Announcing the BDFX project
A bit of a stealth announcement/preview about a new direction for PI based on recent discussions between @Rafael Fernández and myself, and things we’re planning and building for PI which you’ll start to see early next year. There are several moving parts but the first thing we want to offer a peek of is a project called BDFX.

What’s BDFX? That will take a spot of context.
PI is evolving from a grants program (SoP) into a place for risk-sharing partners who participate in collaborative opportunities that require contribution of leadership and risk-taking energy, not just labor, research, or service energy in exchange for “funding.” To this end we need suitable systems that enable more people to lead, lower the activation energy for leadership contribution, and underwrite risk-taking to some extent.
So this is increasingly going to be a place where you come to learn to fish and find fish, not be fed fish.
Scholar-entrepreneurs of protocolization
But unlike purely entrepreneurial spaces like startup incubators, there is also a public mission here — to build the knowledge commons of Protocol Studies and drive protocolization in the world, at every level from basic research through practical applications and pedagogy. That makes it more like a university. So to participate fully in PI you need to do research, apply it, teach others, and contribute with a service mindset too. And do all four not just as a contributor but as a leader, creating more value than you yourself consume. Something like a scholar-entrepreneur. We’ve been gradually headed this way for almost two years now. Our research group leaders/hosts are the early archetypes. We want to level up that type of role to become financially load-bearing, and build systems that help more people get there.
In other words this is about being more than a solopreneur or indie, a mode which the internet made possible, but which didn’t really scale, and just barely worked at n=1 scale. Most people in the “indie creator” economy are still highly precarious, and barely manage to make ends meet. But infrastructure can change sustainability. Substack and YouTube made a lot more people less precarious and get closer to n=1 self-sustainability for example.
But with AI you can be much more ambitious. You can aim for more than n=1 self-sustainability.
The AI-enabled Sawai leader
To be a leader of the type PI needs, you don’t need to be Steve Jobs. You just have to have a mindset of creating (say) 25% more hard value than you need for yourself. That means money and other hard resources, not just ideas (which we still need too!). That 25% can go towards helping bootstrap someone earlier in their trajectory, build shared infrastructure and resources etc.
In the old Mughal empire, the title “Sawai” was given to people who proved to be particularly capable senior leaders, where sawai comes from the Hindi sawa, meaning “125%”. We don’t believe in “special people” or “chosen ones” around here, but we think anyone can step up to “sawai” level at least some of the time, especially with AI help, and we need more people to do so. All our programming in 2027 will be aimed at that. PI as a pipeline for AI-powered part-time sawai people. We need sawai leaders more than we need money. If we solve the sawai leadership problem, the money problem will solve itself.
125%, some of the time
The good news is that you don’t have to be at 125% all the time, in all your participation. We want to build infrastructure that allows you to be **Benevolent Dictator for X, ** where X is some bounded scope and time. The idea is based on the open-source community leadership idea of BDFL — Benevolent Dictator for Life. In some areas, some of the time, you can BDFX. In other areas, you can follow and benefit from other people BDFXing. And in some areas you can just free-ride on the efforts of others.
I (Venkat) wrote about the basic concept of BDFxing in 2024, but with AI, we can finally build systems to support the pattern more powerfully.
So we’re building BDFX—a leadership OS, built on the premise of activating more people to being partners who participate in leadership — some of the time, in some areas. Here is the initial spec for the first version. This is not a manifesto. It’s a spec for a software system we’re building to allow everyone willing to learn to BDFX and “go Sawai.” You can read it here (or point your LLM at it and ask it to explain the idea). The agents are already busy building it. The system will be rolled out in January.
PS, thanks to buddies at the yak collective where this idea first took shape years ago, and is already practiced on a smaller, more informal scale.