Evil Weekly (9/26/2022)
An interesting forum post shared with me by @postpostpostr. It’s a critical reading of Sudhir Venkatesh’s book Gang Leader for a Day, about a UChicago grad student who crosses the street from his department— this is literally true— and conducts an ethnographic analysis of a Chicagoland gang. I read it around the time it was published and wouldn’t be able to recall many specifics, other than that it outlines some of the gang’s organizational features. (I do remember that the gang leader nets a low six-figure income but the organization thrives mainly on low-level operatives who perceive their miniature world as a wholly natural, automatic state of affairs).
Anyway, the forum post discusses what Venkatesh leaves implicit: that the Chicago gang functions essentially as a State. It’s more surprising that a social scientist didn’t quickly come to this conclusion: Tilly famously called the State a ‘stationary bandit’, plenty of scholars like David Skarbek have described the emergence of state-like functions in prisons and gangs, and other illicit organizations like ‘terrorist groups’ are often known to become states-within-a-state where the de jure state has retreated. Perhaps the realization that the State and criminal enterprises are not very different is now verboten in the social sciences. (Naturally, this happens by omission. Why, after decades at the forefront of social science did analysis of the State fall out of favor? Surely it’s because there’s nothing left to learn here).
But, an even more interesting subtext here: the United States is no longer a State in the Weberian sense. That definition was always and deliberately an ideal-typical construction. And yet, is the US moving closer to or further away from that ideal type? It’s ambiguous, as captured in the concept of ‘anarcho-tyranny.’ In many respects the US inches closer to totalitarianism, with its massive party-state apparatus that interpenetrates more of society every day, its surveillance against ordinary citizens, its capricious (okay, it’s ‘logical’) application of the Law. At the same time we have hundreds of little microstates opening up, whether it’s Chicago gangs, neo-hippie communes in Washington, or anarchy in Jackson.
I will say this: our political concepts require constant revision because social reality quickly adapts to them. Even ‘totalitarian states’ have pockets of lawlessness to whom violence is outsourced or that at least forms an uneasy alliance with the state. Nazis had the SS, einsatzgruppen, and police reserves (who did most of the killing in the Holocaust, even without direct command); neither Saddam nor Assad could much control the paramilitaries they nominally operated; China contracts out violence to ‘gangs for hire’, which vaguely smells of an institutional legacy from the Cultural Revolution. What’s easier for a state than to extend bureaucratic control over every square-inch of its territory? Alliance, penetration into society through its vassal states. Maybe it’s worth mentioning that Chicagoland would collapse without direct payments to its inhabitants.
My latest short idea, added to my portfolio last Thursday, is Wolfspeed, Inc., with the following thesis:
WOLF is too cool a stock name. Tickers ought to be boring, like R or QFTZ. Cool tickers are liable to become hyper-inflated during a speculative run as inexperienced, well, speculators crowd into legible tickers and invent neat narratives about their potential as confirmation bias that their minds don’t really operate at the level of a fourteen year old anime consumer. Why not WLF? It was available, so I can only imagine that the marketing team that shouldn’t exist was in on the operation.
It’s a cash Holocaust. If the Nazis burnt Jews like WOLF burns cash, today there would be no State of Israel or New York City. Q4 alone saw a loss of 6 million USD, and it suffered a negative adj. EBITDA for yet another quarter with no sign of reversal on the horizon.
It currently trades as nearly 20 times revenue, in contrast to most of its peers in the semiconductors space. (Unless there’s good insider information to the contrary, there’s no good reason for this performance).
Earlier this year was a grand opportunity to short; you could make 30 to 70 percent returns betting against just about any overinflated tech stock, particularly software and de-SPAC’ed listings. That time has passed. Even as I am convinced the market will continue its decline till at least the end of 2023, I expect volatility to wreck traders jumping too late onto the inverted bandwagon. My strategy for the next few months will be to concentrate my short positions in several stocks I expect to have the greatest downside risk. WOLF recently climbed back to the neighborhood of its speculative height without any improvement in its underlying fundamentals, which makes it a good candidate for a dramatic reversal. It is possible WOLF launches another 100 percent as a new deal to power the Apple IDildo is unveiled, but I doubt it.
Albuquerque will probably be (at least based on the timing of this news article) the latest city to beat its record number of homicides after a record-setting 2021. A stabbing turned into its 100th homicide, putting it on track to surpass the 119 it had last year. Social scientists are apparently uninterested, occupied as they are with important questions like ‘are women interrupted more on a per-second basis in faculty meetings?’, and the FBI stopped trying to gather and report data on homicides altogether. I wonder when data collection and social science will be taken over by private companies. It wouldn’t be hard— you could do it with a team of ten to twenty— to create a ‘Google Society’ that aggregates all the relevant data and does the analysis. Perhaps the money isn’t there. But it would be a good public service: that the data aren’t accessible to you isn’t an indication of greater privacy, simply that the government will conceal the data that it still has and uses to inform its policies. (Get ready for “where’s your data for that?” in response to obvious claims to accelerate as the data becomes more restricted).
I previously typologized ‘data work’ into three roles: Data Astrologer, Data Janitor, and Data Nigger. I’ll return to this another time, but this week I received three DMs from people pursuing a career as a Data Analyst. My understanding of this role, which is limited, is that it’s a distinct category that falls under business operations or accounting rather than ‘data work’ proper, although I am happy to be corrected on this matter. It seems positions of this type combine elements of the aforementioned: Data Astrologer, in that you’ll often be asked to weave together fanciful narratives based on divination with the ancient art of barcharting; Data Janitor, in that you’ll often be tasked with plumbing databases; and Data Nigger, in that an anointed MBA will make you prune the wild fields of Excel spreadsheets. Sometimes, from what I can tell, the Data Analyst is primarily tasked with creating ‘dashboards,’ which I understand to be clickable horoscopes (hence, a Junior Data Astrologer not yet permitted the sacred weapon called the neural network).
I have no particular recommendations on how to pursue this job, other than to communicate skills in SQL, Excel, PowerBI, and BS, and to apply relentlessly. My impression is a common trajectory here is to job hop into a management role (e.g., ‘VP of Data Analytics’), although the job market may be constrained enough in the upcoming years to make aggressive hopping ill-advised.
I did find this figure from Robert Half’s Technology Salary Guide:
A national midpoint (by which they mean ‘median’— good to see they have confidence in aspiring tech workers) of 106,500 is not bad considering the skills involved. To compare, their quoted median for ‘Data Scientists’ is 135,000. If you’re looking to break into ‘data work,’ then a Data Analyst position is probably a reasonable pursuit if you have an undergraduate degree only and/or are not interested in software engineering or graduate level statistics. Outside a few top companies and coastal CA, the salary distribution is pretty flat across these occupations (my guess, almost everyone falls between 90 and 180k until they move into management). Landing a position in the salary tails is probably as much a function of hustle as skills— I’ve encountered many an autist happily but unknowingly underpaid by hundreds of thousands of dollars.
To put it another way, what’s the easiest path in these roles to 200k per year in middle-city USA? A Data Analyst’s best bet, again from my outsider view, is to climb into mid-level managerial roles centered around data reporting and its communication to exectives, often with a focus on company decision-making (e.g., based on wisdom from the barchart Gods, how can we increase sales?) A Data Scientist should be able to achieve this salary with a few year’s experience in a solely technical role (advanced astrology— statistics, experiments, machine learning, whatever), although running a small team normally helps. Data Janitors, the data / ML engineers, are hopefully earning this salary immediately.
One thing to consider if you plan to become a ‘Data Analyst’ is path dependence. You risk walking into a plateau if you lack the disposition or inclination for business management, since there are only so many ways to advance one’s barchart engineering skills. (And, obviously, the best option if you’re business inclined is to create your own business). It is not impossible to transition to more technically sophisticated roles— don’t let me or anyone else stifle your ambition— but it is a fact of reality that most people self-educate poorly and worse the more difficult the subject. More common is the reverse. In my first job, at a large yet boring fortune 100 company, half the newly hired data scientists grew frustrated figuring out GLM, or tensorflow, or whatever and promptly self-sorted into the Data Analytics team. No harm in just wanting to grill, especially in one’s job, but know thyself, nigga.
Related: barchart science from the BLS that came to my attention this week. Projected job growth rate over 30 percent for both statisticians and data scientists (which are, in many cases, identical jobs). Good news for us, and while I already extolled the virtues of the ‘just grill’ lifestyle of the ‘data analyst,’ the reality is you should grow up and learn difficult subjects if you want to secure your income and job stability. I am not entirely sure how the BLS comes up with its job title designations, though, since these median salary estimates seem appropriate at the entry level (you should expect a salary of 100k, give or take ten or twenty thousand dollars, your first year as a data scientist). Perhaps they have a narrow operational measure of ‘data scientist’ that effectively excludes people with multiple years of experience. Also, look at the projected job growth on restaurant cooks. I hear Mariachi.
Had a good conversation with a friend who I convinced to read Hermann Hesse’s novella, Demian. Here’s a spoiler if and only if you’ve read the book or never intend to: at the end the narrator has died (in WWI) and awoken in Hell. (I have some issues with this author’s interpretation but on the fundamental conclusion I have no disagreements). Hesse is one of the many authors whose work has been whitewashed to scrub of all Christian influence. Think about what the ending means, why it doesn’t occur to contemporary readers, and what this says of the impoverished state of internal worlds stripped of historical-symbolic meaning.
More from my reading this week. Between this and House of Leaves, 1999/2000 may have been the last year in recent memory for exceptional literature (cinema, as a more popular medium, seemed to stagger on a little while longer). What’s the last thousand-page novel since that doesn’t make you want to jam a fork in your eyes? (No dragons or warlocks).
Tweet from the Big Autist. Did WFH increase the absolute number of jobs? More likely it disclosed redundancy in many sectors, where a large segment of the white collar workforce exists solely to create meetings and encumber productive people (if not worse). How to reconcile this claim with the projected growth in data scientists is not obvious, as a large portion of that workforce won’t return to the office absent an apocalyptic scenario (which I won’t entirely rule out).
I suspect more tech-adjacent roles will remain remote but the bottom 20 percent of performers will be culled and re-allocated to other work. One problem is the growth of borderline-fraudulent ‘data science’ Master’s degree programs whose graduates cannot calculate a t-test or code a Python class. Many of the graduates will flame out of intensive work and become SQL-monkies or transfer to on-site business roles that require little math or coding— an inversion of the migration into data science. Remember, skills are scarcer than titles (and no, the universities will never be held responsible for their crimes).
Another relevant subgroup: ex-academics who become data scientists. Many companies offer research roles and should, in fact, create more (I have seen an inability to evaluate simple interventions or field experiments cost major companies millions of dollars). But the supply probably won’t be sufficient to meet growing demand for these jobs. Unemployment and underemployment of these academics is likely too low and will be exacerbated by the unending production of PhDs, dwindling government jobs, and decrease in college enrollment at the undergraduate level. Prepare accordingly, lest you end up with a 4-4 teaching load for 40k at the University of Northwest South Dakota.
Aristotle believed thoughts are formed in the heart. You know that experience you have of thoughts taking place inside your head? Yeah.
I saw this tweet right as I was about to click ‘Publish.’ It displays with beautiful clarity that, for many more people than you realize, society is a magical system of automatic wish fulfillment. Can you really blame Skinner for his otherwise offensive behaviorism? Are these free people? Do you want them to vote? All the passive-aggressive, behavioral-economic death-by-one-thousand nudges totalitarianism that is so preposterous to free men is downstream of widespread enfranchisement.