Blog

2020 2021 2022 2023 2024 2025 2026
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
2004 2005 2006 2007 2008 2009

Summer Projects

  1. Teach my Organizational Change workshop in Manchester in July. This is booked, and I’m looking forward to it.

  2. I’d like to run my workshop on shutting projects down for the first time. If you’d like to host or take part, please give me a shout.

  3. I’d also like to run How to Not Be Wrong About AI for the first time. Again, if you’d like to host or participate, please let me know.

  4. Revise and deliver Managing Research Software Projects, although this will probably take a back seat to the three workshops above.

  5. Finish writing Sex and Drugs and Guns and Code, which in practice means doing the drudge work of editing 70,000 words of blog posts into something coherent.

  6. I’ve started writing Lean for Python Programmers as a way to (finally) learn a pure functional programming language. It’s slow going, in part because most of the existing documentation is more focused on the teacup than on the tea, but I’ll try to stick with it.

  7. Finish writing The Cloudherd and the Tiger’s Boy, a YA fantasy novel I’ve started and abandoned several times. I have about 24,000 words against a target of about 90,000, but only three quarters of a plot.

  8. Get an agent. Some publishers still accept unagented submissions, but having someone advocate for your writing is more important than ever in an age of AI slop. I’ve sent over a dozen inquiries in the last year without luck; if I want any of the stuff I’m writing to get on the shelves, I need to push myself on this.

Enshittification

In 2005, a Dutch startup called Booking.com offered hotels a deal: list your rooms on our platform for a 12% commission, and we will send you customers you would not otherwise reach. Hotels signed up; travelers followed, because the inventory was there, and by the early 2010s, Booking.com was the dominant hotel search platform across Europe and much of Asia.

Then the commissions started climbing. By 2019, many hotels were paying 25-30% per booking, plus additional fees for “preferred placement” near the top of search results. Hotels that declined to pay for placement found themselves buried behind those that did. The traveler experience degraded too: search results increasingly reflected who had paid for prominence, not which hotel best matched the search.

Hotels understood what had happened, but they were locked in. Their repeat customers now booked through Booking.com rather than directly, because that was where travelers looked. A hotel that left the platform did not take its customers with it—those relationships belonged to the platform. Leaving meant losing access to a market the platform now controlled.

Cory Doctorow named this pattern enshittification. A platform enters a market by offering a service below cost to build a user base. Once users are locked in, it begins subsidizing business customers, using the captive user base as leverage. Once business customers are also locked in, it harvests both by degrading service quality, raising prices, and extracting the maximum value from a market it now controls. The platform can do this because the switching costs that lock users in also protect it from competitive consequences.

Enshittification depends on two things: network effects and structural lock-in. Neither is specific to digital platforms. The record club Columbia House ran the same play in a different era. Launched in the 1950s in the United States and Canada and later extended to the United Kingdom, Australia, and Brazil, it offered new members twelve records or cassettes for a penny. The first transaction was a real deal. The extraction came later: members committed to purchasing eight more titles at “regular club prices,” which were two to three times the retail price of the same albums in a shop. If a member forgot to decline the monthly selection, it arrived automatically and the cost appeared on their bill. The penny offer built the membership; the commitment structure extracted the value. Columbia House recruited around sixteen million members in the United States alone before the model collapsed when digital music eliminated the inventory advantage.

In 1986, a corporation better known for making women’s underwear acquired JanSport and, over the following decades, bought nearly every backpack brand with a reputation for durability. Controlling more than half the US backpack market, it had no competitive pressure to maintain quality. Fabric thickness dropped, cheaper zippers replaced better ones, and stitching density fell. The products looked identical on the shelf; customers discovered what they had actually bought when the stitching pulled apart at the stress points.

JanSport continued to advertise a lifetime warranty, and the suggestion “just use the warranty” sounds entirely reasonable. In practice, using it required paying $12 to $25 in return shipping, waiting three to six weeks, and arguing that a failure qualified as a “defect in materials and workmanship” rather than “normal wear and tear.” (The warranty language was not incidental; it was written to exclude precisely the kind of failure that was now designed into the product.) One customer, when told their zipper failure was wear and tear, got quotes of $50 to $100 from local tailors, then bought a used bag at a thrift store for four dollars rather than a new one.

Grab, the ride-hailing and delivery platform dominant across eight countries in Southeast Asia, followed the same trajectory at scale. It entered markets including Malaysia, Indonesia, Vietnam, Thailand, and the Philippines with driver incentives and passenger subsidies that made rides cheaper than local alternatives. It acquired Uber’s Southeast Asian operations in 2018, eliminating its main competitor outright. With competition gone, driver commissions rose and passenger fees increased. The platform began requiring restaurants and drivers to pay for placement in its food-delivery and services listings. The pattern was identical to Booking.com’s, conducted in markets where regulatory capacity to respond was considerably thinner.

Investor dynamics accelerate enshittification. Platforms in their subsidy phase operate at significant losses, funded by venture capital in expectation of eventual monopoly returns. Once the platform achieves dominance, those investors demand extraction. Losses during the subsidy phase are booked as investment; extraction during the harvest phase is booked as profit. The users who benefited from below-cost service in year one fund those returns in year ten.

Antitrust enforcement has provided limited relief. Because acquisitions of potential competitors are evaluated on whether they raised consumer prices immediately—not whether they reduced competition structurally—dominant platforms can acquire dozens of rivals before those rivals threaten their market share. The result is that enshittification faces no real competitive check: alternatives are acquired or driven out during the subsidy phase.

Two structural interventions have produced results. The European Union’s Digital Markets Act, which took effect in 2024, requires platforms designated as “gatekeepers” (those with market capitalizations above €75 billion or monthly user bases above 45 million in Europe) to allow interoperability with competing services, to refrain from self-preferencing their own products in search results, and to allow users to uninstall pre-installed software. Fines for non-compliance reach 20% of global revenue. The act is the first regulatory framework designed specifically around the leverage that platform dominance creates, rather than around the consumer prices those platforms charge.

India’s Unified Payments Interface uses a different model: intervene before dominance rather than after. Instead of regulating private platforms that have already achieved lock-in, India built public payment infrastructure that any platform can connect to on equal terms. Google Pay, PhonePe, and Paytm compete on the same rails; none owns the customer relationship—that belongs to the user’s bank account. No single platform can raise fees on the underlying infrastructure because the infrastructure is public. Brazil’s Pix system follows similar principles, as do comparable approaches adopted by central banks in Ghana and Sri Lanka. The question today is not whether enshittification can be stopped, but why regulators in Canada, the US, and elsewhere choose not to stop it.

see the whole series · email me

Doctorow2022
Cory Doctorow and Rebecca Giblin: Chokepoint Capitalism: How Big Tech and Big Content Captured Creative Labor Markets and How We’ll Win Them Back. Beacon Press, 2022, 9780807007068.
Doctorow2025
Cory Doctorow: Enshittification: Why Everything Suddenly Got Worse and What to Do About It. Farrar, Straus and Giroux, 2025, 9780374619329.
Sapp2026
Keyana Sapp: “Your Backpack Got Worse On Purpose.” Worse on Purpose, March 23, 2026, https://www.worseonpurpose.com/p/your-backpack-got-worse-on-purpose.
Shapiro1999
Carl Shapiro and Hal R. Varian: Information Rules: A Strategic Guide to the Network Economy. Harvard Business School Press, 1999, 9780875848631.

Women's Work

On October 24, 1975, ninety percent of Icelandic women refused to work at paid jobs, in the home, or anywhere else. They called it a “day off”, though what they demonstrated was how much labor had been invisible. Schools closed, factories shut down, flights were cancelled, and men brought their children to work because there was no one else to care for them. The economy of Iceland effectively stopped. The women named what had been unnamed: everything they did without wages was work, and its absence was catastrophic.

The standard way economists measure an economy is through gross domestic product. GDP counts what is bought and sold, but not what is done without money changing hands. A woman who cooks dinner for her family contributes nothing to GDP, but if she hires someone to cook dinner and goes out to work herself, GDP goes up twice: once for her wages and once for the cook’s. The economic system counts the second arrangement as more productive. This is not an oversight. Marilyn Waring, a New Zealand politician and economist, spent years documenting the specific decisions embedded in the UN System of National Accounts that made domestic and care labor invisible. Her 1988 book showed that these decisions were choices made by people who had never performed most of the work in question.

The consequences of that invisibility are far-reaching. Women have fewer pension entitlements because pension systems are built on formal wage labor. They have less access to credit because banks valued declared income, not the labor of raising children or caring for elderly parents. When Waring raised these issues in the New Zealand parliament, she was told that economics was a technical matter and not a feminist one. She pointed out that someone had already decided that preparing food counted as productive activity only if someone received a wage for it. If that is not a political decision, it is hard to know what is.

In Japan, the term sengyo shufu—“full-time housewife”—became widespread only in the postwar period, when rapid industrialization in the 1950s and 1960s created male industrial jobs and pushed women into domestic roles. The image of the sengyo shufu as a traditional Japanese institution is largely the invention of corporations and government policy from a period of perhaps thirty years. Before industrialization, Japanese women had worked extensively in agriculture, small business, and textile production. The “tradition” was assembled recently under identifiable political and economic pressure.

The sociologist Arlie Hochschild spent years tracking what happened when both partners in a household entered paid employment, which most households in wealthy countries had done by the 1980s. She found that women in dual-income couples were coming home and working a second shift of domestic labor while their partners rested. Over the course of a year, the average woman in her study worked roughly a full additional month compared to her partner. Hochschild called this the second shift, and the pattern was consistent across Europe, Australia, South Korea, and Latin America: when women entered paid work, they tended to add paid hours on top of unpaid hours rather than trading one for the other.

Hochschild also found that the women she interviewed had largely rationalized the inequality by developing what she called “family myths”: stories about why the division made sense in their particular case. The myths varied but the pattern was constant: women interpreted inequitable arrangements as choices, and blamed themselves when the choices felt unsustainable. This is the same mechanism by which the “passion principle” (discussed earlier) operates: exploitation is most durable when those experiencing it explain it to themselves as freedom.

The historian Silvia Federici argued that the witch trials of the sixteenth and seventeenth centuries, during which tens of thousands of women were murdered, were primarily a way to discipline female labor. The women most often accused were those who lived outside household structures, like widows, healers, and midwives. The trials coincided with the enclosure of common land and the creation of a landless wage-labor force. Federici showed that enclosing women’s reproductive and domestic labor inside the household, unpaid and culturally devalued, was an economic project as much as a cultural one; it produced a class of workers whose labor was essential and whose compensation was called love.

Today, the global economy depends on nursing, childcare, eldercare, and domestic service, which are all systematically undervalued because they are feminized. The workers are underpaid not because care work is unskilled but because the labor market for care work is structured differently from the labor market for work that men predominantly perform. In India, the national health system relies heavily on informal women’s health workers, the ASHAs, who are classified as volunteers rather than employees and receive “incentive payments” rather than wages. This is not an oversight; it is a legal and administrative decision with direct economic consequences for nearly a million women. The ASHAs have organized and struck repeatedly for employee status and wage protections. The government has consistently refused.

Platform labor has reproduced the same structure with new interfaces. Apps coordinating domestic cleaning, childcare, elder companionship, and household maintenance present themselves as markets connecting independent contractors with clients. What they actually do is take informal, feminized labor that was already undervalued and formalize it while keeping the wages and protections of the informal economy. The worker bears the risk, and the platform takes the margin; the work is performed overwhelmingly by women (often migrant women) working for rates that reflect the historical devaluation of care.

Invisible labor is endemic to tech companies. Research has shown again and again that women perform more of the relational maintenance that holds teams together, such mentoring junior colleagues and managing interpersonal conflict, without receiving credit for it in performance reviews or promotion decisions. A study of a large software company found that women spent significantly more time on what the researchers called “non-promotable tasks” that benefited the organization but was invisible to the criteria used to evaluate careers. When it was pointed out that women were doing more of this work, the response from management was to encourage everyone to do less of it, which predictably meant that women continued to do it and men continued not to.

Five years after the 1975 strike in Iceland, Iceland elected its first female president, Vigdís Finnbogadóttir, who went on to serve four terms. The organizing that preceded the strike built the political infrastructure that made the election possible. Fifty years on, though, the gap between men’s and women’s wages in Iceland has not closed, and women still perform more unpaid domestic labor.

see the whole series · email me

Federici2004
Silvia Federici: Caliban and the Witch: Women, the Body and Primitive Accumulation. Autonomedia, 2004, 9781570270598.
Folbre2001
Nancy Folbre: The Invisible Heart: Economics and Family Values. The New Press, 2001, 9781565846555.
HochschildMachung1989
Arlie Hochschild and Anne Machung: The Second Shift: Working Families and the Revolution at Home. Viking, 1989, 9780142002926.
Waring1988
Marilyn Waring: If Women Counted: A New Feminist Economics. Harper & Row, 1988, 9780062509338.

Setting the Standard

In May 1886, American railroads completed the largest coordinated industrial operation the country had seen. Over two days, work crews across the South pulled up thousands of miles of track and relaid it three inches closer together, converting the region’s idiosyncratic gauge to match the rest of the country. The process was planned, organized, and executed in less time than most tech companies take to schedule a product launch.

The southern railroads had resisted this change for decades because the incompatibility was not an accident. Different gauges meant that northern rolling stock could not run on southern track, so freight crossing a gauge boundary had to be unloaded, transferred, and reloaded. Every cargo shipment lost time and money at the border, and that money went to the carriers who controlled the boundary. The technical choice was a market weapon.

A technical standard is an agreement about how things connect. The value of having a standard usually far exceeds the value of any particular outcome in the format war that precedes it, which is exactly why control of the format is worth fighting for.

Thomas Edison understood this in the 1880s, when he and George Westinghouse fought over whether the United States would run on direct or alternating current. Edison had bet on DC infrastructure and stood to lose if AC won, so he electrocuted animals at exhibitions and lobbied for AC to be used in the first electric chair, hoping to fix the association of alternating current with execution in the public mind. None of this changed the underlying physics, though; AC transmits over longer distances at lower cost, and Westinghouse won.

A century later, Microsoft repeatedly used standards capture as a business strategy. The company would adopt an open standard like HTML, Java, or Kerberos, and then add proprietary extensions that were convenient but tied developers to Microsoft’s version. Once enough code was written against the extensions, interoperability with non-Microsoft systems degraded. The strategy had a name inside the company: embrace, extend, extinguish. (The “extinguish” referred to both the open standard and Microsoft’s competitors.)

Google’s Accelerated Mobile Pages, launched in 2015, was presented as an open standard for fast-loading mobile web pages. To receive preferential treatment in Google’s mobile search results, however, pages had to be hosted on Google’s own servers and delivered through Google’s infrastructure. Publishers who adopted AMP handed Google control over their content delivery in exchange for a favorable page rank. Google eventually backed away from the hosting requirement, but only under regulatory pressure.

Apple took a different approach. Rather than capturing an open standard, it simply refused to implement one. RCS is a modern successor to SMS that supports end-to-end encryption, read receipts, high-resolution media, and group chat. For years after Android phones had adopted it, Apple kept cross-platform messaging on unencrypted SMS, so messages between iPhones and Android phones appeared as green bubbles. Internal documents revealed in US antitrust litigation showed Apple executives explicitly acknowledging that adopting RCS would reduce the social cost of switching from iPhone to Android, particularly for teenagers for whom the green-bubble distinction had become a social marker. The EU’s Digital Markets Act eventually compelled Apple to support RCS in 2024.

When a technology becomes embedded in a formal standard, the patent holder typically commits to license it on FRAND terms: fair, reasonable, and non-discriminatory. This sounds reassuring, but the word “reasonable” has funded decades of litigation. Qualcomm’s licensing practices for mobile baseband patents (the technology that connects a phone to a cellular network) were subject to antitrust proceedings simultaneously in the United States, Europe, South Korea, China, and Taiwan in the late 2010s. The core allegation was consistent across jurisdictions: Qualcomm was using its position as the patent holder o charge fees that bore no relationship to any recognizable interpretation of “reasonable.” The legal outcomes varied by country; Qualcomm remained profitable throughout.

During the browser wars of the late 1990s, Microsoft bundled Internet Explorer with Windows, deliberately undercutting Netscape’s business model. IE’s market dominance then allowed Microsoft to implement proprietary HTML extensions that worked only in its browser. What finally broke this cycle was not antitrust action—which produced a consent decree and little structural change—but the arrival of Firefox and then Chrome, backed by organizations with different incentives.

If you are a developer building a product, every choice between an open standard and a proprietary API is a bet on the future behavior of the platform that owns the API. History suggests the bet is a bad one. From Instagram’s third-party API and Twitter/X’s developer API to Google Reader and Firebase’s original pricing structure, platforms have repeatedly changed terms after developers are so invested that they can’t walk away. Each time, the platform points to its terms of service, which always reserves the right to change.

The structural remedy for this is interoperability mandates: legal requirements that platforms accept connections from competing services on terms that do not depend on the platform’s goodwill. Phone number portability, which required carriers to let customers keep their numbers when switching providers, eliminated one of the most effective lock-in mechanisms in telecoms. Messaging interoperability requirements in the EU’s Digital Markets Act are attempting to do the same thing for social platforms. Whether they succeed depends on implementation details that are actively contested, and on enforcement that companies have historically resisted every way they can.

see the whole series · email me

Russell2014
Andrew L. Russell: Open Standards and the Digital Age: History, Ideology and Networks. Cambridge University Press, 2014, 9781107612044.
Shapiro1999
Carl Shapiro and Hal R. Varian: Information Rules: A Strategic Guide to the Network Economy. Harvard Business School Press, 1999, 9780875848631.
Wu2010
Tim Wu: The Master Switch: The Rise and Fall of Information Empires. Knopf, 2011, 9780307390998.

Twelve Ways to Be Wrong About AI-Assisted Coding

Suppose your manager asks you next week to demonstrate that the AI coding tools your company signed up for are worth the subscription cost. Would you measure lines of code generated, or tickets closed? Or would you send out a survey asking whether developers feel more productive? Each of those approaches is flawed in a different way; the sections below explain why.

Note: this post is about how people are assessing AI, not at LLM-assisted coding itself; with a little rewording, these criticisms could be applied to a lot of the claims that have been made about agile development, test-driven development, and other practices. If I’ve learned anything in the last twenty years, it’s that software engineering would be a lot further ahead today if we had been willing to let our peers in the human sciences teach us how to study these kinds of things properly.

Also, if you’d like a one-day introduction to the research methods you should use to avoid making these errors, please reach out. I’m not qualified to teach it, but I know people who are, and I could probably talk them into doing it…

Counting Lines of Code Generated

Proxy metrics stand in for concepts that are hard to measure directly, and lines of code is one of the oldest. LLMs generate more code, but not necessarily better outcomes: a team that sees a 40% increase in lines of code per developer after adopting LLM tools has measured verbosity, not productivity. Deleting 2000 lines of tangled logic and replacing it with 200 clean ones is an improvement that looks like a loss on this metric [Sadowski2019]. More code also means more to read, maintain, and debug, and AI’s contribution to that future burden does not appear in the line count.

Timing Artificial Tasks

A widely cited study found that developers who used GitHub Copilot completed a task 55% faster than those who did not [Peng2023]. The task was implementing an HTTP server in JavaScript from scratch, in ninety minutes; the developers had no other obligations that day. Real software development involves navigating a large codebase you did not write, understanding a requirement described ambiguously in a ticket, coordinating with colleagues, and attending meetings. Speed on a greenfield toy task does not predict speed on any of that. A randomized controlled trial with experienced open-source developers found the opposite of what participants themselves predicted: giving them access to AI tools increased task completion time by 19% [Becker2025].

Before/After With No Control Group

You start using LLMs in January; by June, pull requests are shipping faster, so the tools must be working, right? But between January and June you hired twelve engineers, refactored the CI pipeline, and switched your cloud provider. Without a group that did not adopt the tools, you cannot separate the effect of LLMs from any of the other changes that happened at the same time. Internal validity requires a credible counterfactual, i.e., some way of knowing what would have happened otherwise.

Asking Developers If They Feel More Productive

Survey results like “87% of developers report feeling more productive with AI tools” are regularly cited as evidence that the tools work [Liang2024], but three things make self-report systematically misleading:

Counting Commits, Pull Requests, and Tickets

In 2023, McKinsey proposed measuring individual developer productivity using counts of commits, pull requests, code reviews, and similar activities [McKinsey2023]. Goodhart’s Law states that when a measure becomes a target, it ceases to be a good measure [Goodhart1984]. When developers know their commit count is tracked, they make more, smaller commits; when ticket counts are tracked, tickets get split. The numbers improve while the underlying work does not [Beck2023]. Activity is not output; output is not value.

Measuring Only the Easy Half

LLMs make code generation faster, and that half is easy to measure. The other half is harder: time spent reviewing LLM-generated code for correctness, time lost debugging confidently wrong suggestions, security vulnerabilities introduced by plausible-looking but insecure code, and technical debt from suggestions that solved the immediate problem while ignoring the surrounding design. A study of GitHub Copilot’s code found that a substantial fraction of generated code contained security vulnerabilities, and that developers under time pressure accepted insecure suggestions at higher rates [Pearce2022]. A 2025 evaluation of five major LLMs found that none produced web application code meeting industry security standards [Dora2025]. A large-scale analysis of over 300,000 AI-authored commits found that more than 15% introduce at least one quality issue, and nearly a quarter of those issues persist in the codebase long-term [Liu2026]. Measuring only the inputs that go up while ignoring the costs that also rise is not measurement; it is marketing.

Treating Adoption Rate as a Success Metric

“We have achieved 90% AI tool adoption across engineering” is a procurement outcome, not a productivity outcome. Adoption measures whether the tool is installed and opened; it says nothing about whether suggestions are useful, whether developers accept them thoughtlessly, or whether the accepted suggestions are correct. High adoption combined with low suggestion quality produces a workforce spending time managing a tool rather than benefiting from one. A study of IBM’s enterprise AI coding assistant found that while the tool often provided net productivity increases, those gains were not experienced uniformly across its user base [Weisz2025]. But adoption is easier to measure than benefit, which is exactly why it gets reported instead.

Comparing Volunteers to Non-Volunteers

Studies that compare developers who chose to use LLMs against those who did not are comparing two different populations, not two conditions. Early adopters differ from late adopters and non-adopters in ways that directly predict productivity: they are more motivated to experiment, more comfortable with new tooling, and more likely to already be high performers. Selection bias means any observed difference between the groups may be a property of the person rather than the tool. This is the most common design flaw in industry AI productivity reports, because it is the cheapest study to run. A two-year longitudinal study of Copilot use at a large IT organization found that developers who used the tool were consistently more active than non-users even before it was introduced [Stray2026].

Measuring the Individual Instead of the System

Individual coding speed is the easiest thing to measure, so it gets measured. But if AI tools help developers write code 30% faster and the team’s time from ticket to production does not change, the bottleneck was not writing code. More code generated also means more code to review: if AI increases code volume without increasing review capacity, cycle time may worsen [Forsgren2021]. An empirical study of professional developers found that while AI tools boosted output for less-experienced contributors, senior developers experienced a 19% decline in their own productivity as they absorbed a 6.5% increase in code review load from AI-generated code [Xu2025]. Optimizing one stage of a pipeline while ignoring the others is a systems-thinking failure dressed as a productivity study.

Measuring During the Novelty Period

A four-week study that finds a productivity boost has found a four-week productivity boost. The novelty effect is real: developers are more engaged with new tools during the initial period, which inflates observed performance relative to the long-run baseline. Effects that actually matter emerge over months, not weeks, including skill atrophy for tasks now delegated to the AI, accumulation of technical debt from wrong suggestions, or changes in how teams collaborate. A study designed to detect a short-term benefit, has not told you anything about what happens after the study ends. An analysis of 807 open-source repositories adopting Cursor found exactly this pattern: adoption produced a large but transient increase in development velocity alongside a substantial and persistent increase in code complexity and static analysis warnings [He2026].

Treating Suggestion Acceptance Rate as a Quality Signal

LLM coding assistants commonly report what fraction of their suggestions developers accept, and higher acceptance rates are presented as evidence that the tool is useful. Acceptance measures whether the generated code looked plausible enough for a developer to press Tab; it does not measure whether the code was correct, secure, or maintainable. Developers under time pressure accept more suggestions, including insecure ones [Pearce2022], so a tight deadline makes acceptance rate rise for exactly the wrong reasons. An enterprise study of 400 developers found a 33% average acceptance rate alongside high developer satisfaction, but tracked no measure of the correctness or security of accepted code [Bakal2025]. A metric that rewards looking good enough is not a metric that rewards being good.

Comparing AI to Nothing

Studies that compare AI-assisted developers to a control group using nothing have chosen a baseline that does not exist in practice. Developers without LLM assistants use documentation, colleagues, and the time they would otherwise spend thinking through the problem themselves. The relevant question is whether LLM tools outperform the alternatives developers already have, and that comparison is rarely made [Peng2023]. Choosing a weak baseline makes any tool look good; it does not make the tool useful.

I am profoundly grateful to everyone who has taken the time over the years to explain this stuff to me. Any errors or over-simplifications in what I’ve written are entirely my fault.

Bibliography

Bakal2025
Gal Bakal, Ali Dasdan, Yaniv Katz, Michael Kaufman, and Guy Levin: “Experience with GitHub Copilot for Developer Productivity at Zoominfo.” arXiv:2501.13282, 2025.
Becker2025
Joel Becker, Nate Rush, Elizabeth Barnes, and David Rein: “Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity.” arXiv:2507.09089, 2025.
Beck2023
Kent Beck: “Measuring Developer Productivity: Real-World Examples.” Medium, 2023. https://tidyfirst.substack.com/p/measuring-developer-productivity
Dora2025
Swaroop Dora, Deven Lunkad, Naziya Aslam, S. Venkatesan, and Sandeep Kumar Shukla: “The Hidden Risks of LLM-Generated Web Application Code: A Security-Centric Evaluation of Code Generation Capabilities in Large Language Models.” arXiv:2504.20612, 2025.
Flournoy2025
John C. Flournoy, Carol S. Lee, Maggie Wu, and Catherine M. Hicks: “No Silver Bullets: Why Understanding Software Cycle Time is Messy, Not Magic.” arXiv:2503.05040, 2025.
Forsgren2021
Nicole Forsgren, Margaret-Anne Storey, Chandra Maddila, Thomas Zimmermann, Brian Houck, and Jenna Butler: “The SPACE of Developer Productivity.” ACM Queue, 19(1), 2021.
Goodhart1984
Charles Goodhart: “Problems of Monetary Management: The U.K. Experience.” In Anthony Courakis (ed.), Inflation, Depression, and Economic Policy in the West, Rowman and Littlefield, 1984.
He2026
Hao He, Courtney Miller, Shyam Agarwal, Christian Kästner, and Bogdan Vasilescu: “Speed at the Cost of Quality: How Cursor AI Increases Short-Term Velocity and Long-Term Complexity in Open-Source Projects.” Proc. MSR ‘26, 2026.
Hicks2025
Catherine M. Hicks, Carol Lee, and Kristen Foster-Marks: “The New Developer: AI Skill Threat, Identity Change & Developer Thriving in the Transition to AI-Assisted Software Development.” https://osf.io/preprints/psyarxiv/2gej5_v2, 2025, https://doi.org/10.31234/osf.io/2gej5_v2.
Liang2024
Jenny T. Liang, Chenyang Yang, and Brad A. Myers: “A Large-Scale Survey on the Usability of AI Programming Assistants: Successes and Challenges.” Proc. ICSE 2024, 2024.
Liu2026
Yue Liu, Ratnadira Widyasari, Yanjie Zhao, Ivana Clairine Irsan, Junkai Chen, and David Lo: “Debt Behind the AI Boom: A Large-Scale Empirical Study of AI-Generated Code in the Wild.” arXiv:2603.28592, 2026.
McKinsey2023
Nora Elsayed, Tarek Elhounsri, and Sven Blumberg: “Yes, You Can Measure Software Developer Productivity.” McKinsey & Company, 2023. https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/yes-you-can-measure-software-developer-productivity
Pearce2022
Hammond Pearce, Baleegh Ahmad, Benjamin Tan, Brendan Dolan-Gavitt, and Ramesh Karri: “Asleep at the Keyboard? Assessing the Security of GitHub Copilot’s Code Contributions.” Proc. IEEE S&P 2022, 2022, doi:10.1109/SP46214.2022.9833571.
Peng2023
Sida Peng, Eirini Kalliamvakou, Peter Cihon, and Mert Demirer: “The Impact of AI on Developer Productivity: Evidence from GitHub Copilot.” arXiv:2302.06590, 2023.
Sadowski2019
Caitlin Sadowski and Thomas Zimmermann (eds.): Rethinking Productivity in Software Engineering. Apress, 2019, 978-1484242209.
Stray2026
Viktoria Stray, Elias Goldmann Brandtzæg, Viggo Tellefsen Wivestad, Astri Barbala, and Nils Brede Moe: “Developer Productivity With and Without GitHub Copilot: A Longitudinal Mixed-Methods Case Study.” Proc. HICSS-59, 2026.
Weisz2025
Justin D. Weisz, Shraddha Kumar, Michael Muller, Karen-Ellen Browne, Arielle Goldberg, Ellice Heintze, and Shagun Bajpai: “Examining the Use and Impact of an AI Code Assistant on Developer Productivity and Experience in the Enterprise.” Proc. CHI‘25, 2025.
Xu2025
Feiyang Xu, Poonacha K. Medappa, Murat M. Tunc, Martijn Vroegindeweij, and Jan C. Fransoo: “AI-Assisted Programming Decreases the Productivity of Experienced Developers by Increasing the Technical Debt and Maintenance Burden.” arXiv:2510.10165, 2025.

SDGC Glossary

I will update this as new entries are added.

see the whole series · email me

A

accounting fraud
The manipulation of financial statements through aggressive use of permissible accounting choices like early revenue recognition, optimistic asset valuations, or buried liabilities to misrepresent a company’s financial position.
algorithmic amplification
The process by which platform recommendation systems systematically promote content that generates high engagement, regardless of its accuracy or harm.
algorithmic hiring
The use of automated software systems to screen, rank, or select job candidates, typically trained on historical hiring data.
anchoring
The cognitive tendency to rely too heavily on the first piece of numerical information encountered when making estimates or decisions. Even when people know that an initial number is arbitrary or irrelevant, it systematically pulls their judgments toward it.
at-will employment
An employment doctrine, dominant in the United States, under which either party may terminate the employment relationship at any time, for any reason not prohibited by law, without notice or cause. Most wealthy countries impose statutory notice periods, severance requirements, or just-cause protections that significantly limit at-will termination.
availability heuristic
The mental shortcut of estimating how likely or common something is by how easily examples of it come to mind. Vivid or recent events are overweighted relative to more common but less memorable ones.

B

backfire effect
The phenomenon in which presenting people with evidence that contradicts a strongly held belief causes them to hold that belief more firmly rather than updating it.
Basel Convention
The 1989 international treaty restricting the transboundary movement of hazardous waste, particularly from wealthy to developing nations.
bertillonage
The system of criminal identification developed by Alphonse Bertillon in the 1880s, based on recording eleven precise body measurements. Widely adopted by police forces in Europe and the colonial world, it was displaced by fingerprinting around 1900.
biometrics
The measurement and statistical analysis of unique physical or behavioral characteristics to identify or verify an individual’s identity.
blood libel
The accusation that Jews murdered Christian children to use their blood in Passover rituals.
boryokudan
The Japanese government’s official designation for organized crime groups, meaning “violence group,” used in anti-organized crime legislation since 1992 as an alternative to the self-identifying term yakuza.
bounded rationality
Herbert Simon’s term for the observation that human decision-making is rational only within the limits of available information, cognitive capacity, and time. People do not optimize; they search for solutions that are good enough and stop, a process Simon called satisficing.
brand protection
The enforcement of exclusive commercial identity through legal mechanisms such as trademarks and trade dress, used to prevent competitors from trading on an established name.

C

civil disobedience
The deliberate, nonviolent refusal to comply with laws or government demands as a form of political protest, aimed at changing the law rather than evading it.
codetermination
A labor relations system in which workers hold formal representation rights on corporate governing boards, as in the German Mitbestimmung model.
cognitive bias
A systematic pattern of deviation from rational judgment in which people reach conclusions through processes that are predictably and consistently skewed.
cognitive pollution
The systematic degradation of the shared information environment through the spread of misleading, manipulative, or low-quality content at scale.
collective bargaining
The process by which workers negotiate wages, hours, and working conditions as a group through their union rather than as individuals.
commons
Resources held in common by a community rather than owned privately, subject to collective governance rather than market allocation.
conspicuous consumption
Thorstein Veblen’s term for spending on goods and services primarily to signal social status rather than for their practical use or pleasure. The signal only works if the expenditure is visible and sufficiently costly to exclude lower-status competitors.
conspicuous leisure
Thorstein Veblen’s term for the visible display of not working as a signal of social rank. In contexts where most people must labor, idleness signals that one can afford to abstain; in modern professional contexts, the signal has inverted to highly visible overwork and constant availability.
consumer cooperative
A cooperative owned by its customers rather than its workers, in which members pay a joining fee or purchase membership shares and receive a dividend proportional to their purchases, with governance following a one-member-one-vote principle.
content moderation
The practice of reviewing and enforcing community standards on user-generated content on digital platforms, including decisions about removal, restriction, and amplification.
control fraud
A form of financial fraud in which the executives of an institution use the institution itself as a vehicle for personal enrichment at the expense of depositors, investors, or clients.
cooperative
A firm owned and governed by its members in which profits are distributed in proportion to participation rather than capital invested, and governance follows a one-member-one-vote principle rather than a one-share-one-vote principle.
copyleft
A licensing approach that requires derivative works to be distributed under the same open terms as the original, preventing proprietary appropriation of open-source contributions.
credit union
A member-owned financial cooperative that accepts deposits, makes loans, and returns surplus to members rather than distributing profits to outside shareholders.
crop lien
A credit arrangement in which farmers borrowed against their future harvest, typically from merchants who charged high interest and kept borrowers in perpetual debt.

D

dark triad
A cluster of three overlapping personality traits—psychopathy, narcissism, and Machiavellianism—that researchers treat together because they tend to co-occur and share a common pattern of callous self-interest, manipulation, and disregard for others.
decriminalization
The removal of criminal penalties for a behavior, typically while other civil or regulatory restrictions remain in place.
Delaware flip
The process by which a non-US startup reincorporates by creating a Delaware holding company above the existing domestic entity and transferring intellectual property to it, typically required by American venture capital investors as a condition of investment.
deplatforming
The removal of an individual or group from a digital platform as a content moderation enforcement action.
diffuse decision-making
The distribution of accountability across many individuals in an organization such that no single person can be held responsible for harmful outcomes.
disposition effect
The tendency of investors to sell assets that have increased in value while holding on to assets that have declined, the opposite of what a rational strategy would recommend. It follows from loss aversion: selling a losing position requires acknowledging a loss, which is more painful than an equivalent gain is pleasurable.
divestment
The disposal of financial assets in a company or sector as a form of political pressure, imposing reputational and economic costs on targeted firms or governments.
donor-advised fund
A financial vehicle in which a donor contributes assets, receives an immediate tax deduction, and retains the ability to direct grants to charitable recipients at any pace over time, while the contributed assets remain under the donor’s effective control.
downline
In multi-level marketing, the network of distributors recruited by a participant and by those distributors’ own recruits, whose sales generate commission income for everyone above them in the chain.
dual-class shares
A corporate share structure in which two or more classes of stock carry different voting rights, typically giving founders or insiders shares with ten or more votes each while public investors receive shares with one vote, allowing the company to raise capital without ceding effective control.
due process
The legal requirement that government or institutional action affecting rights must follow fair procedures and provide a meaningful opportunity for challenge.

E

electronic waste
Discarded electronic devices containing toxic materials such as lead, mercury, and cadmium, much of which is exported for processing in developing countries under hazardous conditions.
Enclosure Acts
Parliamentary legislation in Britain from the seventeenth through nineteenth centuries that converted common agricultural land to private ownership.
Enclosure movement
The historical process by which common agricultural lands in Britain were converted into private property, displacing rural communities that had depended on shared access.
engrams
In Scientology, traumatic memories believed to be stored in the reactive mind and to impair mental and physical function; the target of the auditing process, which aims to discharge them through conscious recall.
enshittification
The process by which platform businesses progressively degrade the quality of their service as they shift value extraction from users to advertisers and away from both to shareholders.
equity compensation
Payment in the form of stock or stock options rather than cash, used extensively in the tech industry to align worker incentives with company value while deferring cash costs.
excommunication
Formal exclusion from membership in a religious community and its sacraments, historically used by the Catholic Church as a disciplinary and political tool against individuals.

F

face turn
In professional wrestling, a character’s transition from villain to hero, typically marked by a moment that reveals underlying decency and produces crowd approval where there was previously hostility.
Fair Game
A Scientology policy, first articulated in 1965, stating that members who leave the organization and publicly criticize it may be subjected to any action by other members without organizational discipline. It was officially rescinded in 1968 but documented as continuing in practice under different names.
Flexner Report
The 1910 report that restructured American medical education around elite university-based schools, simultaneously professionalizing medicine and closing institutions that had trained women and Black physicians.
Framework Convention on Tobacco Control
The 2003 World Health Organization treaty on tobacco regulation, the first international public health treaty negotiated under WHO auspices, committing signatory countries to measures including advertising restrictions, health warnings, and plain packaging.
FRAND
Fair, Reasonable, and Non-Discriminatory: the licensing commitment a patent holder makes when their technology is incorporated into a formal technical standard. The commitment is intended to prevent the patent holder from using a standards-essential patent as a chokepoint to extract monopoly rents from everyone implementing the standard. What counts as “reasonable” is routinely contested in litigation.
front-running
The practice of executing trades based on advance knowledge of pending client orders before those orders are filled, exploiting information asymmetry in financial markets.

G

general partnership
A business structure in which two or more people share ownership, management, and unlimited personal liability for the firm’s debts and legal obligations.
GNU Public License (GPL)
A free software license that grants users the freedom to use, study, modify, and distribute software, with the key requirement that any distributed modifications or derivative works must also be released under the same terms.
greater fool theory
The idea that an overpriced asset can still be a rational purchase if the buyer believes they can sell it to someone else at an even higher price before the market corrects. A bubble collapses when no greater fool can be found.

H

harm principle
John Stuart Mill’s principle, articulated in On Liberty (1859), that the only legitimate basis for restricting individual liberty is to prevent harm to others.
harm reduction
A public health approach that aims to reduce the negative consequences of drug use without requiring abstinence, through measures such as needle exchanges, naloxone distribution, and supervised consumption sites.
heel
In professional wrestling, a character designated to play the villain, whose role is to behave in ways the audience finds objectionable so that sympathy and approval flow toward the opposing character.
heuristics
Mental shortcuts or rules of thumb that people use to make decisions quickly under uncertainty. Heuristics often work well but produce systematic, predictable errors in specific circumstances, which behavioral economists call cognitive biases.
Hicklin test
The 1868 British legal standard defining obscenity as material tending to deprave and corrupt those whose minds were open to immoral influences, superseded in the US by the Miller test.
horizon problem
In cooperative economics, the tendency of member-owners to underinvest in long-term assets because their claim on the cooperative ends when their membership ends, giving them a shorter time horizon than the firm’s productive life.
hyperbolic discounting
The tendency to discount rewards in the near future much more steeply than rewards in the more distant future, producing time-inconsistent preferences.

I

impression management
The ongoing effort by individuals to control how they are perceived by others. In large organizational hierarchies, where decision-makers have limited direct observation of a candidate’s actual work, impression management skills become a primary determinant of career advancement, independently of underlying performance.
information asymmetry
A condition in which one party to a transaction has significantly more or better information than the other, enabling exploitation of that advantage.
interdict
A Catholic Church sanction that suspends religious services in a territory or for a community, historically used as collective punishment to pressure secular rulers.
interoperability
The ability of systems from different vendors or platforms to exchange information and work together without requiring special adaptation by the user. Interoperability mandates are legal requirements that a platform accept connections from competing services on non-discriminatory terms.
investor-state dispute settlement (ISDS)
A mechanism in trade and investment treaties allowing foreign investors to bring arbitration claims against governments in private tribunals, bypassing domestic courts, when government actions are alleged to have damaged their investments.
invisible labor
Work that is essential to the functioning of an organization or household but goes unrecognized, uncompensated, and uncounted because it is performed by people with less power (typically women, caregivers, and low-status workers). Examples include emotional labor, domestic work, and the administrative coordination tasks that fall to junior employees or support staff rather than the people who take credit for outcomes.

J

joint-stock company
A firm whose ownership is divided into transferable shares held by multiple investors, each of whom bears risk only in proportion to their investment; the organizational predecessor of the modern publicly traded corporation.

K

kayfabe
The collective maintenance, by performers and audience alike, of the fiction that professional wrestling matches are genuine athletic contests; by extension, any shared pretense that all parties know to be false but sustain because the fiction serves a social or commercial function.
keiretsu
A Japanese network of interlocked companies that hold shares in one another, maintain long-term supply and trading relationships, and are often anchored around a central bank, insulating member firms from hostile takeovers and short-term market pressure.

L

legibility
The degree to which a state or institution can observe, measure, and administer a population or resource through standardized categories.
legitimacy
The quality of being recognized as having rightful authority, distinct from power exercised by force or coercion alone.
limited liability corporation
A firm in which shareholders cannot be held personally responsible for the company’s debts or legal obligations beyond the amount they have invested, separating the financial risk of the owners from that of the enterprise.
limited partnership
A partnership structure with two classes of partners: limited partners, who risk only the capital they have contributed and take no part in management, and at least one general partner, who retains full personal liability and operational control.
long firm
A fraud in which a business establishes credibility through legitimate trading before ordering large quantities of goods on credit and disappearing with them.
longtermism
The belief that the most important thing people can do is to ensure that the long-run future of civilization goes well, on the grounds that a future containing trillions of people vastly outweighs present-day welfare in moral calculations. Longtermism is used to justify present harm in the name of speculative futures, and that its practitioners disproportionately assume that they themselves should be making those decisions.
loss aversion
The finding from prospect theory that losses produce roughly twice the psychological impact of equivalent gains.

M

Miller test
The 1973 US Supreme Court three-part test for obscenity, requiring that material be evaluated by community standards, appeal to prurient interest, and lack serious literary, artistic, political, or scientific value.
Mitbestimmung
The German system of codetermination, in which workers in large firms hold legally mandated representation on corporate supervisory boards with formal authority over management.
Mittelstand
The layer of medium-sized, often family-owned manufacturing and engineering firms that form the backbone of the German economy, characterized by long time horizons, specialist expertise, strong regional roots, and formal worker representation under the codetermination system.
Montreal Protocol
The 1987 international treaty requiring signatory countries to phase out the production and use of ozone-depleting substances, including chlorofluorocarbons (CFCs). It is notable for including differentiated timelines for developing and developed countries, a technology transfer fund, and trade sanctions against non-signatories.
moral disengagement
The psychological mechanisms by which people disengage their moral standards to engage in or tolerate harmful behavior, including displacement and diffusion of responsibility, dehumanization of victims, and advantageous comparison with worse acts.
moral licensing
The psychological phenomenon in which past virtuous behavior reduces the felt need for future virtuous behavior, sometimes leading to more harmful subsequent choices.
moral panic
A disproportionate social reaction to a perceived threat, typically focused on a specific group or behavior, amplified by media coverage and political entrepreneurs.
motivated reasoning
The cognitive tendency to evaluate evidence in ways that support conclusions already preferred, rather than following evidence to its logical conclusion.

N

narrative economics
The study of how popular stories and narratives spread through populations and drive economic behavior, including investment manias and market crashes.
neoreaction
A political movement, also called the Dark Enlightenment, arguing that Enlightenment values such as democracy, equality, and human rights have made societies weaker and more chaotic. The movement is distinct from classical fascism in its rejection of nationalism and its preference for corporate over state power, though critics note this distinction matters less to those governed by the result.
network effects
The phenomenon in which a product or service becomes more valuable to each user as more people use it, creating strong advantages for established platforms and barriers to entry for competitors.
normalization of deviance
The process by which organizations gradually come to accept risk thresholds that would initially have been unacceptable, through repeated exposure to near-misses that did not immediately produce catastrophe.
nudge
A change in how choices are presented, such as a default setting or the order of options, that predictably influences behavior without restricting options or changing financial incentives. Nudges exploit the same cognitive tendencies that produce bias, and may serve the choice designer’s interests rather than the chooser’s.

O

obscenity law
The body of law governing the prohibition of material deemed indecent or harmful to public morals, whose legal standards have shifted substantially across time and jurisdiction.

P

paradox of tolerance
The argument that unlimited tolerance must eventually lead to the disappearance of tolerance: a society that extends tolerance to movements committed to destroying tolerance will itself be destroyed.
Parkinson’s Law
The observation that work expands to fill the time available.
partnership
A business structure in which two or more people share ownership and, in a general partnership, unlimited personal liability; a foundational organizational form for professional services firms, law firms, and investment funds.
passion principle
The cultural norm in certain fields that workers should pursue their work as a vocation rather than merely a job, used to justify poor conditions, low pay, and long hours.
patent assertion entities
Companies that acquire patents not to produce goods but to extract licensing fees or litigation settlements from operating businesses; commonly called patent trolls.
payola
The practice of record labels paying radio stations or disc jockeys to play their songs without disclosing the payment to listeners.
penny press
The mass-market newspapers of the 1830s that lowered cover prices and relied on advertising revenue rather than subscriptions, establishing the commercial model that now dominates media. See also “yellow press”.
peonage
A system of debt bondage in which workers are legally required to continue working until a debt to an employer is repaid, effectively a form of coerced labor.
philanthrocapitalism
The use of private charitable foundations and large-scale philanthropic giving by wealthy individuals to shape public policy, research priorities, and social institutions, without democratic accountability or electoral legitimacy.
pollution haven
A jurisdiction with weak environmental regulations that attracts investment in polluting industries relocating from places with stricter standards.
Ponzi scheme
A fraudulent investment scheme that pays returns to earlier investors using funds from later investors rather than from actual profits, collapsing when new investment stops.
positional goods
Toods whose value depends on how few other people have them. Unlike material goods, which can be democratized as production scales, positional goods lose their value if everyone has access to them, producing self-defeating competition in which each round of gains returns participants to the same relative position.
precarious workers
Workers in contingent, part-time, contract, or gig arrangements who lack the employment security, benefits, and legal protections associated with standard employment.
predatory pricing
Setting prices below cost to drive competitors out of a market, with the intention of raising prices once competitive pressure is eliminated.
predictive policing
The use of algorithmic tools to forecast where crimes will occur or who is likely to offend, criticized for encoding and amplifying existing racial and socioeconomic biases.
propaganda model
The observation that mass media systematically filters news to serve elite interests through ownership structure, advertising dependence, and other institutional pressures.
prospect theory
A model of how people actually evaluate risky outcomes. Key findings are that people evaluate outcomes relative to a reference point rather than in absolute terms, and that losses loom roughly twice as large as equivalent gains (a phenomenon called loss aversion).
Psychopathy Checklist
A clinical assessment instrument for measuring psychopathic traits in individuals. The checklist evaluates characteristics including callousness, grandiosity, manipulation, lack of remorse, and failure to accept responsibility. The Framework is also used to show that corporate legal structure produces psychopathic behavior at the organizational level.
public benefit corporation
A corporate form that legally requires directors to consider the interests of employees, communities, and the environment alongside shareholder returns, providing some protection against purely financial acquisition pressure; available in several Canadian provinces and US states.
publication bias
The tendency for academic journals and researchers to publish positive or statistically significant findings while null or negative results go unreported, distorting the scientific record.
putting-out system
A pre-industrial manufacturing arrangement in which merchants supplied raw materials to workers in their homes and collected finished goods, a precursor to the factory system.

Q

R

racial formation
The process by which racial categories are created, transformed, and destroyed through political, economic, and social forces rather than biology.
redlining
The systematic denial of mortgages and insurance to residents of non-white neighborhoods in the US by government agencies and private lenders from the 1930s onward. It was formally outlawed in the US in 1968 but its effects persist in wealth and credit gaps.
regulatory capture
The process by which a regulatory agency comes to serve the interests of the industry it is supposed to regulate rather than the public interest it was created to protect.
regulatory compact
An implicit or explicit agreement between a regulated monopoly and a government regulator, in which the monopoly accepts rate and service regulation in exchange for protection from competition.
revolving door
The movement of individuals between roles in regulatory agencies and the industries those agencies oversee.

S

satisficing
The strategy of searching through available options until a good-enough solution is found, rather than evaluating all possibilities to find the optimal one. Satisficing reflects bounded rationality: people do not have the time, information, or cognitive capacity to optimize.
scrip
Company-issued currency or vouchers that could only be spent at company-owned stores, historically used by employers to keep workers’ wages circulating within the company economy.
selection bias
A statistical error that occurs when the sample used in a study is not representative of the population it is meant to reflect, leading to skewed or misleading conclusions.
selectorate
In selectorate theory, the full set of people whose nominal support a leader depends on to hold power, from which the winning coalition is drawn. The selectorate is large enough to give the leader alternatives if individual coalition members defect, but most members receive only public goods rather than private benefits.
selectorate theory
A framework explaining political and organizational behavior in terms of the size of the winning coalition a leader must satisfy to remain in power. Leaders with small winning coalitions can maintain power by distributing private benefits to a few key supporters; those with large winning coalitions must provide broader public goods.
sharecropping
A system in which tenant farmers gave a portion of their crop as rent to a landowner, often perpetuating poverty and dependency through tied credit arrangements.
sideloading
Installing software on a device from sources other than the platform operator’s official app store, bypassing the operator’s review and commission systems.
slotting fees
Payments made by consumer goods manufacturers to retailers for guaranteed shelf placement, functioning as a barrier to entry that favors established brands with resources to pay.
social legibility
The degree to which an individual’s characteristics, credentials, or identity are visible and interpretable within the categorization systems used by institutions making decisions about them.
sodomy laws
Laws criminalizing certain sexual acts, historically applied primarily to same-sex relations; declared unconstitutional in the United States by the Supreme Court in 2003.
sole proprietorship
The simplest business structure, in which a single individual owns and operates a firm with no legal separation between personal and business assets, meaning the owner bears unlimited personal liability for the firm’s debts and obligations.
speculative bubble
A rapid, self-reinforcing rise in the price of an asset driven by expectations of further price increases rather than underlying value, followed by an equally rapid collapse when those expectations reverse.
stand your ground
Laws that remove the traditional legal duty to retreat before using lethal force in self-defense, applicable wherever a person has a legal right to be.
standards body
An organization that develops, publishes, and maintains technical standards through consensus processes among industry participants, governments, and other stakeholders.
standards capture
The process by which a dominant firm shapes a technical standard to its own advantage, either by controlling a standards body, by implementing proprietary extensions that become de facto requirements, or by refusing to implement open standards that would reduce switching costs.
structural lock-in
The condition in which technical, contractual, or network dependencies make switching away from a platform or system prohibitively costly, regardless of dissatisfaction with it.
suppressive person
In Scientology, the designation applied to individuals who publicly criticize the organization or leave it in ways deemed harmful to its interests; those so designated may be cut off from contact with current members, including family members who remain in the organization.
surveillance economy
The business model in which user behavior data is systematically collected, analyzed, and monetized through targeted advertising and data brokering.
switching costs
The financial, practical, or social costs incurred when changing from one product, service, or platform to another, which incumbents exploit to retain users.

T

tax haven
A jurisdiction that imposes low or zero taxes on foreign income, profits, or wealth, attracting capital and corporate registrations that have little or no genuine economic activity there.
TESCREAL
An acronym for Transhumanism, Extropianism, Singularitarianism, Cosmism, Rationalism, Effective Altruism, and Longtermism. These movements are not identical, but they share a belief that a small group of enlightened individuals can and should steer humanity’s long-run trajectory, which tends to position democratic institutions as obstacles rather than safeguards.
tithe
A mandatory contribution of one-tenth of income or produce, historically collected by the Catholic Church as a form of taxation with legal enforcement in many countries.
tontine
A historical financial arrangement in which subscribers contribute to a shared fund and receive income from it; as each subscriber dies, the remaining subscribers’ shares increase until the last survivor inherits the whole. Banned in most jurisdictions because the structure creates incentives to hasten the deaths of other participants.
tort liability protection
Tort liability is legal responsibility for harm you cause someone outside of a contract; tort liability protection (also called liability shield or immunity) is when a law specifically exempts you from being sued for such harms.
transfer pricing
The setting of prices for transactions between subsidiaries of the same multinational corporation, which can be manipulated to shift profits to low-tax jurisdictions by under- or overcharging for goods, services, or intellectual property licenses.

U

union avoidance
A formal management discipline focused on preventing workers from forming or joining unions, employing specialized consultants to structure workplaces, communications, and compensation to reduce organizing incentives.

V

vertical integration
A business strategy in which a company controls multiple stages of its supply chain, from production through distribution and retail, reducing dependence on external parties and raising barriers to competitors.

W

waqf
An Islamic charitable endowment in which property is dedicated in perpetuity to a specified purpose and cannot be sold or inherited; historically used to fund mosques, hospitals, schools, and public infrastructure across the Middle East, Central Asia, and North Africa.
winning coalition
In selectorate theory, the minimum subset of supporters a leader must keep satisfied to remain in power. The smaller the winning coalition, the more each member must be rewarded with private benefits rather than broad public goods, creating strong incentives for corruption and cronyism.
workplace politics
The process by which individuals and groups within an organization attempt to influence decisions in their favor when there is no shared agreement on goals or priorities. Workplace politics is not exceptional behavior by bad actors; it is the normal mechanism by which organizations resolve disagreement.

X

Y

yellow press
Late nineteenth and early twentieth century sensationalist newspapers known for emotionally manipulative coverage that prioritized circulation over accuracy, named after the Yellow Kid comic strip.

Z

Why Don't You Just…

In 2013, the United Kingdom launched Universal Credit, a welfare reform ostensibly intended to simplify the benefits system by merging six separate payments into one. The new system was designed to be applied for online, but many claimants had no reliable internet access. Those who did often lacked fixed addresses, which the system required before it would register them. without registration they couldn’t receive payments; without payments they couldn’t maintain an address, and without an address they couldn’t register.

Government officials and advice workers, when presented with this loop, would sometimes suggest that claimants “just go to the library” to use a computer. They did not know—or had not thought through the fact—that some libraries require a membership card to use their computers, that membership cards require proof of address, and that the Universal Credit application times out and loses your work if you do not complete it in a single session. The word “just” was doing an enormous amount of lifting.

“Just” does that in a lot of conversations. When someone with relative power is told about a problem they have not personally faced, a common response is to suggest an individual solution: “Why don’t you just move to a better neighborhood, just report the harassment to HR, just open a bank account, just apply for a scholarship.” The suggestion is not usually made in bad faith. It is made because the person offering it has, at some point in their life, moved, or reported, or opened, or applied, and found the process manageable. What they cannot see is what made it manageable for them and what makes it unmanageable for someone else.

Peggy McIntosh described this as an invisible knapsack: a set of advantages so routine to those who carry them that they are not experienced as advantages at all. You do not notice that your accent marks you as non-threatening, that your name gets you callbacks, that your neighborhood has a library, or that the official you need to speak to treats you as a legitimate claimant rather than a probable fraud. You don’t notice it because none of it has ever caused you a problem. The person offering the “just” solution is usually describing what they would do, which is not the same thing as what the other person can do.

The most obvious thing “just” conceals is cascading prerequisites. Many systems assume that the person using them already has a set of prior resources in place. FIXME: give an example other than Aadhaar

A second thing “just” hides is the cost of the transaction. Taking a day off work to stand in line costs money that people without savings cannot spare. Challenging a decision by a government agency or a platform company requires knowing how to challenge it, having the literacy and the time to fill in forms, and being willing to risk the relationship with the institution you depend on. The last point matters more than it might seem: if the institution denying you a benefit is the same institution you are hoping will pay your rent next month, asserting your rights has a price.

A third thing the word hides is cognitive load. Sendhil Mullainathan and Eldar Shafir spent years studying what poverty does to decision-making, and their conclusion was not what most people expect. Poverty makes people worse at decisions in the same way that keeping someone awake for twenty-four hours makes them worse at decisions: it depletes a finite resource. Managing an unpredictable income, keeping track of which bills are overdue, or calculating whether buying the cheaper item in bulk saves more than it costs to store consumes cognitive bandwidth that is then not available for navigating bureaucratic systems. The person telling someone to “just” do something is typically not paying this tax, so they have the mental capacity to think about the solution that the person being advised does not.

The invisible knapsack metaphor is useful, but there is a complementary image that captures something different: the invisible staircase. People who have always lived in buildings with functioning elevators do not think about stairs. When they suggest that someone in a wheelchair “just take the stairs,” they are not being cruel. They have simply never had to think about stairs, because stairs have never been an obstacle for them.

This is why the “why don’t you just” response so often lands badly. It is not merely that the solution is wrong; it’s that is reveals that the person offering it has never had to think about the problem. They are announcing their distance from it, and the person who has been living inside the problem knows, at once, that the advice is coming from outside.

South Africa’s apartheid pass laws required Black South Africans to carry a document showing they had permission to be in a given area. The document had to be produced on demand by any police officer. If a man had left his pass book at home, or lost it, the advice “just get a replacement” required traveling back to the township administrative office where the original had been issued—not the office nearest to where he was stopped. This meant taking time off work and navigating a bureaucracy that had been designed to make compliance difficult.

A further problem with the “just” response persists even when the individual solution actually works. If a person successfully navigates the broken system, the system sees this as proof that it is functioning correctly. Virginia Eubanks documented this pattern across multiple American welfare systems: people who managed to get the automated system to approve their application had often spent dozens of hours learning its quirks, calling hotlines, finding advocates, or simply applying repeatedly until an error cleared itself. None of that effort was visible to the system. The system recorded a successful outcome. The structural problem—that the system was failing the people it was supposed to serve—was continuously obscured by the individual successes of the most resourceful claimants.

The pattern is particularly vicious on tech platforms. Companies routinely advise users to “just report” harassment, “just use a VPN” to avoid surveillance, or “just switch platforms” when a service degrades. Each suggestion assumes that the user has time, technical literacy, and social capital that many do not have, and none of them fix the underlying system. They transfer the cost of a structural failure onto the person least able to bear it.

Eubanks2018
Virginia Eubanks: Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor. St. Martin’s Press, 2018, 9781250074317.
McIntosh1989
Peggy McIntosh: “White Privilege: Unpacking the Invisible Knapsack.” Peace and Freedom, July/August 1989.
MullainathanShafir2013
Sendhil Mullainathan and Eldar Shafir: Scarcity: Why Having Too Little Means So Much. Times Books/Henry Holt, 2013, 9780805092646.

Who Are You?

In 1879, a young police clerk in Paris named Alphonse Bertillon proposed a solution to a problem that had plagued law enforcement for decades: how do you know if the person in front of you is who they say they are? Before photographs were cheap to reproduce and before fingerprint databases existed, professional criminals could simply give a false name and walk free. Bertillon’s answer was anthropometry: measure the skull, the length of the forearm, and other bodily dimensions that, taken together, were statistically unlikely to be identical in any two people. The system spread across Europe, the United States, and colonies in Asia and Africa, and was the first large-scale attempt to use the body as a database.

Anthropometry had a fundamental weakness: measurements had to be taken correctly by trained operators. By 1900 fingerprinting was replacing it almost everywhere because fingerprints were more reliable and required less skill to record. But the desire of states and institutions to pin individuals permanently to a record did not die with the calipers.

The history of identity management is partly the history of states trying to solve what the political scientist James C. Scott calls the legibility problem. A state cannot tax, draft, or police people it cannot identify. Medieval English peasants might know themselves as “John the Miller’s son from the village by the ford”, but that description doesn’t survive a move to a city or a change of occupation. Surnames became standardized in Europe partly because governments needed them. The same logic produced house numbers in Paris and Vienna, censuses across the colonial world, and passports that began as occasional travel documents issued by monarchs and became, by the twentieth century, a requirement for crossing most international borders.

The problem is that the state’s desire for legibility doesn’t have a built-in limit. The state that registers births so it can provide schooling can use that register to conscript soldiers. The government that issues identity documents to allow people to vote can use those documents to deport people. Once a population register exists, every subsequent administration can use it for whatever purpose it finds useful.

Colonial governments exploited it systematically. The British introduced population registers, caste certificates, and tribal designations across India, Africa, and Southeast Asia that served double purposes: administration, taxation, and census on one hand, and on identifying potential troublemakers and restricting movement on the other. South Africa’s pass laws, introduced gradually from the eighteenth century and formalized under apartheid after 1948, required Black South Africans to carry a reference book at all times. The book recorded their employer, their designated “homeland,” and their permission to be in urban areas. Police could stop anyone and demand the book; failure to produce it meant arrest. The pass system was one of the most sophisticated identity management infrastructure projects in history, and it was designed entirely to restrict freedom of movement and force labor into mines and factories at wages set by the government.

Keith Breckenridge’s history of South African biometrics traces a direct line from the pass system to the world’s first large-scale biometric population register, introduced in South Africa in 1986. Fingerprints were added to the passbook because fingerprints are harder to falsify than signatures and do not require literacy. A system designed to prevent forgery of internal passports became, almost automatically, one of the most comprehensive biometric databases in the world at the time.

Simone Browne’s analysis of the surveillance of Black people in the United States and elsewhere argues that modern surveillance practices such as facial recognition predictive policing, and airport pat-downs are not new inventions applied to race as an afterthought. They grew directly out of earlier practices designed to control enslaved and colonized people. The identification and tracking of freed people after the Civil War and the systematic photography of colonized people both influenced later uses of photography in criminology. The tools of identity management and the tools of racial control were never separate projects.

This creates a dilemma that does not have a clean solution. Democratic participation depends on being able to identify voters. Electoral systems need to verify that voters are eligible residents and prevent people from voting twice. This requires some form of voter registration, and voter registration requires identity documentation. In the United States, the history of voter registration is entangled with the history of voter suppression: poll taxes, literacy tests, grandfather clauses, and, since the 1990s, photo ID requirements that are nominally neutral but fall disproportionately on communities less likely to hold a driver’s license: the poor, the elderly, and people of color. The registration systems needed to enable political participation have been weaponized to prevent it.

But the dilemma goes deeper than that. Communities that have historically been targeted by state identity systems have rational reasons to distrust those systems. If the government has used population registers to intern Japanese-Americans to deport undocumented immigrants in mass raids, people are right to be suspicious of any new identity system. Even if today’s government only intends to use it for good, it cannot constrain promise the behavior of future governments.

This dilemma is now one of the biggest challenges facing democracy. The communities most harmed by identity management in the past have the best reasons to distrust registration systems, but that distrust keeps them from participating in the political processes that would let them constrain those systems. Fear of being rounded up is itself a tool of disenfranchisement.

India’s Aadhaar system, launched in 2009, has enrolled over 1.3 billion people in a biometric identity database based on fingerprints and iris scans linked to a twelve-digit number. Its designers argued that a universal, biometric identifier would eliminate fraud and exclusion in government benefit programs by making identity verification objective. The argument was partly correct: Aadhaar has reduced certain kinds of fraud and brought some previously excluded people into the formal economy. It has also created new forms of exclusion. People who cannot authenticate are cut off from food rations and pension payments, and the database itself represents a concentration of sensitive biometric information that, if breached, cannot be reset: you can change your password, but you cannot get new fingerprints.

The European Union’s General Data Protection Regulation (GDPR), which came into force in 2018, treats biometric data as a special category requiring explicit consent and restricts its collection and processing. The regulation is imperfect and inconsistently enforced, but it exists because privacy advocates organized across member states over roughly two decades to demand it. It is the most significant constraint on commercial biometric data collection in the world. The tech industry’s response has been intrusive pop-ups designed to make users blame governments for safeguarding their privacy rater than companies for trying to collect information for resale.

What abusive identity management has been constrained, those constraints have been the result of organized political activity. South Africa’s pass system was not abolished because its administrators became enlightened; it was abolished because the resistance movement grew strong enough to make it unsustainable. The constraints on Aadhaar came from litigants and activists who brought cases, and the Voting Rights Act of 1965 in the United States came from marchers, organizers, and a political coalition that made the status quo more costly than change. The center always wants more information. The question is always whether the people most likely to be harmed by that information have enough power to say no.

see the whole series · email me

Breckenridge2016
Keith Breckenridge: Biometric State: The Global Politics of Identification and Surveillance in South Africa, 1850 to the Present. Cambridge University Press, 2016, 9781107434899.
Browne2015
Simone Browne: Dark Matters: On the Surveillance of Blackness. Duke University Press, 2015, 9780822359388.
Cole2002
Simon A. Cole: Suspect Identities: A History of Fingerprinting and Criminal Identification. Harvard University Press, 2002, 9780674010024.
Scott1998
James C. Scott: Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed. Yale University Press, 1998, 9780300078152.
Torpey2000
John Torpey: The Invention of the Passport: Surveillance, Citizenship and the State. Cambridge University Press, 2000, 9780521634939.