📖 Here’s what I’m reading this week

• Will you forget what it means to be human? That’s the contention of a great essay arguing that the AI poses a real threat not because of rogue robots but because “it offers answers faster than humans can ask the questions that help them contemplate their existence.” Live the the existential questions, don’t passively accept the antiseptic answers of an algorithm. (More)

• Are you willing to rely on a technology that will always be wrong some of the time? The future of the web, commerce, service delivery, etc. are all being reoriented by unreliable chatbots. In making this shift, we’re building upon a new infrastructure with a lot of pot holes, requiring a kind of on-going vigilance and default doubt. As Matteo Wong writes, “AI products could settle into a liminal zone. They may not be wrong frequently enough to be jettisoned, but they also may not be wrong rarely enough to ever be fully trusted.” (More)

• Did a chatbot tell you to say that? According to new research, we’re starting to mimic the language of chatbots, morphing ourselves to sound more the machines. Along the way, we’re losing “verbal stumbles, regional idioms, and off-kilter phrases” that communicate our humanity. Take this (totally relatable) example from the piece: “I’m sorry you’re upset” versus “Hey sorry I freaked at dinner, I probably shouldn’t have skipped therapy this week.” Don’t skip out on therapy, and don’t give up on the authentic, messy communication that makes you human. (More)

• spent hours talking to people who believe large-language models are an all-knowing god. Check out what she learned (More)

This week’s guide is about how to do the next right thing amidst uncertainty. As a reminder, Untangled Guides offer strategies to see your system clearly, anticipate its behavior, and ensure technology works for (not against!) it.

⚒️ How to do the next right thing

In complex systems, you don’t set targets, you set directional goals. Targets foreclose real-time learning and sense-making. As Dave Snowden puts it, “targets become an entrained pattern of expectation and people cease to question.” This is particularly problematic in complex systems that don’t behave how you might expect.

Once you’ve set a directional goal, you don’t build a long-term plan, you build adaptive strategies rooted in sense-making. In my guide, “How to determine if you’re dealign with a complex system,” I distinguished between complicated systems (e.g. a car engine) and complex systems (e.g. traffic). Well, in the former, it makes sense to ‘sense-analyze-respond’ to use Snowden’s terms. A car engine is based on well defined principles of physics and engineering, so if you understand how the components interact, you can predict the engine’s behavior and fix it if it breaks down. You simply need to collect the relevant facts, analyze them, and apply the appropriate procedure.

But complex systems don’t abide by cause-and-effect. Traffic changes its behavior based on information from its environment. Focusing on one car won’t teach you much about the entire system because what matters is the interactions between them, and the patterns they form. In this case, you would follow the protocol of ‘act-sense-respond,’ so that you can sense how the system responds to your intervention and then adapt accordingly.

Read more

Last week, I outlined how to determine if you’re dealing with a complex system, which behave in unpredictable and uncertain ways, and wherein participants adapt to each other’s actions. Today, I’m sharing a Guide that will help you engage your system amidst uncertainty.

Uncertainty is uncomfortable. We don’t like it. We want to control it. As a result, we often conflate it with risk so that we can ‘manage it.’ But, as Vaughn Tan makes clear, risk is the management of an unknowable thing; whereas uncertainty is more akin to ‘not knowing’ or when you can’t manage that unknowable thing. This is a subtle but important difference because we get in trouble when we mistake the two. For example, as I’ve argued in the past, ‘AI alignment’ work is predicated on the idea that we can align AI with our needs and wants — gotta stop those robots from going rogue! But we hold conflicting values and needs that cannot be optimized away. In this way, alignment isn’t a risk companies can manage, it’s true uncertainty hiding in plain sight.

Subscribe now

In mistaking uncertainty for risk, we play a dangerous game. We start to behave as if we just know what to do; that all we need are experts making analytical, reasoned decisions. But if we’re dealing with true uncertainty, we need collectives that bridge across different sectors and areas of expertise (more on this in a future Guide!) and constraints. As Alicia Juarrero explains in Context Changes Everything, constraints are “entities, processes, events, relations, or conditions that raise or lower barrier to energy flow.” In short, constraints modify the context by which decisions are made, actions are taken, information is shared, behaviors are expressed, etc.

Read more

I write about ‘systems,’ ‘sociotechnical systems,’ and ‘complex adaptive systems,’ which, is, uh, confusing. Let’s untangle these terms and help you determine what you’re dealing with.

A ‘system’ exists wherever people are trying to meet a need or solve a problem. (More) For example, a group of funders working to align the strategies of their grantees toward collective action is a system. If this system is mediated by technology — say the grantees are working to advance ‘fair and responsible AI’ — it becomes ‘sociotechnical.’ Sociotechnical is just a fancy word for how technology shapes people, and people shape technology.

But at a systems level, we’re not looking simply at how people shape technology, we’re looking at how systems of race, gender, status, and power shape technology (Untangled Deep Dive), and in turn, how technology shapes these systems (Untangled Deep Dive). This mutually shaping process creates dynamics within a system that contribute to its overall behavior. To use our example, if the funders and grantees don’t understand how these hidden dynamics affect the pursuit of ‘fair and responsible AI,’ they’ll either design programs that create ‘unintended consequences’ or worse, perpetuate the status quo in the name of technological innovation and progress.

Subscribe now

But here’s the crux: most systems aren’t just sociotechnical, they’re complex and adaptive too. This means that they behave unpredictable and uncertain ways and participants adapt to each other’s actions. As detailed in the Cynefin Framework by Dave Snowden, complicated systems respond to cause-and-effect in a linear and predictable manner, wherein if I do X, Y will occur, and we can verify that X caused Y. By contrast, in a complex adaptive system:

• Data and/or knowledge are incomplete or contradictory.

• You can’t predict the outcome. Cause-and-effect doesn’t exist. Outcomes are more often the result of non-linear dynamics. (Untangled Deep Dive)

• The component parts of the system are interdependent (More) and the macro behavior of the system emerges from how the parts adapt to one another, and the dynamics this creates (More)

An oft used example is the distinction between a car and traffic. A car engine is complicated. Its operation is based on well defined principles of physics and engineering. If you understand the components of how they interact, you can predict the engine’s behavior and fix it if it breaks down. Whereas a car’s engine might be complicated, traffic is a complex system. Traffic changes its behavior based on information from its environment. Focusing on one car won’t teach you much about the entire system because what matters is the interactions between them.

Read more

This issue is part of a series of practical guides. They offer you the lenses to see your system clearly, and the levers to collectively change it. The series includes:

• Guide 1: How to see data as socially constructed

• Guide 2: How to see technology as socially constructed.

• Guide 3: How to analyze the frames and metaphors that hide power in technology

• Guide 4: How to anticipate technological shifts in power.

• Guide 5: How to anticipate technology’s impact on different communities.

• Guide 6: How to see your system clearly.

• Guide 7: How to anticipate different community uses of technology.

• Guide 8: How to take interdependent (not independent) action.

• Guide 9: How to leverage feedback loops in your system.

• Guide 10: How to think & act ecologically.

• Guide 11: How to map backward from the future.

Today, I’m launching Guide #12: How to define your system. Drawing boundaries around a system — defining it! — is really hard.

Sometimes, we’re too broad — everything is connected to everything! But the bigger risk is in being too narrow. The team repeatedly misses its goals, so we focus on the team, when it turns out, the underlying reason they continue to miss their goals is connected to organizational structures, how decision rights are allocated, what is rewarded and given status within the organization, etc. Or maybe the system extends beyond the organization because the team dynamic is actually influenced by a partner relationship or funder requirement.

When we define the system too narrowly, and then intervene, it surprises us — and we call these ‘unintended consequences’! But the deeper truth is that we inserted artificial boundaries and didn’t see the complete picture.

Subscribe now

The ‘boundaries’ we insert often reflect our position in the organization, our training, and our life experience. For example, say an algorithmic decision making system repeatedly harms the same communities. In response, product teams and engineers might work together in an attempt to design the technology in a way that is more ‘responsible’ or ‘fair.’ This narrow focus on the technology itself is understandable given the role and training of engineers and product managers.

Moreover, it’s not totally insane to consider fairness a reasonable goal if you’re a white guy who believes in meritocracy. If unfairness is simply the result of “fallible human biases on the one hand, and imperfect statistical procedures, on the other” as I wrote in “Beyond Minimizing Harms” (Untangled Deep Dive), then it’s plausible to make meaningful change by focusing on the algorithmic system. Together, they’ll draw a nice li’l boundary around the technology and the process of creating it, and then try to make it fairer.

Read more

Overview

Changing your system requires mapping backward from the future.

We know that social systems are uncertain, unpredictable, and complex (Untangled Deep Dive).

Two things from this follow. The first is that other actors in your system will not respond to your strategic intervention passively or predictably. Your actions will shape their actions and the system you’re in, and vice versa. This requires taking interdependent (not independent) action. (Untangled Guide)

The second is that strategic interventions that map forward in time -- as if the future is a neat theory of change away -- must be replaced by those that map backward. This is both because complex systems don’t follow cause-and-effect assumptions, and it’s because our beliefs, values, and stories constrain our conception of what’s possible. Right, the past is always alive in the present, and systems seem impervious to change until they are not.

Think of it this way, the only way to get to the future is through the present moment. We’re creating our future each day through our decisions, actions, choices, etc. This path has been grooved by those who came before us — the relationships they formed, the institutions they built, and the norms, values, and beliefs they propagated. These grooves are subtle but they reflect and then instantiate power relations. Who has it, and who doesn’t? Who benefits from this version of reality? Who has something to lose if we changed society’s default settings, and charted a radically different path? If you start with questions like this, you’re likelier to see the grooves and figure out how they were created in the first place. It’s also in seeing these grooves clearly, we can see how the past is alive in the present, constraining what we perceive as possible.

Read more

🔗What I’m reading this week

• Google is starting to embed ‘AI mode’ in search. (More) Want to understand the social consequences of swapping search for chatbots? (Untangled Deep Dive)

• Elon Musk & Sam Altman want to bring chatbots, finance, gaming, and a social network together. (More) Why? They need your data. (Untangled Deep Dive)

• Meta’s new ‘digital companions’ are engaging minors in sexually explicit discussions. (More)

• Researchers ran deeply unethical experiment on Reddit users. (More)

• “From Artificial Intelligence to Abundant Imagination,” a lecture by scholar Ruha Benjamin. (More) How separating technical elements from social ones like race, gender, and power helps create radical different futures (Untangled Deep Dive). How to work our way towards utopia (Untangled Deep Dive).

Share

Overview

Technology alters ecosystems. I’m not the first to make this point. Neil Postman argued long ago that “Technological change is not additive; it is ecological,” explaining that “A new medium does not add something; it changes everything.” Postman uses the printing press as an example:

“In the year, 1500, after the printing press was invented, you did not have old Europe plus the printing press. You had a different Europe. After television, America was not America plus television. Television gave a new coloration to every political campaign, to every home, to every school, to every church, to every industry, and so on.”

In short, systems don’t remain the same, but just with ‘more technology’; they become altogether different systems. For example, the use of image and video-based algorithmic recommendation systems isn’t just society plus those technologies. TikTok is changing the kind of relationships we have with one another — disrupting the social graph, ushered in by Facebook and earlier social media sites, and encouraging more para-social relationships.

Read more

🔗What I’m reading this week

• A debate over whether chatbots truly understand between Emily Bender, professor of linguistics, and Sébastien Bubeck from OpenAI. I’m squarely on Team Bender in this debate. (More) (Untangled Deep Dive)

• A sharp essay on the relentless attempt to override the complex, messy world of human beings with computational logic. (More) (Untangled Deep Dive)

• What if we all agreed to use the phrase ‘computational information processing’ instead of ‘artificial intelligence.’ Please? (More) (Untangled Deep Dive)

• Chatbots are flattening language. (More)

• Herbert Simon’s 1971 seminal essay on attention economics theorizes that information consumes attention. So the critical question when assessing a new technology becomes, “how much information will it allow to be withheld from the attention of other parts of the system.” (More)

• Did you talk to a chatbot on Instagram this week that lied to you about being a licensed therapist? (More)

Subscribe now

⚒️Overview

Ever been stuck inside a feedback loop?

If you see a recurring behavior or output, a feedback loop is likely at work. For example, if within ten-ish minutes of waking up, I don’t have a cup of coffee in my hand, something has gone terribly wrong. As Donella Meadows explains in Thinking in Systems, that’s a feedback loop: my energy levels are low, so I increase my coffee intake toward my desired state. “It is the gap, the discrepancy, between your actual and desired levels of energy for work that drives your decisions to adjust your daily caffeine intake,” as Meadows writes.

Sometimes I might have a third cup of coffee in the afternoon, and I realize I’ve gone too far. I didn’t just close the gap between my actual level of caffeine and desired state, I surpassed it. In those moments, I’ll take a beat, and lay off coffee, until I get closer to my desired state. This is balancing loop. Balancing loops lead to some equilibrium or set-point. They are self-regulating, stabilizing, and goal seeking, and they are all around us. A thermostat is a balancing loops. How planes fly — adjusting its course every so slightly toward the end destination in response to data — is a balancing loop. Our systems of checks and balances are (in theory!) a balancing loop.

The second type of feedback loop is a reinforcing loop. Reinforcing loops accelerate change, good or bad. For example, an interest bearing savings account is a good reinforcing feedback loop — left untouched, the balance of the account earns interest, which increases the balance, which increases the total generated from interest etc. That’s a magic li’l loop that, as Meadows points out, might look linear at the start but grows faster and faster over time. It becomes exponential.

Read more

Why do system strategies fail all the time?

• We see only part of the system. (More)

• We think data are objective reflections of reality. But they’re made through interactions in the world. So we misunderstand what data say, and what they don’t, and in turn, what problem we’re actually solving. (More)

• We think technologies are ‘a solution.’ But this engineering mindset misunderstands the role of hierarchies, power dynamics, status, cultural norms etc. So power and culture — two very big things! — aren’t incorporated into our analysis of the system. (More)

• We misunderstand how technologies shape a system — they aren’t additive, they’re ecological. (More)

• We think we can predict how our intervention will impact the system but the system is complex, uncertain, and behaves in non-linear ways. (More)

I could go on.

Share

But one of the biggest reasons system strategies fail is that we don’t see the interdependent nature of the system. We don’t see the connections between the parts — technology, people, organizations, etc. — that shape the system’s behavior. We’re entangled in relationships that shape our actions, which shape our relationships, which shape the system, which shapes our actions. Around and around we go!

In biology, this is called “structural coupling.” As Patrick Hoverstadt and Lucy Loh explain in Patterns of Strategy, it is “the process whereby an organism interacts with its environment in such a way that the organism changes the environment and the environment changes the organism.” The macro behavior of the system emerges from these interactions and the dynamics they create.

But we’re not often aware of how these relationships — without any intervention on our part — shape the direction of our strategy and the evolution of our organization. “Just as in biology,” Hoverstadt and Loh explain, “evolution doesn’t demand the hand of a conscious designer, so in organizations a structural coupling view of strategy doesn’t demand that there is any conscious strategy at work.” Systems strategy requires that we get to know the relationships that shape our behavior, and the path we’re on.

Read more

Overview

The same underlying technology can be used by very different communities shaped by different cultural norms, leading to very different outcomes. Often then, the best way to understand technology’s likely impact on society is with an anthropological lens — it’s in understanding the beliefs, rituals, and practices of communities that we can anticipate future use cases more clearly.

Share

E.g.

For example, some crypto communities see blockchain technology as a mechanism for reinvigorating democratic governance, while others see it through a libertarian, anti-state lens.

• In my interview with Nathan Schneider, ‘From Co-ops to Crypto,’ we discussed how the Occupy Wall Street movement shaped his imagined use cases of blockchain technology. In particular, OWS encouraged Schneider to see blockchains as a mechanism for reinvigorating democratic governance.

• By stark contrast, in ‘Crypto is not decentralized,’ I showed how early crypto projects like Bitcoin were underpinned by a libertarian, anarcho-capitalist belief system, shaped by the values and beliefs of Web 1, which were themselves rooted in the counterculture movements of the 1960s-70s.

• In ‘What connects crypto and the Netflix show ‘Love is Blind’?' I argued that as crypto is becoming fully financialized, speculation is increasingly underpinning the cultural practices of crypto communities.

• In “The Church of Number Go Up,” I argued that while some crypto projects represent casino culture, and others reflect the culture of a ‘strict church,’ there are new entrants that are far more pragmatic and secular who are focused on generating value over policing collective belief.

Across these instances, we see the same underlying technology, but very different communities shaped by different cultural norms and movements, leading to very different use cases.

Course: Sociotechnical Systems Change, April 26-27*

This is the last week to register for cohort #2 of my course. After two live, online, interactive sessions, you’ll be able to:

• See clearly how power shapes your system.

• Anticipate how emerging technologies like AI shape (and are shaped by!) your system.

• Map the complex dynamics of your system.

• Develop strategies to reimagine and change your system.

You’ll come away with key skills to analyze and change the system you’re in and a 90-page workbook full of 30 exercises and tools to help you put them into practice. (Register)

Resources

• From Co-Ops to Crypto

• Crypt is not decentralized

• What connects crypto and the Netflix show ‘Love is Blind’?

• The Church of Number Go Up

Approaches

Think like an anthropologist

The next time you see a technological phenomenon take off, ask what community beliefs, rituals, and practices are likely to shape it. It’s possible you might even notice the rituals first — because these are a key part of how the new phenomenon travels through networks. This requires thinking like an anthropologist — making the familiar strange, and the strange familiar. While it’s not always obvious what’s strange and what’s familiar — we are not separate from social and cultural systems — Gillian Tett offers a great question in her book Anthro-Vision: A New Way to See in Business and Life: “If I was to arrive in this culture, as a total stranger, or as a Martian or child, what might I see?”

Now, adopt the role of alien and apply this thought experiment to artificial intelligence in the U.S. in 2025. What do you see? (Add your answer in the comments!)

Okay, that’s it for now,

Charley

Overview

Once you start interrogating and poking holes in the frames and narratives that hide power in technology, you’ll find a system. Basically, a ‘system’ exists wherever people are trying to meet a need or solve a problem. So a neighborhood trying to foster community or improve safety is a system. A constellation of service organizations trying to reduce homelessness is a system. A group of non-profits trying to ensure ‘AI’ serves the public interest is a system.

Systems are all around you but often invisible. You just see a recurring output, and, if you care about changing the system, you have to make visible how the interactions of the different parts of the system (people, technologies, organizations, cultural norms, etc.) produced it. And you have to do so in a way that gets the key stakeholders in the system on the same page about what it looks like, how it behaves, etc. Seeing your system clearly is about developing enough awareness so that you can strategically intervene amidst uncertainty — and, intervene in a way that takes account of how the system and its stakeholders are likely to respond. Which. Is. Hard.

Share

Some groups only see part of the system. Different groups will have different ways of seeing (e.g. different backgrounds, expertise, epistemologies, etc.), which will shape how they see and what they ignore. Others will deny that they’re part of the system -- ignorance really is bliss when the system is tilted in your favor. How different groups make sense of data (e.g. what they can say, what they can’t) and technology (e.g. the values and beliefs reflected in it, how it impacts different communities, how it shifts power, etc.) will fundamentally shape how they see their system. Moreover, some systems are harder to see clearly because they’re complex and uncertain, behaving in ways you can’t predict. If people are involved, then the system includes hierarchies, power dynamics, cultural norms, and status, which tend to hide in plain sight.

Seeing the system clearly requires not just that we interrogate the frames and narratives that hide the real problem but that we map the stakeholders, cultural norms, technologies, and power dynamics that make up the system, and keep it in place. (This overview doubles as a plug for my course, Sociotechnical Systems Change, because we can’t meaningfully change our system without seeing it clearly.)

Learn More & Sign Up

E.G.

In 2014, Amazon developed an algorithmic recruitment tool to support its hiring process. Amazon hoped the tool would increase efficiency while avoiding bias against women by eliminating social identity markers. The recruitment tool was trained on the previous 10 years of employment data, and technology was —and is — a male-dominated sector. So the algorithm assigned higher scores to men.

While we’re often quick to focus narrowly on the algorithmic impact, there is a system producing this outcome, including things like hiring practices and corporate culture, data that reflect and reproduce societal biasses, gender norms, ideas about meritocracy, etc. But if you’re part of the system, it might be challenging to see your hiring practices clearly. Or if you’re a trained engineer or statistician, you might see the biases as statistical, not social and systemic. Or if you’re a man who benefitted from this system, you might not see the social and cultural norms that nudged you along the way and defend the existing hiring practices as ‘fair’ or meritocratic.’ Or you might see them clearly, which creates a different set of dynamics.

Read more

Overview

How one experiences a new technology depends on their societal position or status in society. Indeed, technologies tend to disproportionately harm those at the margins of society. This is exacerbated by the propensity to design for scale and, in turn, the median user, rather than leaving space for contextually sensitive technologies, which might account for a range of societal positions. Or, heaven forbid, design products or govern systems from the perspectives of those on the margins!

This guide will help you anticipate technology’s impact on different communities and includes:

• An example that illustrates the idea.

• Additional resources to dive deeper.

• Practical approaches to apply the Guide in your working context.

E.G.

Blockchains record your transactions on a public ledger. But those transactions aren’t linked to your identity — they are linked to an alphanumeric, private address. Anyone can see your transactions but they don’t know that it is you buying those medications. That’s the idea of pseudonymity.

In ‘Is pseudonymity the answer?’ I drew upon the work of Michelle Lamont and Alice Marwick to argue that whether pseudonymity protects your safety or your power depends on your position in society. When pseudonymity is taken away, it’s worth asking whether further visibility reveals your power or makes you vulnerable. As a white guy who writes a newsletter, visibility is a good thing. But that fact reveals my societal privilege. For many, visibility means vulnerability or precarity.

Resources

• Is pseudonymity the answer?

• Tokenizing Creators

• Let’s imagine an alternative metaverse

• Fragment the media! Embrace the shards!

• AI is a mirror. What can it show us?

Approaches

Center the margins. Since technologies are designed for scale and the median user, those on the margins are often disproportionately harmed, and offer unique perspectives. Ask yourself questions like:

• What groups have historically been disproportionately harmed or excluded by the sociotechnical system?

• How might you center the perspective of marginalized groups and those disproportionately harmed in the design, deployment, and governance of the sociotechnical intervention?

In such an approach, members of marginalized communities would lead the design and governance process. They would make decisions about how their data is governed, which technologies are used and banned, how abuse is handled, and how their identity is captured.

Okay, that’s it for now,

Charley

Overview

Technologies can shift who has power. Lewis Mumford, a historian and philosopher of technology, divides technologies into “authoritarian technics” and “democratic technics.”

• Democratic technologies are “human-centered, relatively weak, but resourceful and durable.” These are tools; they are used by people in service of their objectives. The objectives might still be really bad — a hammer can be used as a weapon — so it’s the user’s objective that matters.

• Authoritarian technologies are “system-centered, immensely powerful, but inherently unstable.” These are technologies that become a system — which then acts on people, rather than the other way around: power shifts from the people, to the system.

Technologies can empower people, putting the objectives of their users first. Or they can shift power from the individual user to the broader technical system. Take the Moses’ overpass again from “How to critically analyze technology.” The story isn’t simply that Moses embedded his politics into the overpasses, but that together, the overpasses became a system that regulated behavior. As I referenced in Technologies encode politics Langdon Winner referred to this as “specific ways of organizing power and authority.” Anticipating technological shifts in power, then, starts by analyzing how technologies become systems that shape power.

Read more

We make a problem through frames, narratives, and rhetorical power. The frames we select — ”fraud,” “empowerment,” etc. — make certain elements of a narrative salient, and hide or downplay others. Frames aren’t neutral — indeed, they’re often intentionally created and marketed. This is why they’re extremely powerful. Ya know how we say “climate change” rather than “global warming?” That was a strategic choice. In tacitly adopting a frame, we’re aligning ourselves with a set of interests, values, and politics, often without knowing it.

Share

Frames and narratives are especially powerful tools amidst technological change and uncertainty -- when our expectations aren’t anchored to a predetermined future. It’s in these moments -- e.g. the rise of ‘AI -- wherein, according to Beckert and Bronk in Uncertain Futures, “some stabilisation of expectations is required for investment to be made in new products or processes.” “If innovative ideas are to corner the resources needed to make them a reality,” they continue, “they require promissory stories that help coordinate investment.” In short, our sociotechnical future is shaped by market, political, and rhetorical power to offer the stability -- and, in turn, close off plausible futures -- required for investment.

Read more

Guide #1: How to critically analyze data. If you want to understand AI — what it is, what it can say, and what it can’t — and anticipate its impacts, then you need to understand data, and how they are made.

Subscribe now

On to Guide #2: How to critically analyze technology.

Overview

Technologies are shaped by social systems like race and gender, and those technologies in turn shape social systems. This is what’s meant by ‘sociotechnical’ — that we can’t view social and technical systems separately. Critically analyzing technology, then, means analyzing the dynamics at the center of this entanglement, and the stakeholders implicated in creating and changing them.

The way technologies shape people is through ‘affordances.’ You can think of an ‘affordance’ as how technology alters our experience of it -- the way they shape how our bodies are represented, where we focus (or don’t!) our attention, the actions we can (or can’t!) take, the decisions we make, etc. Spend a minute listing all of the ways ChatGPT or an iPhone alters your experience of it (e.g. the color of the applications is known to be addicting) and then map backward to consider why that affordance might exist. The quick (and often correct) answer to this exercise is financial incentives, but that’s not always the case.

Technologies are designed by someone, somewhere, whose beliefs, normative commitments, and operating contexts subtly shape its development. This matters because how technologies are designed and their affect on the world aren’t inevitable or pre-determined. For example, the internet of today is nothing at all like the internet of the 1990s, which emphasized affordances like anonymity and ephemerality. What changed wasn’t a technological change but an accomplishment of companies like Google, Facebook, and Amazon.

Read more

The path from the present moment to radically different futures isn’t clear. When everything feels inevitable and the next step isn’t obvious, we need guides. Guides that help us identify hidden power dynamics in a system. Guides that help us anticipate technology’s impact on disparate communities. Guides that help us critically analyze data and technology. Guides that help us develop strategies that maps backward from the future we want to create.

Today, I’m launching Untangled Guides — a new ‘how to’ offering for paid subscribers that will help you interrogate the system you’re trying to change, and align your actions, decisions, and strategies to the future you want to create. They will be short, sweet, and chockfull of actionable insights and practical approaches.

Here are a few I have in the hopper:

• How to critically analyze technology

• How to identify the frames and metaphors that hide power

• How to anticipate technological shifts in power

• How to anticipate technology’s impact on different communities

• How to take interdependent (not independent!) action

• How to think and act ecologically

• How to anticipate different community uses of technology

• How to anticipate technology’s impact on culture

• How to avoid the ‘scale’ trap

• How to avoid the ‘efficiency’ trap

• How to avoid the ‘transparency trap

• How to envision alternative futures

• How to map and measure backwards, not forwards

• How to map complex systems and anticipate emergent dynamics

• How to set directional goals, not targets and outcomes.

The first guide launches today: How to critically analyze data. Let’s get started, shall we?

Overview

Data are made by you and me. We interact and transact with one another. We do things out in the world. We engage with institutions. We click and scroll online. Right, data are never raw, nor are they neutral or objective. They’re descriptive and diagnostic of the practices and behaviors of people interacting with one another and institutions at a moment in time. To put a pin on it, data aren’t predictive, they are socially constructed and historically content.

This matters because when we accept the premise that historical data can somehow predict the future — and all AI systems accept such a premise — we encode historical social biases in the present, and bring them with us into the future. We become trapped in a feedback loop of our own making, exacerbating existing inequities along the way.

I started with this Guide because step one of imagining an alternative future is reconsidering what data can tell us, and what it can’t.

Read more

Subscribe now

👇ICYMI

March was busy at Untangled HQ:

• I analyzed the interventions offered by research papers addressing the anti-democratic state of the internet.

• I argued that automation is a structural and political problem, not an individualized one, and that power lies with those who frame the problem.

• I offered my take on the potential TikTok ban and explained how researchers embed how we misunderstand AI into the scientific process.

• I explained why Google’s Gemini isn’t ‘woke.’ Rather, Google papered over a systemic problem and it backfired.

• I published an essay about a new category of crypto project — Decentralized Physical Infrastructure or DePin — and how everything is a Ponzi in crypto, until sometimes, it’s not.

On to the show!

One year ago, I used 3,000 words to answer the question: what even is technology? This year, I want to follow up with another doozy: what is data? In this special issue, I outline how, especially in the frame of AI:

1. Data are made by us.

2. Data are classified into value-laden boxes.

3. Datasets are contorted by scale, shortcuts, and mental models

This isn’t an academic exercise. AI is all about ‘the data.’ Who has it, whether it is of high quality, etc. Tech giants like Google, Microsoft, and Facebook have had a head start on the sector because they already hoovered up all our data; OpenAI and other start-ups are testing the boundaries of copyright infringement by scraping the open web for data. We’re also seeing crypo projects finding ways to decentralize data scraping.

This data is then turned into a training data set, which is then used to train AI models that offer a guess about the future. If we don’t understand how data are made, we can’t understand AI and its attempts to recreate the world. Let’s dig in.

Data are made by us

Data don’t fall from the sky. Data are made by you and me. We interact and transact with one another. We do things out in the world. We engage with institutions. We click and scroll online. In the collection of essays, “Raw data” is an oxymoron, Lisa Gitelman and Virginia Jackson make clear that data are never raw but situated in a historical and social context. The data generated from the various actions we engage in are constrained or nudged along by the social norms, belief systems, institutional practices, policies, and the cultural context of the day.

Let’s make this practical. In The Condemnation of Blackness, Khalil Gibran Muhammad argues that “From the beginning, the collection and dissemination of racial crime data was a eugenics project, reflecting the supremacist beliefs of those who created them.” As I wrote in Building Alternative Futures, “Historically, while social scientists conflated blackness with criminality, white criminality was explained away by structural inequities and poverty.” Gibran Muhammad sums it up this way:

“Crime statistics have never been just about behavior no matter how obvious it may seem that numbers speak for themselves. They are proxies for beliefs, a way of defining reality and seeing things. Whatever truth they represent in counting actual arrests or real prisoners is itself a reflection of intense social and political struggles.”

What gets recorded as ‘data’ isn’t self-evident or objective, it’s the output of a contestation of beliefs, norms, and power.

Old assumption: Data are raw and self-evident.

New assumption: Data are made via interactions with social systems.

Data are classified into value-laden boxes

Okay, so we take an action — big or small — in the world, and out pops a data point. Then that data gets recorded in a particular way or ‘classified.’ Take the US Census as an example. As I explained in Who are you? the boxes we can check aren’t an afterthought, they’re inherently political.

“In the 1990s, a group of Americans argued that ‘multiracial’ should replace the problematic ‘other’ category in the race and ethnicity section. This was seen as more appropriate by these advocates because, as Bowker and Star write, it would ‘not force individuals to choose between parts of themselves.’ […] But many civil rights leaders disagreed on the grounds that if everyone selected ‘multiracial’ that would mean lost information and lost resources for specific groups. Distinct categories are needed so that oppressed groups and communities receive the political and economic resources they deserve. The Clinton Administration decided that it would allow people to check more than one box but would disallow the inclusion of ‘multiracial.’ When we check boxes in surveys, these decisions may seem inconsequential, but the boxes available to us and the ones we pick are laced with political choices beyond our reckoning.”

Of course, there are other times where the classification occurs without political struggle, in a way that is invisible to us, and impossible to question. That’s right, I’m talking about Google, Facebook, and nearly every other modern technology company that uses algorithms and millions of data points to construct and then situate us in li’l boxes. These companies use our data to classify our gender, needs, and values. But they aren’t actually ‘ours’ in any meaningful sense — they own the data and don’t care how we self-identify. No matter what mode of data collection we’re talking about (a form, an algorithm, or something else) how individual data points get recorded reflects choices made by someone, somewhere.

Once a classification scheme is in place, the not-at-all ‘raw data’ can be labeled to train a machine. Now, in AI models, companies often skip this process altogether and rely on unsupervised training (i.e. trained on correlations/patterns in unlabeled data). Building a ‘supervised model’ (i.e. trained on labeled data) takes time and money. In the great new resource, “Models All the Way Down,” Christo Buschek and Jer Thorp estimate that it would take 781 years — working full time — to look at each image in the influential AI training set, LAION-5B. Rather than do the labeling up front — and to accommodate this ridiculous scale — companies will try to convince us that data are objective and that if we stitch together a big enough data set, it can represent a ground truth of reality. Other times, companies outsource this process: remember when OpenAI paid Kenyan laborers meager sums to label data that included violence, hate speech, and sexual abuse?

While investing time and money to label training data might minimize the problems above, it would generate others. The labeling process is infused with social and cultural biases. Even if the classifier ‘hate speech’ might be agreed upon by a team of developers in Silicon Valley, people across the world are going to have very different perspectives on whether data should be labeled as such.

Old assumption: Data are objective.

New assumption: Data reflect the values, biases, and decisions of people in power.

Datasets are contorted by scale, shortcuts, and mental models

Training datasets are the foundation upon which AI is built. They provide examples upon which systems generate supposed truths. But they don’t materialize out of thin air — they too are created, maintained, and managed by people working in organizations and companies with their own dynamics and incentive systems. In “The social construction of datasets: On the practices, processes, and challenges of dataset creation for machine learning,” Will Orr and Kate Crawford interview dataset creators who expressed how creating high-quality datasets is “a practice that is in flux, and requires considerable individual judgment, hard work and an ongoing struggle for resources and ever-increasing scale.”

Ah, yes, ever-increasing scale. As I’ve written, the promises of scale — that if we just had more data, we could achieve greater intelligence and predict the future, etc. — are pervasive. According to the paper “On the genealogy of machine learning datasets: A critical history of ImageNet,” this belief system and discourse posits “that the more accumulation of data, the better and more accurate the techno-scientific instruments will be.” But this ‘more data, more better’ ideology misunderstands the limitations inherent to scale. The creators interviewed by Orr and Crawford point out that the scale of a dataset separates the creator from the data. At some point, the dataset becomes too large to audit or understand, and the creators feel at “’ arm’s length’ from a sense of accountability.”

Because the datasets are too big to understand, creators take shortcuts. They might create filtering mechanisms to reduce their labor burden: to filter out pornographic images from one dataset, the creators put together a list called ‘Dirty, Naughty, Obscene or Otherwise Bad Words.’ But this required the creators to define these terms, which according to Orr and Crawford, led to the disproportionate removal of content “produced by and for minoritized groups, such as health content for LGBTQI+ communities and marginalized English dialects.” The article distills the many ways creators shape the creation of large datasets, leading the authors to conclude, “Creators’ motivations, practices, and constraints invariably impact the form and substance of datasets which then go on to shape the models used by the ML community.”

Old assumption: Big data is truth.

New assumption: Big data requires people to take shortcuts and make tradeoffs.

Training data is referred to as ‘raw’, or ‘objective’, or a complete stand-in for absolute truths. But data are anything but. The classification of individual data points alters and skews what those points are meant to represent; representations are then skewed further by how these data are used in large datasets. The purveyors of AI systems are veiling these processes with a narrative that tells us there’s just one truly accurate way to describe the world — and it is divorced from how we interact with each other, institutions, and social systems.

As always, thank you to Georgia Iacovou, my editor.

This is my second ever ‘reading list,’ a new type of ‘special issue’ for paid subscribers. I went back through every essay I wrote on crypto and curated and contextualized my favorite news articles, academic papers, and books. Each helps explain how crypto is entangled in social systems like gender, race, and power. I’ll add to this list over time but for now, sit back, relax, and use this list as your guide to saying real smart things about crypto as it creeps its back into the news, and public discourse.

🗞️ Articles

Read more

Read more

• 🖼️ The frameworks and concepts from past essays, which are organized nicely into themes.

• 📚Each theme connects to the original Untangled posts and provides further resources if you want to dive in deeper.

• ✔️With each theme, I will also recommend a daily action for you to try in real life, to make these ideas more practical.

There are now 13 themes in total - enjoy!

☝️ Theme one: technologies and social systems like gender, race, and power are entangled. I’m nothing if not on-brand!

In ‘Crypto is not decentralized,’ I argued that technical decentralization doesn’t do anything to decentralize power. I drew on the work of Angela Walch to illustrate that power concentrates in intermediaries, voting and governance, token ownership, coordinated decision-making, computing power, and in informal authority.

Though technically decentralized, certain nodes can still hold more voting power.

Then in ‘What’s the deal with DAOs?’ I drew upon ‘The Paradox of Meritocracy’ by Emilio Castilla and Stephen Benard to show that removing formal organizational structures doesn’t actually level the proverbial playing field — it only leads to informal structures and hierarchies based on race and gender.

Both of these examples showcase how technical systems cannot be easily separated from social ones. In fact, we separate the two to our detriment. Our assumption that technologies act on society only hides power and existing structural inequities. This obviously is very useful for some, but damaging to many.

Want more examples of this? In Trust me, I’m a smart contract I discuss how Web2 companies (like AirBnB) rely upon a reputation-based trust which is shaped by existing social relations and prejudices. Or, if you have an extra 30 minutes and want to dive into the ways gender, power, and technology shape one another, listen to Os Keyes on The Good Robot podcast.

Daily Action: Reject the premise that data and technology are ‘objective’ or ‘neutral.’ Ask instead, in what ways are data and technology entangled in social systems?

👬 Theme two: technologies have narratives that shape or hide the societal impact of those technologies.

In ‘Technology for … what exactly?’ I drew upon the work of Virginia Eubanks to show that narratives of fraud are used to justify the use of surveillance technologies on poor people even when the evidence isn’t there. In ‘Tokenizing Creators’ I argued that associating social tokens with the narrative frames of “ownership” and “empowerment” will only mask a much more likely impact: that tokens will turn creators into stocks.

In the creator stock market, $UNTGL is a real up-and-comer, $GAGA is down off the Oscar snub, $BRON and the Lakers are likely to miss the playoffs, while $BRENE (Brene Brown) is ... dominating, like usual.

We use narratives to make sense of ourselves and the world around us. The frames we select — ”fraud,” “empowerment,” etc. — make certain elements of a narrative salient, and hide or downplay others. Frames aren’t neutral — indeed, they’re often intentionally created and marketed. This is why they’re extremely powerful. Ya know how we say “climate change” rather than “global warming?” That was a strategic choice. In tacitly adopting a frame, we’re aligning ourselves with a set of interests, values, and politics, often without knowing it.

Still not satiated? Well, if you have the patience for traversing the wonkiest explanation of the power of frames, embedded in a dense text, check out Robert Entman’s paper, ‘Framing: Toward Clarification of a Fractured Paradigm.’ Or just head over to the Untangled Archives and see what narrative frames pop out. For me, it’s ‘connection’ and ‘decentralization’. What stands out for you?

Daily Action: As you read about new technologies, identify the narrative frame used to characterize the technology, and then interrogate it — how does the frame obfuscate power? What values are implicit in the frame? Whose interests might it launder?

🍨 Theme three: the legitimization of narrators who are experts in technology, and not the social systems that the technology is entangled with — and how this reaffirms or exacerbates the status quo.

In ‘Let’s imagine an alternative metaverse,’ I showed that all of the frames meant to describe the metaverse either accepted or entrenched the status quo. If we ever want to challenge the status quo, we need to assess the system from the perspective of those excluded from it or harmed by it as a result of their social locations. Here, I drew upon the work of Sandra Harding, Sasha Costanza-Chock, Andre Brock, and groups like the DISCO Network.

What the different framings of the metaverse really mean…

Then in ‘Venture capitalists are the new narrators of crypto’, I argued that VCs are unreliable narrators because they subscribe to a techno-determinist mindset — or one that sees technological change as determinative of social change. We need different narrators. I highlighted the philosophy of Afrofuturism which suggests that what needs reimagining are the social systems that shape and are shaped by technology.

Going further: Who we select to help us understand the future is ultimately about power in the present — it’s about whose experience, expertise, and epistemology gets to decide what problems matter and how we understand them. If you want to take the next step in reimagining sociotechnical systems, check out this talk, The New Jim Code: Reimagining the Default Settings of Technology & Society by Ruha Benjamin.

Daily Action: The next time you read an article about technology, ask whose perspective and expertise is being privileged, and whose is being sidelined? Who is quoted and who is not? Who has credentials next to their name and what are they? It’s possible a lot of stuff you read would benefit from someone else’s perspective — e.g. not someone whose expertise lies only in technology; or perhaps not a white guy like me.

Read more

• I analyzed a practical proposal designed to address platform governance at scale: citizen assemblies! and then created an audio version of the essay.

• I shared the first major update to the Primer. As a reminder, the Primer is a special issue that consists of the frameworks and concepts from past essays, organized nicely into themes. Within each theme, I offer further resources if you want to dive deeper, and I recommend a daily action for you to make these ideas more practical.

• In a critique of tech-criticism, I applied the concept of “criti-hype” to the public discourse on AI, and then offered a list of my favorite newsletters that transcend this trap.

This month I have something different for you — it’s a new, special issue, called “Technically Social,” which will live under the newly created ‘Special Issues’ tab. Let me know what you think. If you find it valuable, consider subscribing to the paid edition of Untangled.

Upgrade to paid

Now, on to the show.

Every time I write the word ‘technology’ there’s a li’l voice in my head that laments the oversimplification: “what are you even talking about when you say ‘technology’?” Welcome to my brain 🙃

We take what we mean by ‘technology’ for granted. In this special issue of Untangled, I want to ask the uncomfortable question: uh, what even is ‘technology’? What follows isn’t close to comprehensive but it’s a good start for anyone who wants to understand what technology is and the mechanisms by which it is entangled in social systems.

This special issue is somewhat similar to the Primer I wrote in December 2022, in that this is structured as a list of themes or ideas that I will continually add to over time. That primer focuses on how social systems (e.g. race, gender, power, culture, narratives, etc.) shape technology, whereas in this issue — Technically Social — I will focus on how technology shapes social systems: the materiality of the technology, how it is designed and developed, what it affords, whether it is a tool or arranged in a system, how its effects are ecological.

Subscribe now

Like the Primer, I’ll continue to add to this list over time. Anyone will be able to access it for the month of May, after which this document — and subsequent iterations — will only be for paid subscribers. Let’s dig in.

1. Technology evolves in combination with other technologies

One answer to the question, ‘uh, what even is technology’ comes from Brian Arthur in his book, ‘The Nature of Technology: What It Is and How It Evolves.’ Basically, Arthur argues that technologies are created out of existing technologies. This might sound a li’l recursive, but as Arthur explains:

“Any solution to a human need — any novel means to a purpose — can only be made manifest in the physical world using methods and components that already exist in the world. Novel technologies are therefore brought into being — made possible — from some set of existing ones. The jet engine could not have existed without compressors and gas turbines, and without machine tools to manufacture these with the precision required.”

My essay, Are blockchains immutable? gets closest to this idea. Blockchain technology wasn’t created out of thin air — it evolved in combination with myriad other technologies, like cryptography, hashing algorithms, distributed ledgers etc. It’s in part the intrinsic material qualities of the technology — alongside affordances and values (more on that in a minute) — that shape this combination. But we often abstract away from the materiality of the technology, getting lost instead in the narratives and hype around what technology might be able to do one day. So interrogating the material qualities of the technology — what are they, where do they come from, what resources do they require, what other technologies do they depend upon — can actually make us less susceptible to unhinged narratives and “criti-hype.”

Anyway, Arthur’s observations provide some explanation on how technologies evolve out of other ones — which gives us a sense of ‘supply’. But what about demand? Well, ‘demand’ isn’t quite the right term, since some technologies (e.g. penicillin) come into being not because they met a specific demand but because they represented “niches they could usefully occupy.” To capture both ideas, Arthur uses the term “opportunity niches” which reflects human needs and the needs of technologies themselves; it’s not just that technologies are created to meet a specific need, but that those technologies call forth additional technologies. As Arthur writes, “every technology requires supporting technologies: to manufacture it, organize for its production and distribution, maintain it, and enhance its performance. And these in turn require their own supporting technologies.” And there it is — technologies evolving out of other ones.

Moreover, some technologies are extra special; they become building blocks upon which other technologies are created. Neural networks, for example, are used in 3D printing, robotics, cloud services, battery technology, programable biology and therapies. This starts to get at what Arthur calls “combinatorial evolution” or the idea that technologies evolve by combining with other technologies.

2. Technology shapes societal values. But technology itself was shaped by the values and beliefs of its designers. Um, what’s going on?

How the technologies combine and evolve is informed by the values and beliefs of their creators and how those interact with the values in a given society. Stepping back, one theme of Untangled is that the beliefs and values of engineers and designers become encoded in the technologies themselves. Right, technologies aren’t neutral — they don’t originate out of thin air. They are designed by someone, somewhere, whose beliefs and normative commitments subtly shape its development. For example, scholars like Ruha Benjamin and Safiya Noble have shown over and again how engineers and designers encode racism in technical systems.

You might be wondering, “uh, but this special issue is about how technologies shape social systems, not the other way around.” True! But the value-laden technologies are then let loose on the world, and proceed to shape social values, impact people, and call forth aligned technologies:

Technologies create self-propagating values: Since those engineering and designing the technical systems tend to reflect dominant societal views, we often create a self-propagating feedback loop wherein the values or beliefs encoded in the technology are then reaffirmed by values and beliefs out in the world. And around and around we go! This isn’t always the case — conflicting societal values can dampen the effect of those values encoded in the technology, and vice versa, but that depends ultimately on seeing the technology as anything but neutral, evolving our values, and ensuring more people from minoritized social positions are doing the engineering.

The impacts of technologies are the most important part: While the intent behind these values matters, impact matters more. Even if it’s clear that a designer wasn’t malicious, or even just making decisions subconsciously rooted in discrimination, the discriminatory impact of their technology was still very real, and should not be minimized. Even if Robert Moses wasn’t a total racist, the overpasses he built in New York still prevented racial minorities and low-income groups from visiting Jones Beach State Park.

While I think Arthur is correct in saying that technologies evolve in combination with other technologies, I think we can go further. Technologies tend to evolve in combination with technologies that are values-aligned. For example, it’s not just that permissionless blockchains evolved in combination with cryptography. It is that blockchain technology encoded a cyber libertarian ethos that wanted to resist censorship, remove centralized authority, and put the individual front and center — and the values and affordances of cryptography fit with it like a puzzle piece.

3. Technology shapes and constrains human agency

Speaking of affordances, technologies have them in spades. Technologies shape and constrain people — the way our bodies are read and represented, where we focus (or don’t!) our attention, the actions we can (or can’t!) take, the decisions we make, etc. You can understand the comparison I offered between Google Search and ChatGPT-based search as a comparison of affordances — they nurture different kinds of behaviors:

“How we understand search is intertwined — some might say ‘entangled’ — with the technologies of the day and the social behaviors they encourage. One big issue with chatbots like ChatGPT is that it encourages the acceptance of a direct answer. Whereas Google might offer a mix of high-quality and clickbaity information sources from which to choose, ChatGPT generates a single plausible response with an air of authority.”

If you start thinking about how different technologies alter your behavior, you’ll start to see affordances everywhere, hiding in plain sight. has actually outlined a whole bunch of them in this post from (one of my favorite newsletters!). Each question can be read as revealing a kind of technological affordance.

Share

But basically, we can collapse most affordances into two categories: technologies that extend our capabilities and those that restrict them. Here is Ivan Illich, Catholic priest, philosopher, and social critic, making a variation on this point in 1973:

“There are two ranges in the growth of tools: the range within which machines are used to extend human capability and the range in which they are used to contract, eliminate, or replace human functions. In the first, man as an individual can exercise authority on his own behalf and therefore assume responsibility. In the second, the machine takes over—first reducing the range of choice and motivation in both the operator and the client, and second imposing its own logic and demand on both.”

Affordances matter, but to say that ‘technology shapes and constrains’ is too deterministic — it makes it sound like the effects of the technology are baked into the technology itself. As I wrote in Are blockchains immutable? the internet of today is nothing at all like the internet of the 1990s, which emphasized affordances like anonymity and ephemerality. What changed wasn’t a technological change but an accomplishment of companies.

Nevertheless, it’s important to understand the subtle ways in which technologies nudge and nurture certain behaviors — especially because these days, it feels a li’l less subtle and nudge-y, and more controlling. That leads us to the distinction between technologies as tools and technologies as a system.

4. Technology is a tool or a system

The extent to which technologies shape and constrain our agency depends on whether they are a tool or a system. Lewis Mumford, a historian and philosopher of technology, divides technologies into “authoritarian technics” and “democratic technics.” Here’s a brief breakdown:

• Democratic technologies are “man-centered, relatively weak, but resourceful and durable.” These are tools; they are used by people in service of their objectives. The objectives might still be real bad — a hammer can be used as a weapon — so it’s the user’s objective that matters.

• Authoritarian technologies are “system-centered, immensely powerful, but inherently unstable.” These are technologies that become a system — which then acts on people, rather than the other way around: power shifts from the people, to the system.

Take the Moses’ overpass again. The story isn’t simply that Moses embedded his politics into the overpasses, but that together, the overpasses became a system that regulated behavior. As I referenced in Technologies encode politics Langdon Winner referred to this as “specific ways of organizing power and authority.”

Over time, many ‘tools’ have become ‘systems’. Why’s that? Well, code is its own kind of regulatory system, or as Lawrence Lessig put it, “code is law.” Lessig expands on this foundational point in his book, Code and Other Laws of Cyberspace, writing,

“In real space, we must recognize how laws regulate — through constitutions, statutes, and other legal codes. In cyberspace, we must understand how a different ‘code’ regulates — how the software and hardware (i.e. the ‘code’ of cyberspace) that make cyberspace what it is, this code is cyberspace’s ‘law’.”

Code regulates our behavior, constraining the decisions we make, shaping the decisions made about us, and who gets to make them. This is clear when we examine the tools that have historically been used to classify people. Before the internet, we had surveys. Now, the stuff you click on determines ‘who you are’ for an abstract group of entities. But both evolved the broader system or architecture of classification. In Who are you? I wrote:

“Google uses our data to classify our gender, our needs, and our values. But they aren’t actually ‘ours’ in any meaningful sense — Google owns the data and they don’t care how we self-identify. These classifications are used to computationally calculate a profile with a bunch of different categories that they sell to advertisers.”

So essentially, the classification system that Google perpetuates makes self-identification irrelevant, and instead it groups people into marketing categories without their knowledge. Furthermore, both surveys and predictive models are systems that work on codifying judgments made in the past: a survey check-box or audience segmentation tool does not ‘imagine’ how people might want to be classified; it simply creates categories based on past and present information.

Users are especially beholden to algorithmic systems of classification: at least with a state-run survey, there is an opportunity to question why certain check-box classifications were included or excluded. But there is no deliberative process to decide how people should/shouldn’t be classified online. Our algorithmically determined advertising profiles are tweaked and updated on a daily basis, and facilitated by systems that are designed by engineers who are unaccountable to the public.

We hardly ever reflect on the ways in which we are classified. That’s because technological systems fade into the background, often becoming invisible to us — but they’re never not acting on us. Indeed, technology is ecological, rearranging what came before it.

5. Technology is ecological

Technology alters ecosystems. I’m not the first to make this point. Neil Postman argued long ago that “Technological change is not additive; it is ecological,” explaining that “A new medium does not add something; it changes everything.” As technologies combine and recombine with one another — per Arthur’s point above — and an opportunity niche is met, societies evolve too. Systems don’t remain the same, but just with ‘more technology’; they become altogether different ecosystems. Postman looks to history and writes,

“In the year, 1500, after the printing press was invented, you did not have old Europe plus the printing press. You had a different Europe. After television, America was not America plus television. Television gave a new coloration to every political campaign, to every home, to every school, to every church, to every industry, and so on.”

A more recent, Untangled-approved example is how the use of image and video-based algorithmic recommendation systems — like those underpinning TikTok — isn’t just the U.S. plus those technologies. As I wrote in “TikTok & the production of ignorance,” the technology itself compounds the existing sense of societal speculation and uncertainty, and therefore, “encourages a kind of volatile immersion. It feels a li’l chaotic but you’re nevertheless engrossed and stimulated, left to wonder hours later how a string of bizarre videos kept your attention for so long.” It’s a disorienting experience that Aris Komporozos-Athanasiou likens this experience to “the legibility and logic of a narcoleptic dream.”

But this experience, indeed, our relationship to the technology is one-way. We don’t know how the algorithmic system is determining what to show us next. As I wrote then, “It’s this gap — between what the technology helps us see and all that remains hidden — that creates an underlying sense of unease or uncertainty.” U.S. society isn’t ‘the same as it was plus ‘a new way to consume short-form videos’ — its psycho-social ecology is different.

So too is TikTok altering the kind of relationships we have with one another — disrupting the social graph, ushered in by Facebook and earlier social media sites, and encouraging more one-sided relationships. What this new ecology looks like isn’t clear, nor is it fixed — it’s ever-evolving in combination with other technologies that nudge social dynamics and rearrange culture.

That’s it for this special issue of Untangled. Please let me know if you enjoyed this format. If you have an idea for a different format, click ‘reply,’ and send me an email. I’d love to hear from you. As always, thank you to Georgia Iacovou, my editor.