_date: aug 13 2024_ --- [Jonas](https://x.com/mempirate) from [Chainbound](https://chainbound.io/) recently asked how I'd structure an r&d organization. I started doing the exercise and quickly realized it’s hard to structure an org in the abstract without knowing what its mission is. I see structure as a downstream decision for what the mission is. The same way you first decide you want to cross the Atlantic before deciding you need a boat! There are however some invariants to an r&d org that I think stick regardless of the overall mission: 1. thinking of innovation as a system, and optimizing the structure and incentives of the organization to maximize innovation, in line with its mission! - at Bell Labs that was the 3-unit division, and the many cultural choices they made to foster an open, collaborative, multi-disciplinary environment (the ‘open-door’ policies, the opportunities for scientists to enroll in university classes, the long corridors of Bell Labs, the amphitheater that anyone in the company could use to present their work) - see some personal notes on the history of Bell Labs here [[the idea factory, bell labs & the great age of american innovation by john gertner (book notes)]] 2. ensuring leadership are research-minded, maybe researchers/research engineers themselves who strongly believe in long-term research - this is to avoid being lead by people that are fixated on the short-term, compromising the research agenda, the creative capabilities of the org, and ultimately its impact - this is also to avoid being lead by individuals that do not understand how research works, and try to overfit on it performance metrics and expectations that will stifle it 3. consider all modes of funding: vc, token, grants, revenue. Funding is circumstantial and is about survival so there is no hard rule imo. None of these funding methods are mutually exclusive as well. But whichever funding route is adopted, it’s important to have an understanding that they can have serious implications on the nature of the organization and its capacity to innovate. - For eg, raising too much VC money at too high of a valuation too early on can steer the org towards short-termism in order to meet the expectations of investors, and can lead to the wrong company culture. - My preferred sustainability model generally is one where growth is at a measured pace, and where there are inherent feedback loops between the r&d org and the environment it is evolving in. This allows to ground the research agenda and allows for more ‘organic’ growth. Some generally useful links: - [https://www.aria.org.uk/how-were-working/](https://www.aria.org.uk/how-were-working/) — an example of how an r&d org funded by grants organizes (the program manager model), which I like a lot. it’s the [DARPA model](https://benjaminreinhardt.com/wddw) - see also [parpa-2-pager.pdf](https://benjaminreinhardt.com/parpa-2-pager.pdf) — PARPA (Private ARPA) is a new organization that aims to unlock robust technology that can open new frontiers both on earth and beyond. - [https://spec.tech/library/research-leaders-playbook](https://spec.tech/library/research-leaders-playbook) — this is a great resource for an example of how to set up coordinated research programs at different stages of r&d - https://gist.github.com/nicola/0a3001a64320b617c2f7703a95d5b3bb -- a fantastic resource for how Protocol Labs organized their r&d efforts - [https://www.michaeldempsey.me/blog/2024/07/16/the-venturification-of-research-capital-the-resulting-prisoners-dilemma/](https://www.michaeldempsey.me/blog/2024/07/16/the-venturification-of-research-capital-the-resulting-prisoners-dilemma/) — on the “venturification” of research capital resulting in orgs compromising their long-term research agenda to be more short-term - [https://stephango.com/vcware](https://stephango.com/vcware) — on being 100% ‘user-supported’, adapting this mentality to an r&d org is what I’m thinking about when I say ‘feedback loops’ - [https://trohan.com/2023/08/20/raise-less-build-more/](https://trohan.com/2023/08/20/raise-less-build-more/) — on the risks of raising too much VC money - different types of r&d orgs to consider: - (funding r&d via revenue): Google (with Google X, and Google Research), Microsoft Research, Nethermind - (funding r&d via VC/tokens): Flashbots, Protocol Labs, Ethereum Foundation - (funding r&d via grants): 0xParc, The Latest in DeFi Research, [Convergent Research](https://www.notion.so/wip-crypto-strategy-resources-e23900ef5b70408fba7a13f5c8871762?pvs=21) (as a funder of FROs), OpenAI - [The Idea Factory: Bell Labs and the Great Age of American Innovation by John Gertner 8](https://www.amazon.fr/Idea-Factory-Great-American-Innovation/dp/0143122797) – story of Bell Labs and how they nailed r&d for multiple decades. - [https://benjaminreinhardt.com/innovation-org-principles/](https://benjaminreinhardt.com/innovation-org-principles/) Principles For Innovation Orgs - [https://benjaminreinhardt.com/grants](https://benjaminreinhardt.com/grants) On Grants - [Studies on Slack](https://slatestarcodex.com/2020/05/12/studies-on-slack/), Slate Star Codex, on the importance of slack to make organizations more resilient, humane and capable of dealing with unexpected things