad), Elites have a generally negative influence on society, and Gentry are generally positive. After the World Wars, everybody got angry at the Elites for all the war and killing and stuff, which convinced them to lie low for a few decades and forced the Gentry to take over. This was why the country did so well during the 50s and 60s. Whether the country goes in a good or bad direction now depends on whether the Elites manage to take it back or not. One reason Silicon Valley works (used to work?) so unusually well was that it was mostly a native project of the Gentry that hadn’t yet been infiltrated by the Elites.

Reaction to Church on the subreddit was pretty negative, but I find it at least a good nucleus for further discussion. The Gentry/Labor distinction is glaringly obvious. The Labor/Underclass distinction also seems glaringly obvious to me, if only because Labor hates the underclass. The Gentry/Elite distinction doesn’t seem glaringly obvious to me, but maybe that’s just because I haven’t met enough elites. In particular, Church’s “E1” seems caricatured and out-of-place in his otherwise sober analysis. Then again, if those people existed I probably wouldn’t know anyway. Then again, the rest of Church’s blog suggests some paranoid tendencies, so maybe the E1 entry is just those coming out.

Siderea notes that Church’s analysis independently reached about the same conclusion as Paul Fussell’s famous guide. I’m not entirely sure how you’d judge this (everybody’s going to include lower, middle, and upper classes), but eyeballing Fussell it does look a lot like Church, so let’s grant this.

It also doesn’t sound too different from Marx. Elites sound like capitalists, Gentry like bourgeoisie, Labor like the proletariat, and the Underclass like the lumpenproletariat. Or maybe I’m making up patterns where they don’t exist; why should the class system of 21st century America be the same as that of 19th century industrial Europe?

There’s one more discussion of class I remember being influenced by, and that’s Unqualified Reservations’ Castes of the United States. Another one that you should read but that I’ll summarize in case you don’t:

1. Dalits are the underclass, made up of homeless people, chronically unemployed people, drug addicts, etc. They tend to have a lot of trouble with the law, go in and out of jail, never really hold down stable employment. Status is “street cred” that you get from being powerful, wealthy, and sexually successful, eg gang leaders.

2. Vaisyas are standard middle-class people who engage in productive employment. They tend to form nuclear families and try to go to church. Status is having a stable job, a stable family, and being well-liked in your church or social club.

3. Brahmins are very educated people who participate in the world of ideas. They range from doctors and lawyers to artists and professors. Access is conferred by top-tier university education. Status is from conspicuous engagement in progressive politics, eg being an activist, working for an NGO, “campaigning for justice”. They are “the ruling class”.

4. Optimates are very rich WASPs concerned with breeding and old money. Status comes from breeding and an antiquated idea of “nobility”. Optimates used to be “the ruling class”, but now they’re either extinct or endangered, having been pretty much absorbed into the Brahmins.

5. Mentioned elsewhere in the UR corpus: politics boils down to Vaisyas being basically decent people trying to lead normal productive lives, and Brahmins trying to create a vast tentacled monstrosity of useless bureaucrats and petty enforcers of ideological conformity to employ Brahmins in the “knowledge work” they feel entitled to and to protect their interests. Silicon Valley is (used to be?) unusually functional because it maintained some Vaisya values separate from the corrupting influence of the Brahmins.

Michael Church’s system (henceforth MC) and the Unqualified Reservation system (henceforth UR) are similar in some ways. MC’s Underclass matches Dalits, MC’s Labor matches Vaisyas, MC’s Gentry matches Brahmins, and MC’s Elite matches Optimates. This is a promising start. It’s a fourth independent pair of eyes that’s found the same thing as all the others. (commenters bring up Joel Kotkin and Archdruid Report as similar convergent perspectives).

But there are also some profound differences. UR says that the Elites are mostly gone, that everything’s ruled by the Gentry nowadays, and that the Gentry are allying with the criminal Underclass against Labor. MC mentions this same picture, but only as the false facade that the Elites are trying to get everyone else to believe in order to keep them divided.

You could reconcile some of the differences by supposing the two models have different cutoffs. Suppose we rank people from 0 (lowest underclass) to 100 (highest elite). Maybe MC draws the Labor/Gentry and Gentry/Elite borders at 40 and 70 respectively, and UR draws the Vaisya/Brahmin and Brahmin/Optimate borders at 60 and 90. If the world’s being run by 80s, MC could be right to say it’s run by Elites and not Gentry, and UR could be right in saying it’s run by Brahmins and not Optimates. If Silicon Valley is run by 55s but being ruined by 75s, MC could say it’s run by Gentry but ruined by elites, and UR could say it’s run by Vaisyas but ruined by Brahmins. But if there’s this much variability in class boundaries, what’s the point in even drawing them in the first place?

But I think the differences are real and political: MC comes from a liberal perspective, UR from a conservative one. MC wants to locate the source of the cancer in the (mostly plutocrat) Elites, cast the (mostly liberal) Gentry as wonderful people who can do no wrong, cast the (mostly conservative) Labor as deluded and paranoid, and cast the (liberal-aligned) Underclass in a sympathetic light. UR wants to locate the source of the cancer in the (mostly liberal) Brahmins, cast the (mostly conservative) Labor as decent salt-of-the-Earth types under threat from the elite, and cast the (liberal-aligned) Underclass in an unsympathetic light.

And the political angle evokes one more system worth adding here: my own discussion of the Blue Tribe vs. the Red Tribe in I Can Tolerate Anything But The Outgroup. I point out that the group sometimes referred to as “coastal liberals” or “SWPL” and so on are marked not only by Democratic Party beliefs, but by a host of cultural similarities including food, dress, music, hobbies, religion, values, art, etc. Likewise, the group sometimes referred to as “rednecks” or “fundies” and so on are marked not only by Republican Party beliefs, but by a similar set of cultural similarities. I call these the “Blue Tribe” and the “Red Tribe” as an attempt to distinguish them as cultures and not just as sets of political beliefs.

These tribes seem closely related to classes. “Blue Tribe” is similar to Gentry; “Red Tribe” is similar to Labor. I won’t say there’s a perfect 1:1 equivalence; for example, I know some union leaders who are very clearly in the Labor class but who wouldn’t be caught dead in the Red Tribe. But the resemblance is too close to miss.

Some final scattered thoughts:

1. All those studies that analyze whether some variable or other affects income? They’d all be much more interesting if they analyzed the effect on class instead. For example, there’s a surprisingly low correlation between your parents’ income and your own income, which sounds like it means there’s high social mobility. But I grew up in a Gentry class family; I became a doctor, my brother became a musician, and my cousin got a law degree but eventually decided to work very irregularly and mostly stay home raising her children. I make more money than my brother, and we both make more money than my cousin, but this is not a victory for social mobility and family non-determinism; it’s no coincidence none of us ended up as farmers or factory workers. We all ended up Gentry class, but I chose something closer to the maximize-income part of the Gentry class tradeoff space, my brother chose something closer to the maximize-creativity part, and my cousin chose to raise the next generation. Any studies that interpret our income difference as an outcome difference and tries to analyze what factors gave me a leg up over my relatives (better schools? more breastfeeding as a child?) are stupid and will come up with random noise. We all got approximately the same level of success/opportunity, and those things just happen to be very poorly measured by money. If we could somehow collapse the entirety of tradeoffspace into a single variable, I bet it would have a far greater parent-child correlation than income does. This is part of why I don’t follow the people who take the modest effect of IQ on income as a sign that IQ doesn’t change your opportunities much; maybe everyone in my family has similar IQs but wildly different income levels, and there’s your merely modest IQ/income relationship right there. I think some studies (especially in Britain) have tried analyzing class and gotten some gains over analyzing income, but I don’t know much about this.

2. I think Siderea is right that the Right thinks in social class terms more naturally than the Left. To oversimplify, both sides use class warfare, but the Left’s class warfare is economic (“the plutocrat billionaires are ruining everything!”) and the Right’s class warfare is social (“the media and academic elites are ruining everything!”).

3. Closely related: Donald Trump appeals to a lot of people because despite his immense wealth he practically glows with signs of being Labor class. This isn’t surprising; his grandfather was a barber and his father clawed his way up to the top by getting his hands dirty. He himself went to a medium-tier college and is probably closer in spirit to the small-business owners of the upper Labor class than to the Stanford MBA-holding executives of the Elite. Trump loves and participates in professional wrestling and reality television; those definitely aren’t Gentry or Elites pastimes! When liberals shake their heads wondering why Joe Sixpack feels like Trump is a kindred soul even though Trump’s been a billionaire his whole life, they’re falling into the liberal habit of sorting people by wealth instead of by class. To Joe Sixpack, Trump is “local boy made good”.

4. The thesis of “I Can Tolerate Anything But The Outgroup” simplifies to “It is a Gentry-class tradition to sweep aside all prejudices except class prejudice, which must be held with the intensity of all the old prejudices combined.”

5. But “I Can Tolerate Anything But The Outgroup”‘s Grey Tribe sits uneasy within this system. It doesn’t seem to be a class. But it also seems distinctly different from ordinary Gentry norms. And what about minorities? What about the differences between farmers vs. factory workers? If different classes are equivalent to different cultures, well, there are a lot of different cultures that don’t fit easily into the hierarchy. Maybe class is one factor among many that can create a different culture, but other factors can be stronger than class in some groups?

6. Siderea doesn’t want to get into how race interacts with class, and that seems wise. But a related digression: lots of people complain about social justice being classist, in that it’s hard for anybody who hasn’t either gone to college or at least spent a lot of time hanging around social justice people to keep track of which words, opinions, and causes are okay versus will render you radioactive. On the one hand, this is probably true. On the other, it’s probably true of everything, with social justice as an unexceptional example. Yes, the way you refer to trans people shows what class you’re from, but so does the way you order ice cream.

7. Siderea admits she is classist and not ashamed of this. I have a hard time understanding what she means, but I can try to explain my own classism: I think classes probably sort on important qualities and reinforce those qualities. For example, the Underclass and Labor class people I know are much more likely to have high-conflict styles of interaction: if they feel offended, they’ll yell at you and maybe even fight you. Gentry class people would be horrified at the thought; they might respond to the same offense by filing a complaint with Human Resources. I think there are two equally correct ways to interpret this. Number one, people with the maladaptive behavior of starting physical fights don’t make it very far in life and so end out in lower classes, and insofar as these behaviors are either genetic or learned within the family, their families stay in lower classes throughout the generations. Number two, the lower classes have a culture where you defend your honor by fighting people who offend you, and the upper classes have a culture where you defend your honor by submitting complaints, and although in a cosmic sense both of those styles are equally valid, and although indeed a thousand years ago the fighting might have been more adaptive, in today’s society the complaint-submitting is more adaptive and the lower classes are screwed unless they unlearn that behavior – which they probably won’t, because unlearning class is hard. But this means that classism is at least kind of justified – if you want to hire for example a schoolteacher, you might want to look for people who show all the signs of Gentry rather than Labor class to make sure they’re not going to get into physical fights in the classroom.

8. Cellular automaton theory of fashion likely relevant.

9. Siderea’s idea of college as finishing school for the upper classes is interesting, and her own experience is a window into something I never thought about before. But I’m not sure how typical she is; I think most colleges admit students who are already members of the classes their graduates end up in. I felt like I didn’t learn any class culture during my own college experience at all – which isn’t surprising since I was born the son of a doctor and ended up as a doctor myself. I think my story’s probably more typical than Siderea’s, though other people can prove me wrong if they’ve seen differently.