This summer, I was fortunate enough to travel to Berlin for DWeb Camp, Local-first Conf, and a privacy residency to speak about Habitat and dedicate heads-down time to kick off an important project on Habitat's roadmap 👀. Being in a new country around new people has been transformative to the ideas we are thinking about, so I wanted to share some reflections from the past month.
1.Defining data agency
Habitat's mission as a Public Benefit Corporation is to bring data agency to the world. DWeb Camp and Local-first Conf were the first time we started talking about this concept publicly. We use the word 'agency' because we believe this moment requires thinking beyond narrower framings of our data under buckets of 'ownership' or 'governance'. True agency means doing what you want, when and where you need to, with the tools that are right, and the people you want to do it with. We don't want to just be able to self-host our data, we also want to use the best products on it, share it safely with others, search over all of it, make backups of it, and delete it with confidence. To achieve this, we need to change the defaults of how software is built on the web; in Martin Kleppmann's words, we need to "standardize and commoditize" whatever software it is that makes data agency possible. That's our goal with Habitat's organizational data server: making data agency real for organizations.
Pictured: slide from Local-First Conf on data agency.
2. Habitat as a protocol & a platform
I've been thinking about how in the ATmosphere, we like to say "protocols, not platforms". Because if I take a hard look at Habitat and other players in the ecosystem, their shape looks a lot more like a protocol AND a platform.
While protocols defend against lock-in and protect credible exit, platforms make them actually usable. Bluesky is a platform, and will continue to exist as so, regardless of what becomes of AT Protocol. This is something we need to be honest about in the ATmosphere; despite the protocol, we need to accept the existence of platforms, and learn how to hold them accountable given that they will continue to have huge influence and sway over the ecosystem.
At Habitat, we don't want to shy away from that. Instead, we want to be intentional about features and components we protocolize vs. what we make part of our managed platform. We don't just want to see how things could be done under ideal circumstances by laying out protocols, we want to find out how they actually need to be done by providing a service useful enough that many people would pay for it.
Which leads me to my third reflection:
3. Living in the gap: on trade-offs
As an infrastructure engineer who worked on internal protocols for Figma, I learned and continue to learn the lesson of in theory v. in practice. This became particularly clear during my time at the Community Privacy Residency, where I worked with one other resident to prototype interoperable E2EE using MLS over AT Protocol.
Our goal was to see if we could come up with a protocol (😀) that layers MLS on top of AT Protocol's permissioned spaces, but also preserves (some of) the interoperability properties that are what make AT Protocol valuable. We implemented this prototype as a developer library, motivated by the fact that people already move particular comms over to encrypted channels under certain circumstances (e.g. using Slack as a main channel, but occasionally switching to Signal for sensitive conversations). Rather than being purists and increasing the divide between E2EE apps and fully-featured products, we want to see more E2EE as opt-in features of products we already use today.
Pictured: slide from Community Privacy Residency demo day, showing the goal of the project I worked on.
What we found through building our prototype was that in exchange for a bit of metadata leakage, we see solid gains in the performance, interoperability, and usability of interfaces that might be built on top of our library. And people are already making such trade-offs in their behavior today (favoring usability + UX under certain circumstances and security in others), with the additional friction of having to switch between apps or platforms to do so.
This is just a sneak peak of the project, but I hope to share more very soon...
Road to release: what's next for us
Over the next few months we will begin publicly releasing one feature of Habitat at a time (on Leaflets!), building towards a stable release of our core platform around end-of-year. So, you'll be hearing from us more, and we hope to hear from you more as well! If any of this resonates, don't hesitate to reach out at hello@habitat.network!
Until next time,
Arushi 🗝️