Common Fabric and the Future of Resonant Computing
Alex Komoroske introduces a new infrastructure designed to prioritize human-centric, private agentic computing.
1Redefining Digital Experience
Alex Komoroske, CEO of Common Fabric, argues that modern technology is optimized for hollow, extractive experiences. In a conversation with Tim O’Reilly, he proposed the Resonant Computing Manifesto. This framework aims to shift the AI transition toward human-centric norms, prioritizing user nourishment over engagement-based metrics that currently dominate the digital landscape.
2Principles of Resonant Technology
The manifesto outlines five core principles: private, dedicated, plural, adaptable, and prosocial. These pillars seek to prevent central entities from holding excessive power while ensuring users remain the stewards of their own data. By fostering collective meaning rather than individual isolation, the model challenges the prevailing industry focus on extraction.
3Contextual Integrity in Privacy
Alex defines privacy through Helen Nissenbaum’s concept of contextual integrity, where data usage aligns with user intentions. Unlike traditional hard privacy, which acts like bars on a window, this approach focuses on control. It ensures that data remains useful to the owner without creating nasty surprises or harmful, unauthorized secondary uses.
4The Problem with Silos
Current software relies on inadequate mechanisms like complex terms of service and permission dialogs. Alex describes these as responsibility-laundering tools that fail to protect users. Because data is viral and easily copied, existing silos prevent meaningful integration, leaving user data stagnant and disconnected from the services that could actually benefit them.
5Metadata Sidecars for Data
Common Fabric introduces a structural change where every piece of data carries a metadata sidecar. This includes cryptographic provenance, tracing the history of computations, and specific policies governing future access. This architecture ensures that rules travel with the data, allowing for secure interactions without needing a globally consistent, financialized ledger system.
6The Core Reaction Chamber
An open-source runtime acts as the core reaction chamber, enforcing policies on untrusted code. It evaluates proposed computations against defined rules before execution. If a request risks leaking confidential information, the runtime stops the process. This allows users to run code safely in the cloud using hardware-isolated trusted execution environments.
7Patterns as Untrusted Code
Common Fabric utilizes patterns, which are small chunks of untrusted TypeScript code. These patterns can perform computations, produce derived data, and request subpatterns. Because the system enforces strict policy boundaries, users do not need to trust the pattern author. The data itself dictates what operations are permitted, maintaining security throughout the process.
8Proactive Data Requests
Patterns can proactively request data to improve functionality, such as meal planning based on a user's favorite foods. By utilizing LLMs to distill information from private sources like Gmail, the system generates useful insights. Crucially, the pattern author never sees the raw data, ensuring confidentiality remains intact during these automated, helpful interactions.
9Information Flow and Taint
The system employs information-flow control, using the concept of taint to track restrictions on data. When data touches confidential information, it inherits those constraints. The compiler untangles these flows, ensuring that only necessary handles are passed. This prevents sensitive tokens from being exposed, mirroring the safety guarantees found in Rust’s borrow checker.
10LLMs as System Architects
While human programmers might find strict taint checking tedious, LLMs handle these constraints efficiently. They can iterate on code until it compiles, ensuring structural safety. This synergy allows Common Fabric to run patterns speculatively and proactively. If the software compiles, the system guarantees that the data flows are structurally safe for users.
11Social Sifting of Patterns
Inspired by Christopher Alexander’s A Pattern Language, the platform expects a mix of high and low-quality patterns. Alex suggests that social sifting will naturally surface the best tools, similar to how viral content rises on platforms like TikTok. This creates a marketplace where popular, effective patterns become the standard for users.
12Optimism for the Web
Tim O’Reilly expressed renewed optimism regarding the future of the web after reviewing Common Fabric’s architecture. By moving beyond engagement-maxxing and extractive business models, the project offers a path toward more human-centric technology. The focus remains on building a digital home where agents and data work together for the user's benefit.
Original article: O'Reilly Radar — Common Fabric and the Future of Resonant Computing