I think underrated in the “why is working-on-government neglected?” discourse is just rote personal frustration on the part of would-be innovators. It’s much easier to postulate the need to work around it because of black-box failure-states than it is to go and read 85 pages of personnel policy (or whatever) and formulate an actionable thesis over a decadal time horizon. In other words, much ‘venture philanthropy’ presupposes venture-timeline yields. But a 250-year ship don’t always turn like that.
We talk about the greatest generation in awe then take none of the lessons. They were builders, of infrastructure, science, and state capacity. They were so good we've coasted on it ever since. So why do we not do it now?
One can hardly say the context is different, if anything parallels are shockingly similar.
It's not for everyone, but for those that take public service as a vocation they should have every tool available, every arrow in the quiver. We focus on elected officials who make big promises , but even if they are earnest who actually needs to do the work and implement it?
The AI philanthropy question may not be only whether new money helps government work better. It may be who gets to define what “working better” means. Capacity is not neutral. A state can become more capable of delivering benefits and more capable of denying them. More capable of reducing burden and more capable of automated exclusion. That does not argue against philanthropy engaging government, it perhaps argues for being more precise about the terms of engagement. If philanthropy funds not just tools, but the intermediaries and reform language through which public agencies come to understand their own deficits, then the governance of the help matters as much as the help itself.
Amanda's point about capacity neutrality is the right one to press on. A more capable state can automate inclusion or exclusion depending on who controls the optimization target - and "fund better tools" doesn't answer that.
One structural answer worth considering is the public/private corporation model - Nav Canada being the clean case. ATC was transferred from government to a nonprofit corporation in 1996 with a stakeholder board: airlines, controllers, airports, general aviation, government. Operationally independent, self-funded through user fees, legal charter that constrains what the organization can do. The governance of the help is built into the structure from the start, not left to whoever holds the contract at a given moment.
The model seems underexplored in Jen's framework. Where the task is bounded enough to define success and can support a non-appropriations revenue flow - user fees, transaction fees, whatever fits the domain - you can embed the answer to Amanda's question in the governance architecture rather than depending on the goodwill of whoever holds the funding relationship at a given moment.
The philanthropic question then becomes: is there a role for third-wave capital in standing up that governance architecture, rather than the technology itself?
Sitting here at UN Open Source Week, I’m reflecting on the donor fixation with "catalytic" interventions—small, fast-impact tech pilots. The underlying assumption seems to be that if we build a brilliant digital tool in a capacity vacuum, government budget teams will naturally scuttle to finance it.
But we’ve seen this script play out in civic tech, DPI, and now AI pilots. The reality is that "catalytic" funding often leaves governments holding digital public goods they have zero institutional capacity to maintain, or relies on digital goods that lack a viable long-term vendor model.
The issue isn’t the tech; it’s the absorptive capacity of the state. If we want digital transformation to stick, the real "catalytic" work is the slow, unsexy business of building state capability. In concrete terms, this means funding multi-year cohorts of civil servants to hit their own milestones, allowing time for civil service job profiles, procurement policies, and state budgets to catch up to the technology.. while undergoing the shocks of political cycles. Hugh Cole recently called this work in the city of cape town "critical but unglamarous". Glad there is a fund rallying to support unglamarous rather than catalytic..
Philanthropy is an interesting concept. It's been around as long as super wealth has been around. But, is this a reasonable approach in a democracy? Philanthropic funds are directed at and governed at the pleasure of the wealthy donator. Perhaps humanity would be better served if wealth was more appropriately taxed and expenditures then governed by the electorate? Priorities of the wealthy may not be priorities of humanity? Good governance and an effective public service should be a democratic priority.
I'm not sure if it's intentional, but this is showing as black text over a dark grey background, making it impossible for me to read. Is that a substack bug?
I think underrated in the “why is working-on-government neglected?” discourse is just rote personal frustration on the part of would-be innovators. It’s much easier to postulate the need to work around it because of black-box failure-states than it is to go and read 85 pages of personnel policy (or whatever) and formulate an actionable thesis over a decadal time horizon. In other words, much ‘venture philanthropy’ presupposes venture-timeline yields. But a 250-year ship don’t always turn like that.
We talk about the greatest generation in awe then take none of the lessons. They were builders, of infrastructure, science, and state capacity. They were so good we've coasted on it ever since. So why do we not do it now?
One can hardly say the context is different, if anything parallels are shockingly similar.
It's not for everyone, but for those that take public service as a vocation they should have every tool available, every arrow in the quiver. We focus on elected officials who make big promises , but even if they are earnest who actually needs to do the work and implement it?
So well said!
The AI philanthropy question may not be only whether new money helps government work better. It may be who gets to define what “working better” means. Capacity is not neutral. A state can become more capable of delivering benefits and more capable of denying them. More capable of reducing burden and more capable of automated exclusion. That does not argue against philanthropy engaging government, it perhaps argues for being more precise about the terms of engagement. If philanthropy funds not just tools, but the intermediaries and reform language through which public agencies come to understand their own deficits, then the governance of the help matters as much as the help itself.
Amanda's point about capacity neutrality is the right one to press on. A more capable state can automate inclusion or exclusion depending on who controls the optimization target - and "fund better tools" doesn't answer that.
One structural answer worth considering is the public/private corporation model - Nav Canada being the clean case. ATC was transferred from government to a nonprofit corporation in 1996 with a stakeholder board: airlines, controllers, airports, general aviation, government. Operationally independent, self-funded through user fees, legal charter that constrains what the organization can do. The governance of the help is built into the structure from the start, not left to whoever holds the contract at a given moment.
The model seems underexplored in Jen's framework. Where the task is bounded enough to define success and can support a non-appropriations revenue flow - user fees, transaction fees, whatever fits the domain - you can embed the answer to Amanda's question in the governance architecture rather than depending on the goodwill of whoever holds the funding relationship at a given moment.
The philanthropic question then becomes: is there a role for third-wave capital in standing up that governance architecture, rather than the technology itself?
I think it should be up to the voters, and the elected officials that represent them, that decide how that capacity is used
Sitting here at UN Open Source Week, I’m reflecting on the donor fixation with "catalytic" interventions—small, fast-impact tech pilots. The underlying assumption seems to be that if we build a brilliant digital tool in a capacity vacuum, government budget teams will naturally scuttle to finance it.
But we’ve seen this script play out in civic tech, DPI, and now AI pilots. The reality is that "catalytic" funding often leaves governments holding digital public goods they have zero institutional capacity to maintain, or relies on digital goods that lack a viable long-term vendor model.
The issue isn’t the tech; it’s the absorptive capacity of the state. If we want digital transformation to stick, the real "catalytic" work is the slow, unsexy business of building state capability. In concrete terms, this means funding multi-year cohorts of civil servants to hit their own milestones, allowing time for civil service job profiles, procurement policies, and state budgets to catch up to the technology.. while undergoing the shocks of political cycles. Hugh Cole recently called this work in the city of cape town "critical but unglamarous". Glad there is a fund rallying to support unglamarous rather than catalytic..
Philanthropy is an interesting concept. It's been around as long as super wealth has been around. But, is this a reasonable approach in a democracy? Philanthropic funds are directed at and governed at the pleasure of the wealthy donator. Perhaps humanity would be better served if wealth was more appropriately taxed and expenditures then governed by the electorate? Priorities of the wealthy may not be priorities of humanity? Good governance and an effective public service should be a democratic priority.
Really valuable piece for those of us working (and volunteering) on both sides of this. Appreciate the vantage point.
I'm not sure if it's intentional, but this is showing as black text over a dark grey background, making it impossible for me to read. Is that a substack bug?
yes, I think that's substack bug? Maybe check setting?