Working on vs. Solving
General Managers get the job done
Working on vs. Solving
Very few organizations are actually designed to solve a problem, rather than just work on it.
Most of the institutional infrastructure in philanthropy is designed for working on problems — funding research, publishing roadmaps, building fields, coordinating grantees. Consider these mission statements:
Red Cross: “prevents and alleviates human suffering”
MSF: “brings medical humanitarian assistance to victims of conflict, natural disasters, epidemics or healthcare exclusion”
WWF: “to conserve nature and reduce the most pressing threats”
Save the Children: “to give children a healthy start in life, the opportunity to learn, and protection from harm”
American Cancer Society: “to save lives, celebrate lives, and lead the fight for a world without cancer”
American Heart Association: “to be a relentless force for a world of longer, healthier lives”
Nan Ransohoff recently wrote that we need more “general managers” for the world’s most pressing problems — founder-types who feel personally responsible for delivering a specific outcome, not just working on it.
Working on it: This is what I do at Convergent Research. Every day, my team and I show up and try to figure out how to make our existing FRO portfolio successful and how to find new, awesome FROs. We work on it, hard, every day. We’re working on bottlenecks in science infrastructure - but the new institutional model we are building is not a measurable state of the world - there is no quantifiable ‘finish line’. We can’t point to a number and say ‘done.
Solving it: Dr. Donald Hopkins was given a mission by Jimmy Carter for The Carter Center: zero cases of guinea worm. The Guinea Worm Eradication Program is responsible for getting the world to zero. They show up to work every day too. But they are doing more than working on it — they believe they can get to zero, and they are focused on doing it.
Working on a problem means doing excellent work within it. Solving a problem means picking a finish line and owning whether you get there.
Working on a problem is what most organizations do, and it’s valuable. Coefficient Giving has directed over $5 billion in grants across global health, biosecurity, AI safety, farm animal welfare, and scientific research. Save the Children reached 113 million children last year across 113 countries. The American Heart Association has invested over $6 billion in cardiovascular research, and deaths from heart disease have been cut in half since its founding — an extraordinary achievement.
This is important, essential work — but there is an extent to which it is a constant battle. Save the Children isn’t working to end all childhood hunger or create world peace — it is reacting to a complicated and dynamic world putting families at risk. The AHA is trusted and respected across the world, and its research has saved millions of lives. But it’s not designed to figure out the fastest way to end heart disease and get there. It’s a system meant to self-perpetuate — raise funding, identify good uses of those funds, deploy funds, make discoveries, save lives, rinse and repeat. Coefficient Giving funds solvers — but is not itself one. I can imagine a world where any single program is no longer needed. I can’t imagine a world where CG isn’t needed.
Working institutions are valuable and necessary. But Solving institutions promise the rarest of outcomes: a problem ended. A line item removed from the list of things our descendants will have to worry about.... The Montreal Protocol didn’t work on ozone depletion — the people solved the problem (CFCs destroying the ozone layer) by building the international mechanism to ban them. The ozone layer is now on track to recover to 1980 levels by mid-century. The Drugs for Neglected Diseases Initiative set a goal of 25 new treatments for neglected patients by 2028. The Carter Center’s Guinea Worm Eradication Program took on a disease with no vaccine and no cure and drove it from 3.5 million cases to 10.
Scope to something one person can own
So we’ve discovered the key to fixing the world! We just have to get people to switch from working on to solving. That simple.
Unfortunately not.
I don’t think we can solve World Hunger, Global Peace, Inequality, Extremism, or even Biosecurity. They’re all too big, with too many players and too complex a system. For now, these have to be working-on problems.
But you can solve specific, definable problems within them. The trick is finding the right one.
Biosecurity is the work of preventing, detecting, and defending against biological threats — from engineered pathogens to natural pandemics. It matters because advances in biotechnology and AI are making it easier to create dangerous biological agents faster than our defenses are scaling.
Biosecurity decomposes into at least three pillars: prevent biological threats from emerging, detect them if they do, and defend against them once detected. Prevention means things like stopping bad actors from ordering dangerous DNA sequences, verifying who has access to high-risk biological materials, and ensuring AI systems can’t be used to design pathogens. Detection means spotting novel diseases before they spread — through wastewater monitoring, airport surveillance, and rapid diagnostic testing. Defense means protecting people once a threat arrives — stockpiling respirators for essential workers, making buildings resistant to airborne pathogens, developing vaccines and treatments fast enough to matter, and coordinating a global response. Total investment needed: north of $2.5 billion over five years, across a dozen distinct problem areas.
No single person can own all of this - Biosecurity is really a label for a dozen different problems. A GM needs to map the space and define a specific problem she can hold in her head - one she can credibly solve. She decomposes: prevent, detect, defend. Within defend, she picks built environment — transforming indoor spaces to suppress pathogen transmission. Within that, she picks far-UVC: ultraviolet light at 222nm that kills airborne pathogens and appears safe for continuous use in occupied spaces. It’s a real technology with a real evidence base, and nobody owns the problem of getting it into buildings at scale.
Pick a finish line
If a GM wants to solve far-UVC deployment. She needs to figure out what that means — define a finish line.
The finish line has to be a state of the world, not an activity. “Fund 50 grants in biosecurity” is an activity. “far-UVC installed in 100,000 occupied spaces” is a state of the world.
The finish line has to be specific enough to know if you’ve achieved it. You should be able to look at it in five years and know unambiguously whether you succeeded or failed. “Increase far-UVC penetration” is pretty hard to ‘fail’ at. “Zero cases of guinea worm” is pretty clear.
For far-UVC, a finish line could be:
“far-UVC is installed and operating in 100,000 high-traffic occupied spaces in the US by 2035.”
“Indoor respiratory transmission is reduced by 80% in equipped spaces, and far-UVC is installed in 75% of US hospitals, schools, and transit hubs by 2035.”
“far-UVC unit costs are below $500 per classroom-equivalent, and the technology is included in standard building codes for new construction in occupied public spaces.”
Which you choose depends on the theory of change. For this essay, let’s use: “far-UVC is installed and operating in 100,000 high-traffic occupied spaces in the US by 2035.”
Rationally believe you can get there
A finish line alone isn’t enough. The GM has to believe she can get there — and she has to be able to explain why.
Each of Convergent Research’s Focused Research Organizations is solving something specific in science. Andrew Payne, the CEO of E11 Bio, is making mapping a mammalian brain 100x cheaper. Walking into the FRO, he had a credible plan for how to get there. He knew what technology needed to be advanced, what needed to be scaled, what platforms needed to be built. He had a good sense of the resources required and how he would deploy them. There are always unknowns, but a credible plan is needed to solve a problem. Andrew is solving brain mapping, not just working on neuroscience.
A GM for far-UVC deployment would need a similar plan. She’d need to answer: what’s blocking deployment today, and in what order does she address the blockers? The answer might look like:
Close the remaining safety and efficacy gaps — fund the long-term exposure studies and the adequately powered real-world trials that regulators and building owners need.
Drive hardware costs down from $10,000–$18,000 per classroom-equivalent to under $500 through investment in manufacturing and new R&D.
Work with the six standards bodies whose overlapping jurisdictions currently prevent deployment at scale.
Create procurement mandates in federal buildings, hospitals, schools, and transit hubs.
Structure advance market commitments.
Build the installer training infrastructure and HVAC integration protocols.
Run a public awareness campaign so that when the technology is cheap and proven, people want it.
Blueprint Biosecurity’s blueprint covers most of this ground — they’ve mapped the bottlenecks, identified the research gaps, and laid out recommendations. That’s essential working-on work. A GM steps in to solve: to be responsible for all of it getting done in the right order, at the right pace, in the most capital-efficient way.
What the GM needs to be able to do
To solve far-UVC deployment, a GM has to fund clinical trials, invest in hardware companies, create new organizations, build buyer coalitions, lobby for building codes, and run a public awareness campaign - simultaneously.
To close the safety gaps, she needs to make grants — funding university labs to run the long-term exposure studies, paying for the cluster-randomized trials in real buildings that nobody has done at adequate scale.
To drive hardware costs down, she needs to invest — taking equity stakes in excimer lamp manufacturers to fund production scale, or backing the solid-state startups that could transform unit economics entirely.
To fill gaps in the ecosystem, she needs to start organizations — both for-profit and nonprofit. A company to handle building-by-building sales and installation. A nonprofit installer training program. An HVAC integration standards body. A clean-air certification entity.
To unlock volume before the mandates arrive, she needs to structure advance market commitments — assembling a coalition of hospital systems, school districts, and government agencies that commit to purchasing at a target price, giving manufacturers the confidence to invest in scale.
To make deployment self-sustaining, she needs to change policy — working standards bodies, pushing building code reform at the local or state level, driving procurement mandates for federal buildings and VA hospitals, shaping FDA guidance.
And to make any of it stick, she needs to change public perception — a public awareness campaign, influencer engagement, consumer-facing air quality ratings, whatever it takes so that building owners actually want this technology once it’s cheap and proven.
Few institutions can do all of these things at once. A foundation can grant but usually* can’t invest or lobby. A fund can invest but can’t grant or start nonprofits. A nonprofit can advocate but can’t* take equity positions. Fewer still hand all of these mechanisms to a single person to deploy at her discretion.
But it has been done. The Guinea Worm Eradication Program did exactly this. Hopkins and Carter didn’t just fund research on guinea worm. They built country-level eradication programs from scratch in 21 countries. They worked with DuPont to develop a cloth water filter that didn’t previously exist. They created cash reward systems so villagers would report cases — a market-shaping mechanism that made surveillance self-reinforcing. Carter personally pressured heads of state to commit resources and got the World Health Assembly to pass eradication resolutions. The entire campaign ran on behavior change — training hundreds of thousands of community volunteers to teach water filtration in thousands of languages, because there was no drug and no vaccine. Grants, new organizations, product development, political pressure, public engagement — all wielded by a small team accountable to a single number: zero.
Another example is Bill Gates and climate. He framed the problem — five sectors responsible for nearly all emissions, a “green premium” between dirty and clean alternatives in each — and then built a vehicle for each stage of the pipeline. Breakthrough Energy Ventures, with $3.5 billion in committed capital, invests in companies closing the green premium. Breakthrough Energy Fellows supports scientists at the pre-venture stage with up to $3 million to turn lab-stage climate technologies into companies. Catalyst funded first-of-a-kind commercial projects to bridge the deployment valley of death. A policy team worked governments on regulatory frameworks. Five tools, one outcome. But Gates isn’t the GM for climate — he’s also running a foundation focused on malaria, global health, education, and Alzheimer’s. And “close the green premium across five sectors of the global economy” may be too large for any single person to own. Gates built the tools. What’s missing is a person whose entire job is to wield them toward a defined finish line.
*Actually foundations and nonprofits CAN do this, they just often aren’t structured organizationally such that they DO do this.
What does a GM need to get all that done?
Three things: flexibility to act, enough capital to finish, and time to see it through.
Flexibility means she can deploy capital across grants, investments, new organizations, AMCs, and policy work without going back for approval on each one. If she needs a funder committee to sign off on every decision, she can’t solve the problem.
Capital means enough to finish, committed upfront. Some estimate that roughly 20–25% of the total capital needed to solve a problem, held at the GM’s discretion, is enough to quarterback it to completion. If $2 billion needs to flow into a space over a decade, the GM needs $400–500M she can direct. That’s a lot to give one person. But the choice is clear — the GM can spend her time solving the problem or split it between solving and fundraising.
Time means the capital is committed for the duration of the problem, not renewed annually. Annual fundraising cycles are the enemy of the solving posture. They pull the GM toward donor comfort and legible short-term wins instead of the hard, unglamorous, multi-year work the problem actually requires. This is one reason the finish line needs to fit within roughly a decade — much longer and the organization drifts from solving back to working on.
Maniacal dedication means the GM cares about the problem so deeply she can’t leave it alone. Carter asked about guinea worm case numbers from hospice care. Henderson ran smallpox eradication like a wartime operation — traveling constantly, bypassing bureaucrats, pressuring governments publicly when they moved too slowly. A talented, well-resourced person who treats this like a job won’t get it done. The problems that need GMs are too hard, too long, and too full of setbacks for anyone who isn’t obsessed. This is why venture capitalists care as much about the founder as the market — and why the search for a GM looks more like founder selection than executive hiring.
Hold her accountable
All of this raises an obvious question: how do you trust someone with this much capital and this much autonomy?
Talent problems start with search and selection — the more critical and empowered the role, the more careful you have to be. Before capital is committed, the GM maps a credible path to the finish line. Not a theory of change — a sequence of decisions with observable consequences. Here is the end state, here are the eight things that have to happen, here is the order, here is what I’d try first if I’m wrong about the sequence.
Ongoing accountability tracks two things. Outcome milestones measure whether the world is actually changing — are buildings being equipped, is the cost curve dropping, are procurement mandates advancing. Decision milestones measure whether she’s making good calls at key forks — did she kill the program that wasn’t working, did she redirect when evidence changed, did she resist sunk-cost logic when something should have been wound down.
The terminal condition: the capital is committed to the problem, not to the GM. If she isn’t delivering, you replace her. Bet on a person but commit to the outcome.
Funders for a repeatable model
Can we make this repeatable? Can we identify a problem, define the finish line, build the vehicle that gives a GM flexibility and capital and time, find the person, and hold her accountable — rather than waiting for the rare individual who creates the whole thing from scratch?
At Convergent, we did something like this for science. Important research bottlenecks weren’t getting solved because no existing institution was designed to solve them. So we built the Focused Research Organization — purpose-built teams with fixed timelines and dedicated funding to de-bottleneck entire fields. FROs were one answer to the question “what institutional form does this problem need?”
The GM model is another answer, for a different shape of problem. FROs solve science bottlenecks that produce public goods. GMs solve execution bottlenecks that require coordinating technology, policy, markets, and institutions toward a defined outcome.
The problems are real. far-UVC deployment, pathogen-agnostic surveillance, essential workforce protection, lead diagnostics, industrial heat decarbonization — these are solvable problems with describable finish lines and credible roadmaps. The people who could GM them exist. What’s missing is the funders willing to commit capital on these terms: flexible, sufficient, long-duration, and accountable to outcomes rather than activities.
That is a challenge to the philanthropic world. The current model — annual grants, program officer approval, activity-based reporting — is designed for working on problems. It works well for that. But if we want to solve problems, we need to build something different.
Who will be the funders who enable this work?
Thanks to Nan Ransohoff, Genevieve Rogers Buchet, Beba Cibralic, Robbie Barbero, Sasha Chapin, Zander Farkas, Parth Ahya, Caleb Whatney, Ben Kuhn, and Ilan Gur


This was sooooooo good that I immediately had to generalize to one more level of abstraction and write a followup. Exactly the conversation all of us in the sector need to be having. https://zachill.substack.com/p/outcome-shaped-organizations