Prototype Submission » Sparkle Bureaucracy

Field Note - Trust & Safety at Mastodon

It's so interesting that Mastodon, as an organization, is choosing to hire a Community Director and specifically for Trust & Safety.

My top priority is to ensure that Mastodon’s plans for these areas reflect the needs of Mastodon’s many different communities.

Field Note: NYtimes Opinion - video - Goodbye, Price Tags. Hello, Dynamic Pricing.

I loved the way the story was crafted in this piece - about how dynamic pricing is going to target and stalk us in order to squeeze as much money as we're able to pay based on needs the algorithm anticipates through other smart devices or connected services.

I came across this video piece on tiktok and it led me down a rabbit hole of "surveillance pricing" where retailers and platforms will use algorithms and personal data (device type, location, browsing history, past purchases, things you've said aloud to your smart home, etc) to set _individualized_ prices - so two people might see completely different prices at the exact same time for the exact same item!

In many ways, this undermines transparency and fairness since consumers (a) may not be away that they're being charged more or (b) may not be aware why they're being charged more! Most pieces talk about how regulators need to do more - but there might be an opportunity here on signaling that this is likely already at play in the market and will cause real harms.

nytimes.com ↗

Field Note: What makes a community a community? Is it shared membership, or shared responsibility?

I’ve been thinking a lot about the engagement of the cohort and why I feel upset when people disengage. Perhaps it’s a personal thing, such as a feeling of abandoned.

When I really consider that, however, it’s more complex. The fellowship, to me, was a sort of group project to build and govern a small community that is, as Ed often says, like a small country. I believe communities require stewardship, care, explicit communications, discomfort (because we can build anything if we are all comfortable), and collective responsibility.

However, the format by which our cohort is running (in some ways by “default”) - with fluid responsibility and not accountability, to me feels like… the anthithesis to how I implicitly understand communities to be built.

Betrayal is not the right word here, but I certainly feel that when no one is “required” to steward and no one “owes” each other anything, then we can’t build a community we actually want to be a part of, that requires intentional work.

It’s a complex feeling. I empathize with the reasons for disengaging while also feel hurt by the “dropping of the group project”.

Is this grief that the shared project is being abandoned? Grief for the version of the cohort I imagined you we could potentially build together where we would get past the everyday polices that we know to be unhelpful (like codes of conduct) and reimagine what those systems could look like with a new lens to take back to the real world?

I’m trying to decouple my instinct to hold the community together from my excitement about the fellowship itself, but I suspect people like me will always yearn for communities where everyone is bought in. Maybe learning to take small steps… and confronting how people actually engage is part of this political education and I just don’t see it because I’m in it.

And there may still be ways for the core group to model community with more structure or experimentation.

How do I keep modeling the kind of community I believe in, while also protecting myself from “hurt” and accepting that people will engage at their own limits and that this, too, is part of… politics?

_originally posted on Dec 8, 2025_

Field Note: Three Person Teams, Power, and the Shape of Leadership

In our Leading Tech Projects class, Peyman ran an exercise asking what it actually takes to lead a new technology project. We mapped out all the roles and capabilities we could think of, then clustered and prioritized them into tiers. After that, we were given a constraint: _you only get three people on the team. What do those roles look like?_

People proposed many roles, including two developers and one product manager. I strongly disagreed.

My discomfort with that structure isn’t about undervaluing engineering - it’s about power. When you stack a team with developers and remove explicit roles for user research or design, decisions about UX, accessibility, and user experience don’t disappear. They get absorbed by whoever is closest to implementation. In practice, that often means developers making product and design decisions by default, not because they’re best positioned to do so, but because the structure leaves no alternative.

My counterproposal was a team composed of:

- one developer - one product manager - one user researcher

This structure distributes power more deliberately. The developer can focus on technical implementation and system design. The product manager can take responsibility for delivery, coordination, supplier management, disaster recovery, and organizational tooling. And the user researcher can protect the integrity of user needs, grounding decisions in research, accessibility testing, and data rather than intuition or convenience.

This proposal isn’t theoretical. It’s influenced by civic tech fellowship models like Code for America, where small, cross-functional teams are designed to cover not just skills, but overlapping accountabilities. In those settings, the user researcher often acts as a counterweight to purely technical optimization, ensuring that “what can be built” doesn’t eclipse “what should be built.”

What I’m really arguing here is that minimal teams make power dynamics more visible, not less. The question isn’t whether trade-offs will happen, it’s whether we design teams that acknowledge them explicitly.

Field Note: Permission, Invisible Norms, and Acting Inside a "Free" System

After sharing some of my feelings about disengagement and governance with people in the cohort, I found myself in an unexpected conversation about permission. (Well, Ed was responsible for bringing my hopes and dreams and discomfort with informal roles and expectations - but that's a side note)

Newspeak House is designed as a space without fixed rules. Every year, the cohort is meant to define its own norms. We’re all peers. In theory, no one needs permission to act. In fact, many cohort members feel empowered by this - that they don't need anyone's permission, that they'll just do what they want. I really admire that!

And yet, despite understanding the idea logically and wanting to be that person, I notice that I do need permission.

What was difficult to convey is that don’t experience this as a lack of confidence. Although that is feedback that I receive constantly in corporate contexts. I know I can lead. I know I’m capable.

But when it comes to starting programs, proposing policies, or changing how the house operates, I feel a strong pull toward structure: a document, a principle, a process I can point to. Something that legitimizes action.

Part of this may come from my vocation. As an engineer, I’m comfortable with rules, systems, and explicit structure. Ambiguity can feel blocking rather than freeing.

But I don’t think that’s the full story.

I suspect this instinct is more deeply political. As someone who has experienced marginalization, I’ve learned through real consequences that acting without permission in the world often carries risk. In many systems, people like me are penalized more harshly for stepping outside informal or unspoken boundaries. Over time, that teaches you to seek cover: consensus, legitimacy, shared responsibility.

What’s interesting is that this instinct disappears when harm or safety is at stake. In moments involving psychological safety or code-of-conduct violations, I don’t feel the need for permission. I act, because the moral imperative is clear. And I act because I have the experience to know to act in the best interests of the victim - I’ve done this before.

But in cases of governance or experimentation, I hesitate, not because I don’t care - in fact I care a LOT, but because I’m still carrying the outside world into a space that claims to be free of it.

This makes me wonder whether I’m still operating as if systemic oppression is present, even in a place intentionally designed to suspend it. And if that’s true, what does it mean for prefigurative politics? If we’re trying to model new forms of governance inside the house, but we arrive with deeply internalized norms shaped by unequal systems, how do we unlearn those things? Is the design of the course meant to challenge the invisible norms that we bring with us?

Now I’m wondering how invisible histories or invisible norms shape behavior, even in spaces that imagine themselves as neutral. And before anyone says it - no space is ever truly neutral. Even in a space like Newspeak House, with the explict standard of being without rules and everyone can set rules - there are explicit structures of privelege (e.g the people who feel they can act w/o any structural legitimacy prove that).Is learning to act differently is part of the political education I came here for? Because if it is, and if I walk away from this year knowing when to ignore the structures that limit or oppress me - that would certainly be both political and wonderful.

_originally posted on Dec 12, 2025_

Field Note: Shadow Governance

I’ve expressed this in the retro document, but in some of the readings from the course material and in conversations with Ed, I’ve been thinking more about this. In groups like ours that reject formal authority and prefer informal structures, governance doesn’t just disappear, it sort of gets distributed to those who hold the most labor.

Someone is still doing all the work - housekeeping, offering advice, legitimacy, managing ration club or the guest room, or lobbying. When that labor is informal, it often goes unacknowledged unless you see it happen (like someone putting away the dishes every day or running a workshop after you’ve been talking about a thing for ages), and it becomes “Shadow Governance”.

What I’m noticing is that I occupy this role that doesn’t exist. I took it on naturally and I take full accountability for that - I articulate the different stakes and tensions that I hear from people, I lobby the faculty for procedural change and share the implicit conflicts before they happen, I try to warn the institution about frustration when it still feels early.

I believe the group benefits from this work because it legitimizes the documentation that we as a group provide via other channels. This arguably the most political thing we’ve done - collected, synthesized and presented our feedback and demands. I did my best to try and socialize those demands with faculty that I trust - only today was I told that this is called “lobbying”.

The group benefits from this work, but does not recognize it as an official role. At the time when the role was proposed, recognizing it led to several uncomfortable questions about who would have influence, who would have power over information, and why was that needed?

I think this is also why… I take on a larger emotional cost when things go wrong. Not because the failure was in any way my fault, but because I implicitly try to hold the group together. This was painfully visible to me in the governance retro. When two people exited the process, I felt this collapse of my worst case scenario. I had a feeling this retrospective wouldn’t work - I tried to voice those concerns early. I had this instinct that the group needed cohesion and that would be tricky to resolve.

As I’ve been reading some of the papers assigned to us in Six’s class, I’m learning that political theory focuses on formal authority. Our informal systems in the house depend on so much “integration labor” - work that keeps people from Exiting - because they feel they can't use their voice. I guess I felt responsible for (even though I’m actively trying not to do this) the coherence of the group and in turn, their experience with Ed, and in that moment, I could not find a way to repair it.

Someone asked me today if I wanted to be spokesperson for recognition - and I felt a bit strange about that. Maybe that’s what it looks like from the outside to someone who doesn’t understand the concept of spokespersons. I already have that recognition broadly in the industry I work in. To assume that was my primary motivation feels wrong, in a sense.

Being a spokesperson is like living on two islands and you being the bridge. Your role is to transfer as much information across the bridge in both directions as is needed. Too much information and you’ll cause chaos on either side. Too little information and there’s a loss of trust. And sometimes you make decisions to move information across in the best interest of either side. It’s a very tricky position to be in and I have a lot of experience doing it in the corporate sector.

But a shadow spokesperson? That’s just framed as a personality aka “everyone likes her” or “she’s good at explaining” or “she’s just emotionally invested”- rather than understood as labor around governance. In reality, the work of translating so many inputs and then re-filtering them for the context between sub groups, is a process that preserves trust.

_What happens when you try to legitimize a role inside a system that may not believe legitimacy itself exists?_

---

Field Note: On the legitimacy of the governance module

This has been sitting with me since before the governance retrospective, but even more so since then that clarifies why I keep feeling “stuck”.

When I talk about experiments in the governance module, especially the idea of temporarily assigning authority or playing with rules, I often hear some version of:

“Even if you were dictator and there was a signed charter, why would we follow the rules?”

It’s usually said in good faith, sometimes even in good faith debate. When I push on this, my allies in the house say, “well that’s life” or “well that’s politics”

Is it? I’ve been trying to understand why this answer feels unsatisfying to me. This isn’t the politics that I came to practice at the house, this is the politics of convincing people that the legitimacy and value of a thing that I thought they already had bought into.

I came into this fellowship believing that opting in meant opting into the governance module, that we were agreeing, as a collective, by the rules we would design, test, contest, and revise together. I thought that engagement, at the very least was required, not agreement. But other cohort members seem to have opted into a much narrower version of the fellowship: shared house, shared conversations, and this optional play in the sandbox that can’t ever _really_ constrain them.

Those are very different constitutional assumptions.

What I’m noticing is that we keep trying to debate _policies_ without ever settling what would make a rule binding in the first place. Every time a proposal is raised, the conversation snaps back to legitimacy: _why should this count? why should I care? why should I comply?_ That snap-back blocks the kind of political disagreement I actually want.

I want people to fight me on implementation details, on scope, on enforcement, on harm. I want to argue about whether a digital ID at the door produces exclusion, or whether differentiated access rules create power in ways we don’t want. But instead of those debates, I keep encountering a deeper refusal: _the suggestion that no rule here could ever be binding, regardless of how it’s designed._

Reading excerpts from _Max Weber, Economy and Society (authority & legitimacy)_ and _Elinor Ostrom, Governing the Commons_ has helped me name what feels off. Authority doesn’t persist because it’s declared - it persists because people _treat_ it as legitimate. Rules can only stay alive if people who follow them can shape them, when there are shared expectations, and when there are consequences to refusal. I’ve been told that the institution would support compliance to signed charters via forms of exclusion - but I don’t believe in that form of compliance. I would rather have people opt in somehow.

That realization reframes a lot of the frustration I’ve been feeling. I’m trying to run experiments in a system that hasn’t decided whether it wants to be governable at all. The governance _experiments_ aren’t being rejected, they’re just not participating.

That’s a valid choice, but it changes what this place is and it changes the way that the 3-4 of us who believe in the experiments will experience the fellowship as we understood it.

Without that shared commitment to the legitimacy of the governance module, governance becomes this optional performance, this sandbox game that a few are playing in the house that has no real power or impact or authority. That lack of interest in the opportunity and space and resources we have been given to play around with, it actively demoralizes me.

For those of us who came here believing this was a rare, unique space to test rules, norms, and systems with real stakes, that decision quietly redraws the boundaries of the fellowship itself - and I don’t know if I want to work within those limits.

Field Note: Multi-Agent Systems, Architecture, and the Limits of Context

After discussing team structure in Peyman's class on "Do we need developers?", we moved into a second question: which parts of this work can be delegated to AI?

The framing that stuck with me was asking two questions: - What do I care about? - What could I work on with AI?

Right now, the industry seems to be moving toward specialization through tools and agents. We chain different systems together, Figma for design and research, code-generation tools for backend or frontend work, analytics tools, QA tools sometimes all inside one platform (GitLab ;) ), sometimes stitched together across many. Functionally, this already feels like agent-to-agent or tool-to-tool collaboration, even if we don’t always name it that way.

I know that we’re heading toward multi-agent specialization - not just many tools, but many agents with distinct roles, responsibilities, and domains.

This raises a set of questions I've been thinking about: - If you have strong documentation, clean architecture, and well-written code, is that enough for agents to align with an existing system, especially one that’s 10+ years old? - Or do agent-friendly systems need to be designed differently from the start? - What does an agent-based architecture actually look like? Is it microservices? Component-based design? Something else? - How do we design for agent handoff rather than pockets of execution chained together?

Right now, it feels like handoff is the missing piece. We have agents that can do impressive things on their islands of code, but less clarity on how responsibility, context, and intent move between them.

There’s also a scaling problem here. As systems grow, context windows become a hard constraint. At some point, no single agent can “understand” the whole system. Does that mean agents are ultimately responsible only for their own islands of code, communicating through structured interfaces or protocols like MCP? And if so, how much coordination overhead does that introduce?

I'm curious about these tensions as a result of Peyman's class, thinking about the in between of legacy systems and new architectures, between specialization and coherence, and between what we want agents to do and what context actually allows.

Field Note: Multi-Agent Handoff and Context

While presenting about agentic workflows with engineers at a large financial institution, the term that kept coming up in questions wasn’t _reasoning_ or _autonomy_. It was handoff.

That caught my attention.

By handoff, I don’t mean chaining prompts or passing outputs. I mean the transfer of state, intent, assumptions, and responsibility from one agent to another. Not just “here’s the result”, but “here’s why this work exists, what constraints matter, and what you now own”.

Most multi-agent systems don’t really do that. They pass artifacts such as text, summaries, tool outputs and expect the next agent to infer the rest. That inference step is where things become fragile. At first this feels like a gap, but looking closer, it seems more like deliberate design.

This is where context engineering keeps coming up for me, not as prompt tuning but as a systems constraint. As systems grow, no single agent can hold the full picture. Instead of trying to move full context, intent, and responsibility between agents, most systems centralize those things elsewhere.

Anthropic’s multi-agent research system is a useful example here. Their setup relies on an orchestrator agent to plan the work, decompose queries, spin up parallel researchers, and integrate the results. Intent, assumptions, and responsibility are held centrally and re-expressed at each step when they're passed to the next agent. Tools like LangGraph or Agent Squad also do so by centralizing intent and control rather than enabling agents to hand work to each other directly.

So I guess the hard part of multi-agent systems isn’t getting agents to do work. Execution is the easy bit. The hard part is everything that happens _between_ agents. Rather than solving agent-to-agent handoff directly, today’s systems avoid it by routing all coordination through orchestrators as a central point.

There’s a similar pattern with evaluation. Letting agents critique or evaluate other agents can scale quality control, and LLM-as-judge approaches will work well when they’re designed well. Evaluation only works because intent and constraints are held outside the agents themselves. Without that central context, evaluation risks optimizing for surface correctness rather than whether the system is actually doing the right thing.

Protocols like MCP help standardize how agents access tools and shared state, but they don’t really solve handoff either. They make context accessible, not transferable. Intent, assumptions, and responsibility still live in orchestrators, glue code, or prompt conventions.

I’m left with a few open questions:

- What is the smallest amount of context that is actually needed to hand off from agent to agent? - Is centralized orchestration a temporary solution, are there alternatives? - Are there existing systems (or papers) that attempt agent-to-agent handoff of intent or ownership, and where do they break down in practice? - This makes me wonder whether some of the challenge behaviors in agentic systems may be happening in the orchestration layer.

anthropic.com ↗

Field Note: When Optimism Feels Like a Social Risk

This has been sitting with me for weeks. After conversations with Ed and faculty, it’s gotten sharper, but I’m still not sure what this experience really is, or whether it has political viability.

There’s a thing that happens in groups that I care about where critique becomes the dominant point of conversation. Not critique as analysis or disagreement, but critique as posture, where the “system” we exist in whether that’s the cohort or an open source team, is fundamentally broken, the powers that be are “irredeemable” and nothing short of total rejection is acceptable.

And I find that in groups like these, over time, there’s a hostility to hope.

As someone who tries to cradle a tiny flame of optimism, I find it challenged as naivety. Attempts to contextualize, or to reason about what might still be possible, the opportunity space ahead of us, are met with suspicion or ridicule as a gut reaction.

I’ve also encountered this in open source communities I’ve volunteered in. Once an “enemy” of sorts is established, it becomes a very tricky operation to propose any kind of compromise without looking complicit. But without compromise, nothing actually resolves. Volunteers get no closure. The community can’t move forward in a way that preserves the sustainability of either the people or the project.

This dynamic seems to find me in every group, team, or company I’m part of (or perhaps it is a pattern that I find myself drawn to as a result of caring for the building of communities?) And it crushes my hope. Not because I disagree with the critique, I often share it, but because there’s no room left for action. There’s no space to be curious, to imagine moving forward, to repair, iterate, or make partial progress. The only acceptable stance seems to be collective fatalism.

Most recently, I feel this in my sense of purpose in the fellowship. I’m here because I believe deeply in the vision, in building, testing, and intervening in systems and experiments that I know have failed me in the real world, and through that, learning politics by doing. But when optimism becomes socially risky, I start to feel like I don’t fit, like I need to shrink my hopes and ambitions to meet the moment.

If I’m lucky, there are people who replenish my optimism, because they, too, feel unheard and feel the weight of the current political situation pressing on them. The Americans in the cohort, for example, share a kind of implicit understanding of purpose. They understand this political moment, sense that something is shifting, and feel called, like I do, to do more. Those conversations restore my energy and clarity. They remind me why I care and why I moved my entire life to be here. But when I feel replenished and return to conversations at the house, the group norm reasserts itself. I find myself grasping at the threads of hope I had so carefully braided earlier that week, or even earlier that day.

What makes this especially difficult is that I care deeply about the individuals in this group. Their opinions and feelings matter to me. I want to stay engaged, and leaving feels like abandonment or like minimizing their pain. This in-between space is really a rocky road and I find myself feeling drained more often than I am replenished. I also don't want to withdraw entirely.

In the past, I’ve only been “successful” in navigating this when I was given a clear charter of power. In open source, that meant asking maintainers to step away when their positions were no longer serving them or the project, especially when they were silencing newer, curious voices who wanted to work rather than only critique. In corporate environments, it meant reassigning responsibilities and holding people accountable for the work they were meant to do. In situations like this where I didn't have agency to change anything, I usually burned out trying to recapture my hope week after week and eventually left the organization.

Maybe that’s the only way I know how to handle this. Or maybe there are other routes, working at the individual level, explicitly building a group of people who still have hope, and naming what’s happening out loud so we can refill each other’s energy intentionally.

Being able to name this pattern feels important. If I can see it clearly, maybe I can build a skill for living with it or intervening without losing my hope next time.

Open questions:

• How do groups reset norms once they’ve drifted this way?

• What are the early signs that you're headed this way?

• At what point does naming harm stop creating capacity and start eroding it?

Field Note: Before we come for Managers, we need to look at Organizations

I suppose the purpose of this class session was "Do we still Managers?" and will AI replace them? I fully expected Peyman to present us with question early on. Instead, we dove into the meta a bit. Before we can really analyze the role of managers in organizations, need to talk about organizations themselves - and how they evolve (and sometimes don't).

organizational paradigms

In “Reinventing Organizations”, author Frederic Laloux writes that “_[a]n organization cannot evolve beyond its leadership’s stage of [consciousness]._” He describes specific stages of organizational paradigms:

- Red - organizations ruled with a absolute power, limited to the immediate reach and influence of their leader. Red organizations are often described as wolf packs or the mafia. Everyone is motivated by fear, change is management can happen through a change in power dynamics. - Amber - organizations with a top down conformist structure of power with rigid stages of management and hierarchies to pass down stable, replicable processes to large groups of people. Consider the Catholic Church or an army. - Orange - typical organizations that make decisions based on evidence, acting like machines - setting goals, breaking those into benchmarks or OKRS, and then working towards them. These goals can be driven by shareholder value for example. Most companies exist at this orange stage. - Green - a newer model for organizations that is values driven, with a culture of empowerment, multi stakeholders that share in decisions and benefits of those decisions. Individuals in these organizations are able to drive things forward. - Teal - is an evolutionary model for organizations - where everyone is a manager, and everyone has the autonomy and authority to fix issues themselves. Distributed, remote teams are in a key position to take advantage of this new model, because the architecture and real estate of a physical workspace can’t play into reinforcing a hierarchy.

What struck me about this analysis of organizational paradigms was slotting in the organizations that I've worked for and reflecting on how they've moved along the scale - which is certainly not linear. Because in every organization, there are pockets of teal - where employees have autonomy and authority to act, vs pockets of red - where a VP might have absolute power, say, and control. I also considered the fact that organizations don't always evolve forward, they often regress.

Under the pressure of the market, scarce resources, or industry's rapid change, even mature green organizations can slide back into tightly controlled, slower decision processes, and higher amounts of centralization. For example, towards an IPO, even the employees with the most autonomy might find themselves following beauocratic hierarchical processes due to compliance needs.

A question that was posed in class was - _in a genuinely good culture or system, can you leave it up to people to decide what they should do?_ this is an interesting one to think about in reference to AI agents and the future of work.

organizational reshuffles

Organisational reshuffles is also an interesting lens here. When structures change, people often _rethink the organisation itself,_ its values, its rules, what is actually allowed. That moment of ambiguity can feel destabilising, but it might also be one of the few times real reflection becomes possible. Space to rethink isn’t always a failure mode; sometimes it’s the point.

activity - management activities

We ran an activity in class: list “management activities” and then sort them into categories:

things that don’t need management at all, things that are necessary and centralized, things that may be necessary depending on context, and things that aren’t necessary.

The exercise quietly dismantled the idea that “management” is a single role. Instead, it’s a bundle of functions, some structural, some cultural, some political. And some bundles of tasks, we found over discussion, weren't even necessary for the manager to do.

Doesn’t have to be management Organising product Project management (doesn’t need to be centralised) Time allocation Initiative / experimentation

Necessary and centralised Consensus mechanisms Policy enforcement * Core strategy / value-setting (emerging core strategy / value team)

Can be necessary (context-dependent) Product management (depends on scale and cross-functionality) Project management (varies by context) Governance under scarcity Resource prioritisation during limited resources or reshuffles

Not necessary Vision setting “Corporate BS”

This is where the fantasy of fully distributed, future-leaning teams started to fray. Many such models assume abundance: enough time, enough money, enough attention. In reality, teams operate under limited resources. You can decentralise how people allocate their time, but the _consensus mechanisms_ - what ultimately gets prioritised, funded, or enforced - often remain centralised. Scarcity pulls authority back toward the centre.

Product and project management surfaced as examples. In theory, they could be decentralized. In practice, they rarely are, especially when dependencies stack up. Policy enforcement, even more so, tends to centralize by necessity. This isn’t a moral failure; it’s an organizational gravity well.

culture vs process

That tension showed up again in the discussion of culture versus process. Is it more important to have a handbook, or an open form of discussion? A static policy, or a living document? The more interesting framing wasn’t either/or, but _what the handbook actually does_. Handbooks are often ignored and yet they generate discussion. People argue with them, interpret them, feel constrained or protected by them. Their real value may lie less in compliance and more in the sense-making they provoke.

Which leads to a harder question: is there a level of discomfort we should be willing to tolerate in order to create the change we say we want?

Expense policies is another example. Clear rules can feel restrictive, but they also reduce ambiguity and unfairness. Looser, emergent norms can feel freer, but risk uneven application and hidden power. Is it possible to evolve policies that are flexible _and_ fair, freer in practice, without abandoning guidance or reasonableness?

bias for action

Wanting people to try things and learn from attempts isn’t enough. You have to make it _successful to try_. Bias for action turns out to be a skill, not a slogan and it has maturity levels. At its best, it looks like controlled experimentation, not reckless motion. It requires scaffolding.

python vs C

To illustrate this situation, let's take a look at two programming languages.

Python encourages action: forgiving defaults, fast feedback, easy rollback.

C demands precision, upfront correctness, and carries higher consequences for mistakes.

Neither is inherently better (although you might find that enthusiasts disagree with this take) but they produce very different behaviors. Organizations, like languages, encode values into their constraints.

Some conditions for a healthy bias-for-action culture are:

- fast feedback mechanisms and rollback paths, - well-understood impact and blast radius, - frameworks for experimentation with explicit gates for lower risk, - and crucially, structure even in “permission-less” environments.

permission-less cultures

I probably need a whole separate field note on permission-less cultures, because I don't believe they exist - and it hit close to home. Trying to be action-oriented without clear boundaries often creates _more_ ambiguity, not less. Even anarchist systems have rules. “No permission needed” still requires shared understanding of limits.

The phrase “go and act” came up repeatedly, mostly as a warning. It doesn’t work. For many people, playing it safe is more natural, especially in civic or institutional contexts where failure carries reputational or moral weight. Even with frameworks in place, acting can feel risky.

So the real challenge becomes perceptual: how do you show people that the risk is actually small? Could you test something in one year instead of three? Can you make reversibility visible? Signals matter and organizations are constantly sending them, often unintentionally.

culture change takes time

Finally, culture change takes time. It takes months for an organization to acclimate to a new culture and it happens very, very slowly. Institutional safety and psychological safety are distinct but intertwined. And that global contexts complicate everything what feels safe, fast, or acceptable in one region may not in another.

I left class with the sense that the real question isn’t whether we still need managers, but what managers are actually for. If autonomy is the goal, someone still has to design the conditions that make autonomy survivable. Goals, values, guardrails, feedback loops, this is the work that doesn’t disappear just because hierarchy does.

In the next class, we will take a deeper look at the goals, values, and conditions for bias for action, empowerment - and whether managers are no longer needed if processes can be decentralized.

reinventingorganizations.com ↗

Field Note: Growing up amongst my heroes

Civic tech can be quite a small world in many ways and growing up amongst my heroes, was quite difficult. While I shared their ideals, inherited their jadedness, and was inspired by their hopes and dreams - I did not have the experience to follow them into the fight. Instead, I did what I knew best - the coding, the networking, the community building.

I think back to the times that I volunteered at BetaNYC meetups or cleaned up after events Civic Hall. I was in charge of ordering the stickers for Microsoft Civic or picking up enough pizza for the Ambassadors program meetings. I coordinated multiple people, remembered the names of speakers, and at every hack night, checked in with people on the status of their projects.

It wasn't until I came to Newspeak House and I've started to catch up with friends from across the industries that I've worked in that I realized how much of the work that I gave no credit to - because it felt invisible in terms of on paper achievements - was actually so useful and taught me all the variety and depth of skills that I can draw on now.

And everyone's genuine enthusiasm - from the folks that I worked with on drupal community efforts, to the people that I organized civic hack nights with - all are so thrilled for this part of my life, that it really makes me feel both validated and inspired and honestly, full of hope for myself and what I want to achieve.

Back then, I always felt that I could never catch up to my heroes. I always felt it was unfair, we often joked that it was, that I had inherited both their dreams and their jadedness _too_ early in my career, while I was still in college, and had to climb up the ranks before I could do the things that they were doing. And _still_ there was a part of me that was still so inspired that I continued to, despite the burnt out, go back to work in the public service.

Ten years later, my heroes are now my friends, my mentors, and my confidantes. They're no longer people that I once fan girled over after a conference talk or tried to impress on twitter - they're people who I've shared experiences with, talked deeply about issues with, and solved things together with. In so many ways, they are peers - even if they have many more years of experience than me.

Perhaps because I now see them as humans and not as heroes that the perspective has changed. Or perhaps I see myself as competent and capable and able to stand alongside them, rather within their shadows.

John said something interesting to me today and it was that perhaps, one's purpose in life, if not entirely legible internally, will only become legible externally after one's death - and perhaps that means the meaning was more complex - than someone perhaps, who has a clear meaning in life.

And I wondered - ah, perhaps my measurement of success is having one clear thread - a singular purpose in life, by which all of my work and passion weave through. And because my actual purpose in life cannot be contained by a single thread, it doesn't meet my criteria of success.

Ten years later, I'm still inspired by stories from Taiwan's digital democracy

Two weeks ago, I was reviewing the Newspeak House calendar - and in an effort to curtail the loneliness of moving to a new city, I was selecting events that I might be interested in attending. The Decode Taiwan Open Tech: Open source and civic tech meetup in London for latest communities updates from Taiwan immediately caught my eye and as I expected, both g0v and Open Culture Foundation were mentioned in the description.

This really took me back - and it further reinforced why this institution is so unique and brings together so many communities that overlap and do similar kinds of work.

When I first started out in Civic Tech in New York City, I volunteered at Personal Democracy Forum for a ticket to the conference. I was often assigned the role of greeting and guiding speakers. In 2017, 2-3 years into my early journey of civic tech - I'd come in from Boston, where I was working for the Digital Team, to volunteer. And I had the explicit honor of greeting Audrey Tang, the then Digital Minister of Taiwan, and leading her to the speakers ready room. I was an absolute fan girl and the whole interaction was a blur.

Later that day, sitting in the audience - in the gallery above where the volunteers got to look over the stage, I watched - mesmerized as Audrey Tang delivered her keynote - "STORIES FROM FROM THE FUTURE OF DEMOCRACY: TAIWAN" describing the digital ministry's efforts to pioneer "digital democracy", using polis for civic participation and strategies for combatting disinformation.

I was absolutely enamored with the whole thing - from the digital ministry of taiwan to g0v and their decentralized governance for managing the community. In fact, I remember talking to some of the volunteers about wanting to move to Taiwan and join g0v (you know me, I follow the spark).

I made a few friends at the conference who worked in g0v efforts and we kept in touch over the years, even when I dipped out of civic tech into the private sector and then resurfaced as a Code for Canada fellow.

So it was really a walk down memory lane to see Taiwan on the schedule for Newspeak House. I reached out to the organizer of the event for a coffee chat - and they were delighted to meet, and even brought some of their friends.

The coffee chat felt like we'd known each other for years. They asked about the fellowship and then shared some advice regarding research, narrowing scope, and finding your life's purpose. It was so interesting to hear about where the OCF sits - between the civil society, efforts like g0v, and the Digital Ministry.

It reminded me of how I consider Developer Relations to be a bridge - between the product/engineers and the users/customers. And my job is being the runner, to carry information across from side to side, to translate it as best possible for the other side to be able to use of, to maintain the bridge so that they trust it enough to use it, and to make judgements about the urgency of relaying information. DevRel are basically spokespeople in governance, but they also carry so much of the product knowledge and technical expertise to be credible.

Rosalind and I talked about the ambiguity of research and how sometimes in finding one's life's purpose, one must sometimes look to the function and not the theme. It may be that many people I admire know exactly what they want to do with lives, they know what theme they want to impact and they works towards that goal. Other forms of purpose are acceptable beyond themes, for example - perhaps my purpose is being a facilitator of projects and people, or a bridge between systems that do not understand each other.

The visit from the Taiwan delegation gave me so much to think about because it presented a mirror to the past - a time where I was so excited about civic tech and knew very little, in contrast to now - where even though I understand the field and have made an impact, there's still so much more that I want to do.

Stay tuned for a field note about their presentation and all the projects I learned about!

(Rough Draft) Field Note: Reflecting on a Fellowship Where Rules Can Be Rewritten / Who Taught Us the Rules

The first time that I noticed systems meant to support people in fact actually excluded them was the free lunch in elementary school. In first grade, trying to get a free lunch was a bureaucratic rollercoaster. I felt ashamed that my father had to prove his retirement status and income in order for me to be able to eat lunch at school because we could hardly afford lunch otherwise. So that chocolate milk I had every morning when I arrived, it weighed on me heavily because we had to earn it.

Why are systems that are meant to support people so difficult to access?

There is now research that calls this “administrative burden,” the idea that paperwork, verification requirements, and documentation act as gatekeeping mechanisms even in programs designed to support low income families. I did not have that language at the time. I only knew that something about it felt wrong.

Certain systems make it harder for people to access services that they need. And because of the way they are built, people may not even attempt to advocate for themselves.

This is just one of the patterns I have been reflecting on. Systems that I have observed over the years either include some people and exclude others, add friction to the process for the groups most in need, or create an intimidating process that adds too much cost.

My parents were immigrants. They emigrated to Canada and then to the United States in the 1960s, and their perspective on how to interact with systems is so different from mine. My father, in particular, saw rules as flexible, power as attainable, and systems that meant to squash the hopes and ideals of immigrants as unnecessary to pay any mind to. He would bend or navigate around systems that didn't serve him.

For example, he would sign me up for a private elementary school education, negotiate a scholarship that did not exist, and then use my academic abilities to prove it was worth it. He would convince, charm, and negotiate his way into rooms and doors that were historically closed for our people. Part of it was his ability to present as white passing and his British accent, which made him more palatable to the majority, but the bulk of it was his lived experience. Being the other in so many foreign countries gave him this innate confidence that he was owed power, owed a seat at the table, owed an audience.

When it came to high school admissions, my neighborhood in Queens was zoned to one school. On paper, zoning looks neutral. In practice, school zones mirror housing segregation, and housing segregation mirrors income inequality. If you grow up in an underprivileged neighborhood, your local schools are likely to reflect that.

My father was not fond of the assigned school, so we went to school board meetings to try and get an exception without much luck. He would not let this go. He continued to contact the school board and talk to related officials. One day, during the summer before the semester started, he dropped by the high school he wanted me to attend and ran into the new Principal on the grounds in a t shirt and shorts because he was not expecting to meet parents. They became friends. My father explained my situation and suggested that a letter from the Principal inviting me to the school would move the process out of the school board’s hands. He agreed. With that letter, I was enrolled in the school my dad wanted for me.

The admissions process for Queens high schools seems neutral on paper, but it rewards insider knowledge. How to apply. Where to apply. Who advises you. What counts as merit. What paperwork and timing matter. What exceptions exist. Who knows about them.

My parents, and my dad especially, saw this discrepancy clearly because they had to navigate systems repeatedly, first in Canada and then in the United States.

I was different. I grew up in an era where Muslims were othered in society after 9/11. We were not welcome. I internalized rules as fixed and legitimate. To me, breaking the rules was not an option. There were consequences. My father and I often got into arguments about this. I watched him bend systems and get results. I cringed. I assumed he would be punished. He rarely was. He insisted I should do the same, but I had so much discomfort around bending the rules.

Power shows up in small ways. Who gets access to information early. Who understands the process. Who has someone to ask. Who knows the real rules.

Over time, I began to notice that most systems are layered. They are built on precedent and accumulated logic. They persist long after their original rationale. I learned recently about leaseholds in England and Wales, a medieval landholding structure that still shapes modern property ownership. A feudal remnant still determining who owns land and who holds time limited rights.

That was when it struck me again. We live inside sediment. Layer upon layer of inherited logic.

Which is why Newspeak House feels so strange and so special.

Reflecting on these experiences makes me really appreciate the opportunity that the Newspeak House fellowship presents. A house with resources and faculty that we can utilize. We could govern it as a small country, with rules, policies, and processes. We could layer different systems, play prefigurative politics, and build the society that we want to see in the world.

Why is this special?

Because systems in the world exist with such layered inheritance and are hard to change precisely because they are stacked with years of precedent and rules. To change systems like retirement benefits, Medicare, school admissions, or immigration requires thinking out of the box, which is very difficult to do when you are overwhelmed within the system trying to get things done.

Here, at least in theory, we are not bound by centuries of sediment.

At Newspeak House, rules are not treated as fixed inheritance. The system resets every year without full knowledge of the previous systems. It is a space where rules can be created, contested, replaced. It is where you can analyze systems as systems and prototype alternatives because you have the space to design, test, and create better systems than the ones you have interfaced with in the real world.

This feels so special and so unique because out in the real world we are always pushing against sedimented legitimacy.

To change the world in any meaningful way, we need to be able to displace systems, to analyze, shake up, and create purposeful improvements that do not simply get absorbed by the old logic.

The awards exercise is a smaller example of this. This is not about who wins the award. It is about building a system of evaluation. How do we decide, asynchronously and as a committee, what counts as impact and evidence. How do we encode our values into an algorithmic process, with or without AI. And then how do we publicly justify those decisions as part of the system we created to do these evaluations.

The core of this for me, and the part that hits home the most, is the question of legitimacy.

As a young girl, I took rules and systems and laws as the highest form of legitimacy possible. I did not believe they could be bent or modified or made an exception to. I watched my father break rules and somehow still get his way. Later in life, I watched senior officials and leadership do the same.

New systems do not displace old ones simply by being better. They displace them by earning legitimacy.

Most people spend their lives navigating institutions they did not build, arguing within rules they did not write. Here, we are being handed the chance to design systems before they harden.

I sometimes wonder whether that feels different depending on how you grew up. If you were always comfortable inside systems, perhaps designing one does not feel radical. But if you grew up feeling governed by rules you could not bend, watching your family fight administrative friction just to access what was promised to them, then being handed the chance to create rules from scratch feels almost unreal.

And even here, we default to precedent. We reproduce familiar hierarchies. We hesitate to truly unsettle inherited logic.

Designing systems is heavier than criticizing them. It forces the question of authority. What gives us the right to decide what counts. What makes our process legitimate.

This house gives us something rare. The space to practice constructing legitimacy intentionally. The question is whether we will recognize the opportunity in our hands before the time is up.