# Coordinated research program “Coordinated research program” is, quite frankly, a made up term to cover a wide range of activities that nevertheless share enough important similarities that we can talk about them as a group. Like most [[Nebulous]] things, any tight definition would have so many exceptions and caveats as to be useless. So instead, I’ll define a coordinated research program by describing what it is and what it is not so you can start to pattern-match. Coordinated research programs: - Do research work that doesn’t make sense for a single academic lab or startup to tackle for some systemic, institutional reason. There are many reasons *why* it might not make sense: it might not fit in an academic lab because it requires a lot of repetitive engineering and wouldn’t create publishable results; it might not fit in a startup because it would be hard to capture the value it creates or isn’t creating a sellable product at all. Unfortunately “nobody will fund me to do this” is not a systemic, institutional reason. - Have a single, opinionated leader with a lot of control over the program’s direction and actions. - Involve more than one person. - Are finite. - Have a clear goal. That is, when you get to the end of a finite amount of time, you know whether you’ve hit your goal or not. Coordinated research programs can - Very often create public goods. This is generally a big reason that they don’t make a good startup. - Very often involve work being done in more than one organization. They exist on a spectrum between fully *externalized*, where the coordinating organization does none of the hands-on research, work to fully *internalized* where the coordinating organization does all of the hands-on research. In their platonic ideals, ARPA programs are an example of the former and FROs are an example of the latter. Within these broad characteristics, this playbook is primarily meant for people creating coordinated research programs with a few more characteristics: - They are meant to have broad impact in the world, whether it’s through creating new knowledge, tools, or processes. Coordinated research programs can try to unlock esoteric new knowledge but the way you go about starting and running them is quite different. - Have a timeline of ~five years and a budget on the order of ~$10-50m. The way that one plans for drastically more or less time than five years or raises and deploys drastically more or less than ~$10m are just very different. Some examples of coordinated research programs within the scope of this playbook: - ARPA programs. - Focused Research Organizations. - Many of the programs at foundations like The Gates Foundation or The Sloan Foundation. - [Carbon to Sea](https://www.carbontosea.org/) Some examples of coordinated research programs at the edge of or beyond the scope of this playbook: - LIGO - The Rockefeller Foundation’s Molecular Biology program Some examples of things that probably aren’t coordinated research programs: - Most NSF and NIH programs. Importantly, program officers are often beholden to committees about what they fund and have little ability to change the program’s direction once it has started. - Most institutes (like the Alan Institute, the Arc Institute, HHMI etc). Importantly, institutes last an indefinite amount of time and often have broad rather than precise goals.