Grantmaking as investigative journalism

"Whatever idea you have of what a 'grantmaker' is, please take it and throw it out entirely." That's how I closed my latest pitch for someone to join my team.

Over the last four years of nominally being a 'grantmaker', I've spent a minority of my professional attention on grantmaking as such. And even within that share of time, my work has often looked very different from grantmaking as it's usually understood. If I were to retroactively write my job description, the grantmaking section would look something like:

• 30% biosecurity research
• 30% ARPA-like program management
• 30% headhunting
• 10% investigative journalism

That last part shouldn't come as a surprise — after all, we often talk about grant investigations. And yet, I routinely encounter confusion about this part of the job from people who seem to think my job looks dramatically different than it actually is. They picture 'grantmaking' as something out of Moneyball, imagining that I spend my working days split between a whiteboard and a stack of written proposals, which I… don't.

Such confusion is understandable (some grantmakers do in fact operate this way), but we need to get rid of it. If you lack good reporter's tools in your kit, you'll leave value on the grantmaking table. For the new crop of philanthropic leaders, this would be a costly error — and one I expect to compound as AI changes the nature of capital allocation decisions.

The case for investigation: Surfacing private information

The value of investigative chops in grantmaking largely comes down to surfacing important, non-public information from other human beings.

This has probably always been true — it certainly was when I got into this field, before the first generally useful language models launched — and I expect it'll become increasingly salient. For problems that resist solutions from first principles, Claude (and its peers) will remain reliant on available information to think and act effectively. This means that as AI systems approach parity with me on tasks involving public information — research, analysis, planning — working with private information is where I expect to add a lot of my distinct value as a human laborer.

There are two questions for which you, as a human grantmaker, should master the art of gathering non-public information: Why does this human behave this way? Why did this thing happen in the world?

Understanding people: Why does this human behave this way

So much of good grantmaking is, and will be, about understanding people.

My best work has come down to deciphering whether a grizzled industry veteran proposing to build a new org will actually stay focused, or be distracted by the next shiny object. It's been about intuiting whether the grantee you're doing a reference call for is actually, literally "the best person in the entire world for this project," or just a nice guy to work with.

And this goes beyond the evaluation frame, applying equally to the very substantial part of good grantmaking that comes down to management. Good investigation comes down to figuring out whether the two advocacy leaders I'm hoping to wrangle into a coalition just haven't found the time to coordinate, or actually just can't stand each other. Or getting honest feedback on whether I'm screwing something up (I often am), from someone who feels that their entire organizational budget depends on my view of them (it sometimes does).

Doing any of these grantmaking functions well hinges on ably eliciting and reading the emotions of other humans, and here I just think that you (if you're a human reading this) will have an edge over our silicon colleagues for a while longer.

Of course, your favorite transcription app will eventually be trained on visual body language data to surreptitiously read your colleague's anxiety levels during Monday All Hands; even then, I'm betting it won't beat you at prompting people to share what they actually think and feel.

Causal attribution: Why did this thing happen in the world?

Just as often, the job is causal attribution: why did this thing happen in the world? And, often, you'll need to untangle what happened based on data that does not — and maybe should not — become public by default: Why did this lawmaker support that policy? Who was responsible for this AI company making that voluntary commitment? How come this funder is betting on that new area?

Generally, answering these questions well requires fetching context stored only in human minds, out of reach for even the cleverest digital agents. Codex doesn't really have a clue why President Trump's executive order on biosecurity from last year was written the way it was, but I do. No amount of tokens will negate my advantage on questions like this anytime soon.

I've faced million-dollar renewal decisions that hinged partly on whether a grantee shaped legislation in the way their annual report claims. You could tackle that situation by BOTEC-ing the base rate of white papers leading to policy change. But there's another way: find the person who knows the answer and ask them in a way they'll answer honestly! Doing this legwork has often given me faster, better, and correct answers to grant decisions; and you can reap these fruits too, provided that you execute it well.

So you want to be a sleuth?

If you buy my case for developing a nose for investigation, what should you do?

Find the person who knows the answer…

The first and extremely basic answer here is: get out there! You've got to pound the pavements. So much of the actionable intel for evaluating and managing grants resides at happy hours, conference side events, and two and a half minute Signal calls. For as long as I've made grants to shape US and EU policy, I've made regular trips to DC and Brussels to hear from the people at the center of grantee routes to impact.

…ask them in a way they'll answer honestly

Figuring out who knows what you want to know is only half the battle; the rest is to elicit sensitive information skillfully.

First and foremost, you won't get very far without being a good listener. There's just absurd amount of alpha in actually listening carefully to people. People sitting on important, actionable insights are often itching to share them, but they need to feel that you're present enough to receive them. Moreover, if you're interrupting or zoning out after every other sentence, you won't get a full enough picture to read between the lines.

Active, genuine listening is a learnable skill with a simple 80/20 recipe for progress: talk less, smile more, and read this book. Going beyond the basics, my meditation practice has helped tremendously by quieting all the mental chatter that gets in the way of really hearing someone.

Secondly, there's an art to asking questions in a way that permits honest answers.

If you email a lawmaker at their official email, asking "Hey, did you just pass this law because my grantee, the Center for Think Tank Reports, just wrote a report that I funded them to write?", I'm sorry, you're ngmi. There's just absolutely no world where they are going to ever say yes to that question.

First, you've trapped your interlocutor with undesirable implications. If the answer to "Did we successfully lobby you?" is in fact 'yes', you conjure fears of undue influence for a savvy policymaker whose views aren't supposed to come from a single outside advocate. On the flipside, if you effectively asked "Was this NGO report actually just useless?" and the answer is in fact 'yes', you've created a frame where answering truthfully involves dissing the report authors, who generally will be respected, well-liked leaders.

The second aspect to manage here is the medium: generally, people can speak most freely in-person, followed by — in decreasing order of candor — phone calls, video calls, and emails. So, grab a café table near the Eisenhower EOB or Berlaymont, ask questions that can be answered comfortably, and listen up.

Investigating with integrity

The next page you can take from the investigative journo's playbook is carrying yourself with integrity. I'm obviously not encouraging you to emulate the latest hit piece from your Least Favorite Journo; do better.

Step one is respecting your sources. The reason that off-the-record/on-the-record is a somewhat stable equilibrium in journalism (despite not being actually binding, beware!) is that respecting source confidentiality pays off over time. If someone tells you that something is confidential, it's confidential. Don't share it! Extremely few sentences that start with, "I'm not supposed to tell you this…" should end with anything other than "… so I won't."

The next thing you must do is center your digging on actually making something better in the world. Gathering intel has many close relatives — gossip, status games, and palace intrigue — that will both distract and undermine your credibility in the long run. Knowing every blow-by-blow in some beef between two NGOs isn't cool; you know what's cool? Actually finding and using information to make better grants, empower better people, and solve some damn problems for the world.

The limits to journalism-style grantmaking

A big area where investigative journalism won't carry the day is in choosing what to investigate in the first place. For all my respect for the field of journalism (they're heroic), I'm continually flabbergasted and frustrated by its utter inability to prioritize what to cover. With a few rare exceptions, journalistic priorities seem completely detached from what actually matters. You shouldn't have to be a hard-nosed utilitarian to agree that we should focus on issues that actually make a big difference in people's lives. What's worse, I don't think this is simply a matter of clickbait shaping incentives; I see a culture that focuses on the unique and fascinating rather than on the important and consequential.

Don't be like that.

So throw out the whiteboard-and-proposals picture, but don't entirely replace it with a press badge either. Deciding where to point your attention takes cause prioritization, in the tradition of vanilla effective altruism: adding up the numbers, running through the philosophy, taking a cold, hard look at what actually changes the world. Borrow the reporter's craft — the sourcing, the listening, the discretion — but leave their news judgment. Chase the stories that help you make a difference, even if it's not one that's most fun to tell.

Thanks to Emily Ge for our conversations about this idea, and to Max Schons for suggesting a reading-friendly mode for these essays.