Reflyx: Early Preview
My research on the future of computing is intertwined with my active work on our next generation operating system: Reflyx OS. Beyond theoritical and arid articles on why we need a new computing paradigm and the inherent problems of the existing one (see previous articles in this series!), I am working on creating an early prototype to demonstrate my ideas but even more, to myself learn by applying my principles to real problems.
A new paradigm cannot be successful without a reference implementation, some kind of Mosaic equivalent for the Web. We need a way to make this concrete, to use it and learn from our usage, to inherently change our vision of software. My work on this early prototype is also to validate my assumption on how to migrate from our current paradigm to the next, without losing too much of our investments in the current systems. We also need to make sure that we don’t offer a very low user experience (like the initial Web compared to GUI desktop apps at the time). We want this transition to be natural, gradually extending then replacing our current approach with a more fluid one.
High-Level Overview
The following diagram is a high-level map of the major components of the Reflyx OS. Everything is centered on the actor space, which is associated with a strong cryptographic identity is completely private, usable ONLY by this identity. This supports two principles: strict content uniqueness and actor-centric private view of the computing universe.
All the content produced by the actor is stored internally in its space and is no accessible to any other actors by default. Content must be published and explicitly subscribed in order to be introduced into another space. This is true for all actors, not just humans.
Humans actors need an additional layer (represented by the HMI box) to interact with the data as we don’t have native connectivity (yet!). In our current prototype, the HMI is called Cockpit, and it can be visualized exactly like that: a set of tools for a human to perform task in its space,to pilot the actor digital avatar, a thin digital layer to bring us into the computing universe as a peer actor, not as a user.
Each actor identity is managed by an ID service, which can be understood like a blockchain wallet, a vault holding private keys, settings, preferences and used to securely sign transactions between actors. We see a separate hardware device (phone-like, watch or implant) that would be discovered by the cockpit and use to seamlessly prove human actors’ identity.
You can find a very early ID service in our github. This service is just a placeholder for a real identity service, currently used to discover the right API and support the creation of a more formal open specification later on.
Current Focus: Commoditized Computing
One of most radical concept of our new paradigm is the separation of computing from computers. As explained in my previous articles, we need to change our perception of software from an extension mechanism for generic computers to an automation, communication and processing extension to our human capacity. For this vision to succeed, we need to move to a completely commoditized computing grid based on a very different paradigm than the networked physical computer we call the cloud today. We need to get rid of the remote and island nature of computers, each one focused exclusively on its own small hardware realm, and move to fluent ocean of generic and specialised computing capabilities, subscribed under various terms by the different actors (not just humans).
Our current work is how, through the cockpit, a human can interact with this computing grid and start leveraging these resources to support its space and interact with other actors, humans, organizations, objects, systems or AI agents.
As you can see in this early screenshot of the Cockpit prototype, we’re separating concerns around the human. Decentralized ID devices based on similar cryptographic technology than Ethereum or Bitcoin. This ID is detected by our prototype Cockpit, which for now is developed with Electron and run on any of the popular hardware-centric operating systems. The 3rd aspect, is the computing grid, which is detected by the cockpit and is represented with a few examples here.
Local Computing
In Reflyx, you need a computing subscription to exploit the resources of any computing provider, including your laptop! You could use a VR Cockpit or implants with very different UX and integrated computing capabilities and not use the local computing resources at all. Hardware is commoditized, so storage, runtime (CPUs, RAM) and other specialized services can be acquired and used anywhere on the grid.
For our prototype, it’s very convenient to get started on the local computing resources. It’s really not that natural to treat all computing resources as separated, even worst when we still run in hosted mode on devices not designed for Reflyx!
Reflyx Personal Computing Hub Concept
This cockpit has detected a Personal Computing Hub (PCH) on the local proximity grid (LAN). We truly believe that this concept of proximity grid gateway will be a powerful enabler for human centric computing and for the streamlining the computing industry as a whole. Because Reflyx is not based on the concept of networked isolated computers, there won’t be any network providers business model. Instead, we’ll be talking about computing providers, installing hubs in your home, similar to your electrical panel for electricity. This hubs will provide computing resources, either locally on the physical device or as a gateway to grid-wide resources (analog to cloud services today, but way more powerful and fluid)
iPhone Computing Resources
We also work on integrating mobile devices as computing devices in the Reflyx world. In our concept, mobile devices could host the 3 major components for humans, act as an ID token, a cockpit for human interaction and a proximity computing resource provider connected to the grid. With these 3 stereotypes in mind, we expect mobile devices to evolved into something very different as Reflyx get traction. We’ll write more about this vision in future articles!
Contributions!
As you’ve surely realized by now, this is a VERY ambitious project, going against the industry tide and everything we take for granted about software and hardware. But like any challenging journey, we expect it to be very rewarding and a major step forward for humanity quest to emancipation and our ultimate, inevitable, merge with what we call artificial intelligence today.
The best way to stay in touch is to follow us on X or Github for now. We want to use our own platform for most of our internal systems, so it will take a few months before we can get anyone to directly work with us.
Reflyx will be managed like a decentralized foundation (DAO), with gradual decision decentralization as more and more brilliant people grasp the basic concepts and help us bring this to heights beyond our wildest dreams!
We have a temporary WIKI service to aggregate the knowledge, specifications and all the research papers supporting our computing paradigm. Access is very limited. If you feel you can contribute significantly at this early stage, don’t hesitate to contact us to discuss about it.
