Our Model
Our model is meant to bridge the “valley of death” by riffing on the things that made Bell Labs, pre-1972 ARPA, and Xerox PARC great without slavishly trying to copy 20th century organizations in the 21st century.
At the end of the day, our MO is to do what needs to be done to turn big-if-true technologies from a weird thing that seems like it will never happen into an inevitability.
Obviously, that’s a bit vague, so we base most of our work around a portfolio of five-year research programs run by program leads with an ambitious technology vision.
These programs go through four phases:
1. Incubate
We first find a program lead with a vision and skills to build a big-if-true technology that doesn’t have a home in other institutions. (Or they find us!) We then work with them to create a technology roadmap that identifies a definite plan for how the technology can change the world, what needs to happen in the next five years to unlock it, and the biggest questions or risks to this plan.
2. Derisk
The program lead does fast, iterative work to directly address the biggest questions and risks to both know whether we should proceed with the program and to start rallying support for the technology.
3. Build
During the bulk of the program, we actually build the technology. The work here depends heavily on the technology and how we expect it to affect the world: it could involve putting together a team to build everything in-house, working with external partners on several parallel projects, or something in between.
4. Transition
New technology can only be useful if it actually gets out into the world. The final phase of our programs is devoted to making sure this happens in one of several ways (that again depend on the technology and how we expect it to affect the world) — open sourcing a dataset or design, licensing the technology to one or more organizations that can manufacture it at scale, spinning out a nonprofit that carries the technology forward, or spinning out a startup if that makes the most sense.
Infused through all of this is the idea that we need tight feedback loops with the real world, minimal bureaucracy, and laser focus on the idea that success means making something useful.
Creating Public Goods
We are embracing the nature of speculative technology research to end up as a public good and using a permissive licensing stack. It’s incredibly hard to capture the value of exploratory technology research and trying to capture that value often hamstrings the work itself. Imagine if Bell Labs had tried to keep the transistor to themselves: it might have been used for nothing except under-sea signal amplification and Moore’s law may have never been started.
Private Organizations have more institutional Flexibility
We’re a private organization, despite the fact that many view public-goods-oriented research as the role of government. Government research is important, but faces a fundamental tension: we want representative governments to be responsible but in order to do great research you need to be a bit irresponsible (we do try to be as responsible as possible!) A private organization can move faster than the government and can innovate in ways that government organizations cannot.
