How to Apply for an Emergent Ventures Grant

If you're reading this, you're probably a good candidate.

Emergent Ventures is a grant program administered by the economist and blogger Tyler Cowen. It’s awarded for a wide range of reasons. For example, I got one to build out a UAP detection network, but he also just awarded one to the author Helen DeWitt for her creative work.

There’s a wide range of grant sizes, from a few thousand up to apparently $175k at least (though I think awarding that much is relatively rare). If you have an interesting project, it’s worth filling out the application. It’s not very long (<1500 words), reasonably interesting to write, and they get back to you very quickly. I believe Tyler personally interviews every awardee.

The first section is your personal story, and the instructions include:

We probably don't care much about your formal education, credentials, or awards, unless they're particularly germane to who you are or your idea.

I think you should believe them here; they’re not just signaling. I didn’t include much in the way of credentials, but did explain my motivations and, in particular, told a story of ever-growing ambition.

I’m not including my bio in this post, but for your benefit, below are my answers to the second and third parts of the application.

2. A Mainstream View that I Agree With:

Here I answered with something everyone implicitly believes or demonstrates, but won’t actually say out loud. I get the feeling that this type of answer goes over well if thought-through and novel.

Amazon.com is just great. This is mainstream in that it’s widely believed in practice (Amazon has among the highest customer satisfaction among retailers), but saying this at a Brooklyn party will likely give you pushback on workers’ rights and on killing mom-and-pop stores.

However, aside from the ample consumer surplus from cost savings, there are enormous positive externalities from reliable, one or two-day delivery of hundreds of millions of goods, particularly for a hardware developer. Hardware development, compared to software development, is notoriously slow, complex, and costly. This is largely because the inputs are physical goods rather than software packages and libraries. Software engineers can instantly download a tool when needed, often at no cost. In contrast, hardware engineers have to purchase items and wait for them to arrive before they can test anything.

R&D is a continuous cycle of testing and iteration, so the speed of development is roughly the reciprocal of this cycle time. Amazon, with fast delivery and a large catalog of technical goods, has shortened hardware cycle times from 1-2 weeks to 1-2 days. In practice, this means that hardware development starts to approach the pace and expense of software development.

3. Project Proposal:

Though UFOs are reported by the Navy, and top officials like the Secretary of State, quality data is sorely lacking.

The issue is twofold: Respectable civilian organizations find the topic too stigmatized or unserious to devote resources to, while the military classifies relevant data by default (sensor data reveals capabilities).

Crowdsourcing UFO sightings data is a third pathway, and this has been attempted through mobile apps. However, a camera phone isn’t suited for distant, airborne objects, and a 2D image can only give apparent size and velocity. Further, they require line-of-sight, and the camera needs to be pointed at the object (aside from fisheye lenses). The result is generally a single, zoomed-out photo of objects of indeterminate size, distance, velocity, and acceleration.

Passive radar is the ideal solution. “Passive” meaning it uses existing terrestrial radio broadcasts as illumination, so no powerful radio transmitter or FCC license is needed. Like other radars, it uses the delay between the outgoing signal and its echoes to determine range, and the Doppler shift in the frequency to determine velocity. By directly measuring true range and velocity at 30+ miles, each radar node collects robust data across huge areas.

The nodes upload their data to an aggregation server, which generates a global map of radar hits. With enough overlapping nodes, individual objects could be corroborated by several, if not dozens, of detections. The velocity and position data could then be scrutinized for anomalous behavior, such as unusual speeds, turn rates, or accelerations.

Distributed aircraft detection networks are not new. If you’ve ever tracked a flight, you’ve used one. Flight tracking networks like FlightAware and Flightradar24 work by aggregating data uploaded by amateurs using home-built receivers collecting nearby planes’ ADS-B transponder broadcasts. The networks then sell the data via API to airlines and mobile apps. If you live in an underserved area, the networks will often even send you a detection kit for free to improve their coverage.

ADS-B flight tracking requires that aircraft broadcast their location, whereas passive radar can track anything with a sufficient radar cross-section, broadcasting or not.

A system like this is only recently possible. In 2004, a NATO researcher built a passive radar using FM radio, which required a computing cluster for the signal processing. Today, a Raspberry Pi 5 (released late 2023) is fast enough to work and enables full radar kits for under $500.

A grant would accelerate the development of a network for passive radar. I’ve been working on this full-time for the past year, and this will continue to be my primary project in the years ahead. With about $30,000 in grant funds raised and a similar amount personally contributed, most of the individual node development is finished. We’ve established a legal strategy, tested and down-selected the hardware, and polished the software for non-technical users.

4 prototype radar nodes are now live, and their data is being used to develop aggregation and tracking software. A live instance of a node’s output is here: radar3.retnode.com (red dots are radar detections, green dots are plane transponder overlays for validation).

Further funds will be used to accelerate the development of an aggregation server and a 20-node alpha test to known volunteers, beyond what my personal finances could otherwise afford. This will validate the concept, expose the system to external users, and enable scale testing.

I’ve not included the budget section, but suffice it to say, it was only modestly detailed.

If You’ve Made it This Far You Should Apply

You read 1000+ word posts on Substack. This is a positive signal you should listen to.

At the very least, try drafting a project proposal to sharpen your thinking around the topic. Then you can write the “mainstream view” answer because, honestly, it’s an interesting thought exercise. And then you might as well write a bio…