There are still a lot of computer science jobs, but should CS education be just about jobs?

by Aline Normoyle

Programming jobs are presumably disappearing1 but our students are still getting jobs and more broadly, this belief is not supported by the data. In a recent article that analyses data from the U.S. Bureau of Labor Statistics directly, Christine Julien shows that although tech jobs have decreased since COVID, but the total number of jobs is still high.

Left, visualization of the change in jobs, compared to the same month in the previous year; Middle, a clearer visualization that shows the total number of jobs which is still high; Right, a visualization comparing tech-related jobs to pure tech jobs. Plots by Christine Julien, Stop saying computer science is dying, Inside Higher Ed, 2026.

Despite this, I wonder if disenfranchisement with computer science might be due to other factors. First, it’s not as easy to get a job straight out of college: there are fewer job openings for junior developers and the major tech companies have been laying people off. The job market also was bad after the dotcom bust and after 2008 crash. Cycles of hiring and firing have been a fact of American employment for decades — typically done as Jack Welch-style2 corporate restructuring aimed at short-term financial gains. We are in a period of cost cutting through minimal hiring. Secondly, the industry has largely abandoned the humanistic ideals from their founding. There may be less excitement about the types of software that is getting investment: it can feel like technology is making people’s lives worse. More expensive, more isolated, more chaotic, busier and more constrained. People want careers that provide meaning and make the world better.

For example, in the case of recent layoffs at id Software, who make Doom, half the studio was laid off despite it being a profitable and critically-acclaimed studio3. Maybe the layoffs are due to AI or to over-staffing during COVID or to changing corporate priorities or as retaliation for unionization attempts. It is impossible to know from official public announcements. Microsoft’s Amy Coleman, Chief People Officer, was careful to say that “the roles eliminated today are not being replaced by AI. At the same time, what is true is that AI is changing how work gets done.”4 The point I want to make is that in recent years, most layoffs are not about productivity and profits, but about the priorities of management. Management is fickle and particularly vulnerable to hype and FOMO.

AI is affecting employment5. Even when it is not directly replacing people, AI still uses investment capital that could be spent on other projects. Although in some cases, it is replacing people. An acquaintance at Comcast says he babysits AIs now, instead of the junior developers. At the recent SIGGRAPH, a movie director talked about how he normally would have to work with others to talk about ideas and storyboard films. When the purpose of the economy is to help strangers cooperate by exchanging goods and services, I don’t see the point. (On the other hand, if the purpose of the economy is to extract value from others’ labor and appropriate their expertise into a commodity that can be placed behind a paywall, it makes complete sense.)

So far, reality hasn’t caught up to upper management yet: AI (read: LLMS) is actually expensive and unreliable; it’s effect of productivity has yet to be shown; and so just like previous rounds of layoffs, I expect companies will start hiring junior developers again. In my 20s, there was a similar push to “outsource” all jobs to places with lower wages: there was still a demand for engineers in North America.

For students deciding what to study, computer science is a demanding subject with a steep learning curve at the beginning. I think it’s a good thing that students are not pursuing majors if they dislike the subject. However, if students do find computer science interesting, I hope they will persist despite our discouraging economic and political moment. Otherwise, we will never achieve the true potential of our computer technology. Technology facilitates storing and sharing knowledge, communication, self-expression, and creativity. I was a CS major because I liked making things and found the work interesting. So let’s not confuse the tech industry (especially Silicon Valley) with technology. Capital has decided the problems that software will solve and how. To rephrase Lewis Mumford: technology is suffering for the sins of capitalism; contrariwise, capitalism is taking credit for the virtues of technology6

And this brings me back to education: computer science education is not — and should never have been — about just getting jobs. Computer science is about understanding the information systems that underlie our society. And what a lot of people don’t realize is that computers are general purpose machines, capable of doing any task that you can express in terms of an algorithm. The hardware is like paper; the software is like writing. We don’t blame literacy when it is used for manipulation and misinformation. But like learning to read and write, understanding our modern technological systems — hardware, algorithms, software design, network architectures, etc — is difficult and formal education and mentorship helps people overcome the challenges of getting started. We do not have the accept a future where all human activity — shopping, communication, art, food — is mediated through their subscription services. But we can’t build this different future without expertise.

My hope is that the industry readjusts after the AI bubble bursts. We start using LLMs as a normal technology: small models that do not rely on exploiting people and destroying the environment. There is a market for software that do not surveil their users; that are reparable; that are customizable; that you can own rather than rent; that use little energy. Energetic, skilled and idealist engineers are needed to build the better technology of the future. Politics change. Politics can force industries to change. Education is part of how we imagine where we want to go and how we might get there.

  1. For example, Software Engineering Jobs Plummet to 5-year low ↩︎
  2. To learn more, listen to the podcast episode Jack Welch is why you were laid off, on Behind the Bastards. Jack Welch, who became CEO of GE in 1981, was famous for laying off workers at profitable branches of GE to boost shareholder value. He also popularized “rank and yank” employee practices where the bottom 10% of employees would be fired regardless of job performance. In the 1980s, Jack Welch would transition GE away from making real-world products and towards financial offerings such as loans and insurance. By the 2020s, GE would lose half its value as the financial portions of its business lost value. ↩︎
  3. Listen to the full conversation at Blood in the Machine, Xbox workers speak out about layoffs. ↩︎
  4. Read the full press release here. It’s as if Microsoft is trying to say two contradictory things at once: to investors Microsoft wants to signal they are at the forefront of AI while to users, Microsoft wants to signal that these changes are not affecting workers. ↩︎
  5. But if you would like to read more, I recommend Brian Merchant’s reporting on this topic from his substack Blood in the Machine. ↩︎
  6. Here is the original quote in context from Technics and Civilization by Lewis Mumford: “Capitalism utilized the machine, not to further social welfare, but to increase private profit: mechanical instruments were used for the aggrandizement of the ruling classes. It was because of capitalism that the handicraft industries of both Europe and other parts of the world were recklessly destroyed, even when the latter were inferior to the thing replaced:for the prestige of improvement and success and power was with the machine, even when it improved nothing, even when technically speaking it was a failure. It was because of the possibilities of profit that the place of the machine was overemphasized and the degree of regimentation push beyond what was necessary to harmony and efficiency. It was because of certain traits in private capitalism that the machine — which was a neutral agent — has often seemed, and in fact has sometimes been, a malicious element in society, careless of human life, indifferent to human interests. The machine has suffered for the sins of capitalism; contrariwise, capitalism has often taken credit for the virtues of the machine.” ↩︎

Cybernetics: Dan Davies on the science of management that almost was

by Aline Normoyle

In 1948, Norbert Weiner published a book called Cybernetics: Control and Communication in the Animal and the Machine. At the time, it was a best seller and the prefix cyber became eponymous with futuristic technologies. In fact, intellectuals associated with cybernetics were sought after management consultants and public intellectuals.

Jacket with embroidery

In fact, the 1971 Chilean government, under Salvatore Allende, even recruited a cyberneticist, Stanford Beer, to design a new, flexible, and resilient modern system of bureaucracy for Chile. The system was called Cybersym and it aimed to create a flexible administration that preserved the independence and autonomy of its citizens. These goals would be achieved through dense information networks with decision-making authority at every level in a hierarchy, not strictly top-down. Unfortunately, Cybersym never got past the design stage, as Pinochet overthrew the Chilean government in 1973, forcing Beer to flee. Beer would relocate to the UK where he remained an energetic intellectual, attempting to build a computer based on algae in a pond, studying art and poetry, holding lectures, writing, and conversing with other interesting people such as Brian Eno (!). But his career as a management consultant never recovered.

Computer-generated image of Project CyberSyn operations room. by Rama

In his book, The Unaccountability machine, Dan Davies defines cybernetics as “the study and control of systems”. Unlike economics, which historically relies on simplified models of human behavior to describe system effects (e.g. from the inside out), cybernetics instead tries to understand systems by observing its behaviors holistically (e.g. from the outside in). The cybernetic approach to understanding systems treats systems as a black-box: we can only understand the black box by studying how different inputs affect the output.

POSIWID

Large organizations are complex systems that consist of multiple, interacting box boxes. As Davies describes it, the defining characteristic of these systems was feedback, or “the cyclic and recursive use of the output of one black box as the input of another, then back again”. Feedback between many black boxes is what makes the high-level behaviors of these systems so hard to analyze and understand. An organization, such as a school, can consist entirely of people who care a lot about education, but the school system produces credentials (hopefully, but not necessarily tied to learning). Davies describes these systems as “unpredictable, but not random”. Beer summarized this phenomena — the tendency of organizations to produce outputs at odds with its mission — with the phrase “The purpose of a system is what it does”, or POSIWID. Davies gives the following examples.

the purpose of academic publishing is to generate citations for academics and profits for publishing houses; the individuals involved might want to expand and communicate human knowledge, but the system behaves in ways that support its goals, not theirs. The purpose of the pet hotel at Schiphol Airport was to shred squirrels; everyone involved may have cared about animal welfare, but the system had been set up to ensure compliance with import regulation at the lowest cost possible, so that’s what it did.

Homeostasis

The only way to determine the behavior of a complex system is to understand its steady states. In other words, how do the paths of least resistance within the organization combine together to create predictable and habitual responses? Whereas economics has the concept of ‘equilibrium’ where the system ‘stops’ because everyone has achieved an optimal state, cybernetics defined the concept of homeostasis. In homeostasis, the system continues to change, but does so in a way that produces predictable patterns — like the cycles of water or seasons.

Resource Bargains

Stanford Beer believed that good management, with the right structure, could allow large organizations to interact effectively with a complex world. Through Resource bargains, layers within the organization would have areas of responsibility within which they would have complete autonomy. Davies explains as follows.

Beer was trying to adapt hierarchical systems so they were fit for their purpose in a changing world. He saw a strict command-and-control approach as dangerously inflexible, while excessive delegation would destroy the organisation’s ability to act as a coherent system.

Davies uses a military platoon as an example.

It’s a concept that might be most easily illustrated by thinking of a military platoon that’s given an objective – to capture a bridge, say. There is a higher level of the hierarchy that is responsible for coming up with objectives, fitting them into an overall battle plan and making sure that different platoons don’t get in the way of each other. But once the orders are given, capturing the bridge is the responsibility of the platoon, using the soldiers and equipment assigned to the task. The general doesn’t keep on managing the attack; the next piece of communication he or she expects is either that the bargain has been kept (the bridge is captured), or that some external factor has interfered, and the resource bargain needs to be renegotiated by the survivors.

In this setup, both managers and workers have a reciprocal responsibility to each other: manage for providing clear objectives and organizing with other units; workers for getting tasks done. Contrast this with the ubiquitous command and control management structures we see today: call centers where employees can only follow a script and have no authority to help a customer. Their job is to act as an accountability sink — a place where for information to be ignored — or as David Graeber would describe it in Bullshit Jobs, as duck tapers, people who patch broken systems with temporary fixes that keep it hobbling along.

Kate Sekules on the joys of mending

by Aline Normoyle

Give me your tired, your torn, you huddled garments yearning to breathe free, the wretched refuse of your teeming drawer. – Emma Lazarus

Mend! by Kate Sekules is an ode to the joys of slow fashion, letting “clothes become unique to you, a conversation starter, each mend a potential story, like a scar, tied to a story about how the garment became ripped or torn.”

Jacket with embroidery

To be honest, when I was growing up, mending was chore. But when we produce so much waste and are encouraged to buy so much stuff, mending feels subversive. Mending brings us back to a time when crafts were valuable because they were hand-made. Medieval nobility would include garments in their wills. A garment in the 1700s would take nearly have a year to make. Sekules’s book has this account from a girl living in America during the 1790s.

By the 1790s, rural life was still arduous. Elizabeth Fuller, fifteen, of Princeton, Massachusetts, outlined the schedule for making “Pa’s surtout”: Picking wool: three days; breaking the wool: three days; spinning: twenty-four days (by Ma). Then it was sent out to be dyed blue. A woman came in to warp the loom (everyone had one). Elizabeth wove it: two weeks. Then “Mr Deadman” fulled the cloth and Ma cut out and sewed the coat. It took from February to June.

Despite the fact that garments are (artificially) cheap to make, there are many good reasons to mend: the fashion industry is wasteful and exploitative; it makes your style unique and distinctive; it resists commodification and encourages you to have fewer clothes (save money); like veganism and vegetarianism, it can be a protest against an exploitative and wasteful industry; you feel more ownership and agency and meaning when you create the things you use; it’s fun and creative. We often say we take care of the things we love; but it’s impossible not to love the things you care for.

There is a feeling of awe around artifacts of the past that were created with a goal of durability, a desire to last forever. Most garments have not survived, save the occasional 5000 year old dress or the bronze age shirt/skirt outfit of the Egtved Girl. Natural fibers break down unlike metal and stone. But most mass produced garments are created to be disposable (ironically made with petrochemical-based materials that do not biodegrade). Obsolescence means more shopping; it’s part of the business model. Again, from this perspective, mending is subversive. Plus, mending is not hard! From Sekules:

But honestly, there’s only one thought that stops anyone from doing anything, ever, and that is, “I can’t.” But you can mend, even if you’ve never held a needle. Ever since sewing skills melted away, a perception has taken hold that clothes are as impenetrable as a car or an iThing, fixable only by experts—it’s a big reason for our disposable culture.

And she continues:

There were indeed many rules in sewing once upon a time, but we can discount them now. If you like, you can easily stop at the first stitch, running stitch, and still mend for life. Don’t be perfect. Seriously, this is the freest kind of sewing in the world—no wrong way, no pressure, unique fashion item made every time. There’s very little to learn: you practice visible mending by doing it, and in my experience, everyone everyone is always delighted with their very first mend, no exceptions.

Sekules is an artist of visible mending (VM) in which the mends are artistic statements on the original clothing. The book mentions many creative techniques as well as creative upcycling, and for created re-use of existing clothing, such as the following Margiela pattern for making a sweater from many socks.

Or this creative crochet mend for a knit sweater from Sekules’s website.

This book makes mending accessible and playful with a repository of stitch types and patch techniques. It includes advice on fabrics (prefer natural fibers like cotton, wool, and linen); washing (to reduce stress on fabrics, spot
clean with white vinegar, cheap vodka, or baking soda to remove stains and smell); storing (discouraging moths).

Repair and maintenance gives a feeling of agency, competency, and connectedness to the real world. Sekules quotes Steven J. Jackson in the following passage.

Repair, wrote information scientist Steven J. Jackson, “is the secret history . . . that has always sustained (but invisibly) the higher profile stories of exploration, empire, and globalization.” His handy phrase “broken world thinking” suggests an upending of our tired old idea that production and innovation are the be-all and end-all. Instead we could conceive of society as a lot of stuff continuously in flux, being broken down and reconstituted, because, as Jackson says, “the world is always breaking; it’s in its nature to break.”1 The fixer knows this. These visible menders live it.”

Slow fashion shows respect to the designers, drivers, dyers, seamstresses, machinists, and so many more who made the garment possible. Slow fashion is a nudge towards a more sustainable society that is more creative and directly engaged in the tools and processes around them.

  1. Jackson, Steven J., ‘Rethinking Repair’, in Tarleton Gillespie, Pablo J. Boczkowski, and Kirsten A. Foot (eds), Media Technologies: Essays on Communication, Materiality, and Society ↩︎

Dan Davies on accountability sinks

by Aline Normoyle

In his book, The Unaccountability Machine: Why big systems make terrible decisions and how the world lost its mind, Dan Davies offers several compelling reasons for why modern life, particularly in the United States, feels so dysfunctional and inefficient. One of the reasons are accountability sinks. An accountability sink occurs when an organization takes the authority for decision making away from people in favor of rigid rules and policies. Davies gives a typical airport experience as an example:

UnaccountabilityMachine

Someone – an airline gate attendant, for example – tells you some bad news; perhaps you’ve been bumped from the flight in favour of someone with more frequent flyer points. You start to complain and point out how much you paid for your ticket, but you’re brought up short by the undeniable fact that the gate attendant can’t do anything about it.

It is a familiar, frustrating situation but to yell at the airline attendant is unfair: they had nothing to do with the policy and now agency to help you. Davies continues:

Somehow, the airline has constructed a state of affairs where it can speak to you with the anonymous voice of an amorphous corporation, but you have to talk back to it as if it were a person like yourself.

It is easy to see how this situation benefits the corporation. Davies summarizes the situation beautifully:

Meanwhile, the managers who made the decision to prioritise Gold Elite members are able to maximise shareholder value without any distractions from the consequences of their actions. They have constructed an accountability sink to absorb unwanted negative emotion.

When customers get angry, a low-paid employee absorbs the anger. The connection between the decider and those affected is severed. However, there are risks to willful ignorance: an inability to anticipate problems, respond to
emergencies, or even respond to the normal variability of the real world in ways that make sense. To give an extreme example, Davies described how the Dutch airline, KLM, killed 440 squirrels using a shredder machine because of a policies about importing animals. Even the ideology to maximize profits for shareholders is itself an accountability sink. The brainless pursuit of a reductive objective function at the cost of society and the environment is also madness. As a result, our corporations (as well as the many other institutions that have modeled themselves after corporations) are dysfunctional, inefficient, and routinely surprised by predictable crises. Davies effectively invokes Hanlon’s Razer.

The decline in individual accountability for unpopular decisions is not – or not only – a form of moral decline on the part of our rulers. It’s also a consequence of the fact that there are fewer decision makers than there used to be. Nearly all the commands and constraints which afflict the modern individual, the decisions which used to be made by identifiable rulers and bosses, are now
the result of systems and processes.

On the other hand, it is not correct that a world without policies and rules would be better. Davies acknowledges that policies and rules, like laws, are alternatives to systems in which capricious authorities make all the rules on a case by case basis. Good procedures have the potential to be transparent, fair, and flexible in light of common sense special cases. The problem is that in order to handle special cases, someone must take responsibility for that decision if it goes wrong. Thus, we have accountability sinks, e.g. systems whose decisions are officially made by no one. On this point, Davies includes the following quote from George Orwell’s review of James Burnham and the Managerial Revolution (1946):

Capitalism is disappearing, but Socialism is not replacing it. What is now arising is a new kind of planned, centralised society which will be neither capitalist nor, in any accepted sense of the word, democratic. The rulers of this new society will be the people who effectively control the means of production: that is, business executives, technicians, bureaucrats and soldiers.

Also in this book, Dan Davies also likens corporations to artificial intelligences, describes how the abandoned field of cybernetics theorized about effective management, and how the Friedman doctrine has led to an dysfunctional economy.

If you are interested in talking about books like this one with other interesting people, come to the Neo-Luddite Reading Group at Iffy Books!