Coining new terms is fun (if a bit silly), but I do think there is something important to the concept of the Omnifabric - even if the actual term becomes something more sensible.
We’re moving beyond purely static interfaces, into a realm of computer <> human interaction that will be so much more fluid and dynamic that the lines between the two will blur ever further.
AI Summary
We’ve spent decades refining interfaces—from punch cards to terminals to GUIs—but AI is rendering this entire paradigm outdated. The Omnifabric concept suggests we’re moving from static, fixed windows into computers toward dynamic, intelligent interactions that adapt to our needs. Google’s current advantage illustrates this shift: their real power isn’t just Gemini’s capabilities, but how seamlessly it integrates across Workspace, allowing users to synthesize emails, documents, and strategies in natural language. The underlying data remains constant—like embeddings in vector space—but how we access and transform it becomes infinitely malleable. We can turn the same information into podcasts, videos, mind maps, or whatever format serves our immediate purpose. The limitation isn’t technical anymore; it’s our imagination about what’s possible when the interface adapts to us rather than the other way around.
Chapters
(00:00:00) - Introducing the Omnifabric Concept
(00:01:19) - Moving Beyond the Static Interface
(00:03:03) - Why Google Is Winning on Distribution
(00:04:59) - Enterprise Security and Single Sign-On Advantages
(00:05:15) - Notebook.lm and Dynamic Content Transformation
(00:07:01) - Understanding Multidimensional Vector Space
(00:09:11) - Staying Relevant in a Rapidly Changing Landscape
(00:11:47) - From Static Code to Dynamic Computation
(00:13:13) - AI Eating Traditional Codebases
(00:14:10) - Voice as an Emerging Modality
(00:15:21) - The Browser Limitation and OS-Level Integration
(00:16:34) - We’re Just at the Beginning
Key Moments
The Interface Is Fixed, Reality Is Not (00:02:02) - “They are still nonetheless, in a sense, quite static. They’re defined. They are a fixed-sized window into the underlying computer to interact with.”
Distribution Beats Product, Again (00:03:12) - “There is an element of this with OpenAI versus Google at the moment, whereby setting aside the performance of Gemini as a model, the benchmarks, the advances of Nano Banana, etc. It’s actually the distribution that they have is what I’m noticing as being a real game changer.”
The Enterprise Security Nightmare (00:04:20) - “The idea of everyone having different personal logins to ChatGPT or all these services, potentially asking for a recipe and then sharing some confidential business information on the next prompt. This is a bit of a nightmare.”
Same Data, Infinite Forms (00:06:42) - “The underlying information is the same right it’s all my Substack it’s everything I’ve written about over the last however many years but being able to essentially what I’ve been trying to hack on myself and actually Notebook LLM really helps with is I want to bring all this together.”
Your Imagination Is the Constraint (00:08:03) - “The main limitation on how I interact with this information, how I come to understand it and work through it, is as much my own imagination as it is the technical limitations of the tool.”
Malleable for Users, Not Developers (00:13:49) - “It’s something that is dynamic, it’s something that is malleable, changeable, not by the people creating, not by the developers or the people putting it out there, but rather those consuming it.”
Sharing is caring!
If you enjoyed this conversation, please consider sharing it with two friends who might find it useful.
