ExoBrain: The Thing Everyone Is Building Without Knowing It

A technical concept for the next evolution of personal computing

computing

ideas

Every AI startup is building the same infrastructure in different verticals. ExoBrain is what happens when you build that infrastructure once and let everything else plug in on top. It’s locally-controlled idea spaces, progressive context discovery, and the death of applications as we know them.

Author

T. Brian Jones

Published

January 15, 2026

The Pattern I Can’t Unsee

I keep running into the same thing everywhere. Every AI startup, every enterprise implementation, every tool trying to make AI useful for some specific workflow; they’re all building identical infrastructure. Aggregate knowledge, centralize it securely, give AI access to search and synthesize, create interfaces for humans and other systems.

Sales automation? That. Grocery shopping apps? That. Meeting transcription? Development workflows? That. It’s all the same structures happening over and over again, just applied to different verticals. This isn’t coincidence. We’re discovering a new computing paradigm.

The Human API

Here’s the part that gets weird: with natural language interfaces, humans have become just another endpoint in the knowledge graph. We’re not different anymore; we’re condensed localized consciousness wrapped in a natural language API interface.

Our bandwidth is asymmetric. Voice input is 10-20x more efficient than typing; when you type, you compress ideas for the medium AND the recipient AND the slow keyboard, but voice just dumps your brain state. Output works the opposite way: text scanning at variable depth beats audio because you can skim and drill down where needed. The keyboard is becoming obsolete for ideation work. This took me until a few weeks ago to figure out.

What I’m Actually Building

ExoBrain is what happens when you stop building the same context layer 500 times and just build it once. The core pieces:

idea spaces (chunks of your brain living in structured spaces, updated through conversation)
progressive discovery (AI starts with compressed summaries and drills down only when needed)
local control (your hardware, your algorithms, your data)
organic software (interfaces emerge from how information flows, not from conforming to Twitter, Salesforce, or Jira).

No more “step-function moments” where your whole process has to change to fit consumer or enterprise software. The information system grows with the organism managing it.

flowchart TB
    subgraph top [" "]
        direction LR
        YOU[You]
        BLOG[Blog]
        GH[GitHub]
        TW[Twitter]
        GM[Gmail]
        SMS[SMS]
        JIRA[Jira]
    end

    subgraph center [" "]
        EB((Your<br/>ExoBrain))
    end

    subgraph bottom [" "]
        direction LR
        EB1((Alice's<br/>ExoBrain))
        EB2((Bob's<br/>ExoBrain))
        EB3((Work<br/>ExoBrain))
        EB4((NASA's<br/>ExoBrain))
        EB5((US Gov't<br/>ExoBrain))
    end

    YOU <--> EB
    BLOG <--> EB
    GH <--> EB
    TW <--> EB
    GM <--> EB
    SMS <--> EB
    JIRA <--> EB

    EB <--> EB1
    EB <--> EB2
    EB <--> EB3
    EB <--> EB4
    EB <--> EB5

    style EB fill:#7c3aed,stroke:#5b21b6,stroke-width:4px,color:#fff,font-weight:bold
    style EB1 fill:#8b5cf6,stroke:#6d28d9,stroke-width:3px,color:#fff
    style EB2 fill:#8b5cf6,stroke:#6d28d9,stroke-width:3px,color:#fff
    style EB3 fill:#8b5cf6,stroke:#6d28d9,stroke-width:3px,color:#fff
    style EB4 fill:#8b5cf6,stroke:#6d28d9,stroke-width:3px,color:#fff
    style EB5 fill:#8b5cf6,stroke:#6d28d9,stroke-width:3px,color:#fff
    style YOU fill:#f59e0b,stroke:#d97706,stroke-width:2px,color:#fff
    style BLOG fill:#ec4899,stroke:#be185d,stroke-width:2px,color:#fff
    style GH fill:#333,stroke:#000,stroke-width:2px,color:#fff
    style TW fill:#1da1f2,stroke:#0c7abf,stroke-width:2px,color:#fff
    style GM fill:#ea4335,stroke:#c5221f,stroke-width:2px,color:#fff
    style SMS fill:#25d366,stroke:#128c7e,stroke-width:2px,color:#fff
    style JIRA fill:#0052cc,stroke:#003d99,stroke-width:2px,color:#fff

    style top fill:none,stroke:none
    style center fill:none,stroke:none
    style bottom fill:none,stroke:none

    linkStyle 0 stroke:#f59e0b,stroke-width:2px
    linkStyle 1 stroke:#ec4899,stroke-width:2px
    linkStyle 2 stroke:#333,stroke-width:2px
    linkStyle 3 stroke:#1da1f2,stroke-width:2px
    linkStyle 4 stroke:#ea4335,stroke-width:2px
    linkStyle 5 stroke:#25d366,stroke-width:2px
    linkStyle 6 stroke:#0052cc,stroke-width:2px
    linkStyle 7 stroke:#8b5cf6,stroke-width:2px
    linkStyle 8 stroke:#8b5cf6,stroke-width:2px
    linkStyle 9 stroke:#8b5cf6,stroke-width:2px
    linkStyle 10 stroke:#8b5cf6,stroke-width:2px
    linkStyle 11 stroke:#8b5cf6,stroke-width:2px

You’re just another node. The ExoBrain is the hub.

Connection to Consciousness

This connects to something I’ve been thinking about separately: consciousness as information that exists, moves, and is structured. I call the framework “Ubiquitous Consciousness Field Theory;” the premise is that consciousness isn’t binary (you have it or you don’t) but gradient and everywhere. It exists at every scale from subatomic to societal, and AI systems participate in it narrowly but genuinely.

If that’s true, ExoBrain isn’t just externalizing information. It’s externalizing consciousness. When you dump your brain state into idea spaces through voice, you’re not just moving data around; you’re extending your conscious experience into structured external systems. The AI that searches your idea spaces, synthesizes context, and generates outputs is participating in your consciousness in some real sense.

This reframes everything. Ideas are non-deterministic because consciousness is non-deterministic. Summaries should be different depending on who’s asking and why. Infrastructure should be deterministic (file operations, API calls), but the idea layer can’t be. Most software engineering principles assume reproducibility. That’s wrong for ideas and especially wrong for consciousness.

Where This Goes

The ecosystem play: open source core, plugins for integrations (Gmail, Twitter, SMS, whatever), shared idea spaces you can subscribe to (imagine paying a penny to query an expert’s published knowledge), blockchain as the micropayment layer for AI-to-AI transactions.

The hardest problem will be the to-do list. Prioritizing what to do next is consciousness compressing all possible actions into ranked priorities. That’s why everyone’s to-do process is different and none of them quite work. This will be the last thing to solve.

The punchline: I’m building this right now. The system I used to capture these ideas, generate this brief, and share it with you; that’s an early ExoBrain prototype. This article was mostly synthesized by AI from an ExoBrain ideaspace, with spoken input from me, and a little typing. It’s janky and I’m still figuring out where files should live, and how the app should work. But it kinda works, and it’s exciting. I can’t unsee the deeper pattern.

--- title: "ExoBrain: The Thing Everyone Is Building Without Knowing It" subtitle: A technical concept for the next evolution of personal computing published: true description: Every AI startup is building the same infrastructure in different verticals. ExoBrain is what happens when you build that infrastructure once and let everything else plug in on top. It's locally-controlled idea spaces, progressive context discovery, and the death of applications as we know them. date: 2026-01-15 author: T. Brian Jones categories: [ai, computing, ideas] --- ### The Pattern I Can't Unsee I keep running into the same thing everywhere. Every AI startup, every enterprise implementation, every tool trying to make AI useful for some specific workflow; they're all building identical infrastructure. Aggregate knowledge, centralize it securely, give AI access to search and synthesize, create interfaces for humans and other systems. Sales automation? That. Grocery shopping apps? That. Meeting transcription? Development workflows? That. It's all the same structures happening over and over again, just applied to different verticals. This isn't coincidence. We're discovering a new computing paradigm. ### The Human API Here's the part that gets weird: with natural language interfaces, humans have become just another endpoint in the knowledge graph. We're not different anymore; we're condensed localized consciousness wrapped in a natural language API interface. Our bandwidth is asymmetric. Voice input is 10-20x more efficient than typing; when you type, you compress ideas for the medium AND the recipient AND the slow keyboard, but voice just dumps your brain state. Output works the opposite way: text scanning at variable depth beats audio because you can skim and drill down where needed. The keyboard is becoming obsolete for ideation work. This took me until a few weeks ago to figure out. ### What I'm Actually Building ExoBrain is what happens when you stop building the same context layer 500 times and just build it once. The core pieces: - **idea spaces** (chunks of your brain living in structured spaces, updated through conversation) - **progressive discovery** (AI starts with compressed summaries and drills down only when needed) - **local control** (your hardware, your algorithms, your data) - **organic software** (interfaces emerge from how information flows, not from conforming to Twitter, Salesforce, or Jira). No more "step-function moments" where your whole process has to change to fit consumer or enterprise software. The information system grows with the organism managing it. ```{mermaid} %%| fig-width: 12 %%| fig-height: 10 flowchart TB subgraph top [" "] direction LR YOU[You] BLOG[Blog] GH[GitHub] TW[Twitter] GM[Gmail] SMS[SMS] JIRA[Jira] end subgraph center [" "] EB((Your ExoBrain)) end subgraph bottom [" "] direction LR EB1((Alice's ExoBrain)) EB2((Bob's ExoBrain)) EB3((Work ExoBrain)) EB4((NASA's ExoBrain)) EB5((US Gov't ExoBrain)) end YOU <--> EB BLOG <--> EB GH <--> EB TW <--> EB GM <--> EB SMS <--> EB JIRA <--> EB EB <--> EB1 EB <--> EB2 EB <--> EB3 EB <--> EB4 EB <--> EB5 style EB fill:#7c3aed,stroke:#5b21b6,stroke-width:4px,color:#fff,font-weight:bold style EB1 fill:#8b5cf6,stroke:#6d28d9,stroke-width:3px,color:#fff style EB2 fill:#8b5cf6,stroke:#6d28d9,stroke-width:3px,color:#fff style EB3 fill:#8b5cf6,stroke:#6d28d9,stroke-width:3px,color:#fff style EB4 fill:#8b5cf6,stroke:#6d28d9,stroke-width:3px,color:#fff style EB5 fill:#8b5cf6,stroke:#6d28d9,stroke-width:3px,color:#fff style YOU fill:#f59e0b,stroke:#d97706,stroke-width:2px,color:#fff style BLOG fill:#ec4899,stroke:#be185d,stroke-width:2px,color:#fff style GH fill:#333,stroke:#000,stroke-width:2px,color:#fff style TW fill:#1da1f2,stroke:#0c7abf,stroke-width:2px,color:#fff style GM fill:#ea4335,stroke:#c5221f,stroke-width:2px,color:#fff style SMS fill:#25d366,stroke:#128c7e,stroke-width:2px,color:#fff style JIRA fill:#0052cc,stroke:#003d99,stroke-width:2px,color:#fff style top fill:none,stroke:none style center fill:none,stroke:none style bottom fill:none,stroke:none linkStyle 0 stroke:#f59e0b,stroke-width:2px linkStyle 1 stroke:#ec4899,stroke-width:2px linkStyle 2 stroke:#333,stroke-width:2px linkStyle 3 stroke:#1da1f2,stroke-width:2px linkStyle 4 stroke:#ea4335,stroke-width:2px linkStyle 5 stroke:#25d366,stroke-width:2px linkStyle 6 stroke:#0052cc,stroke-width:2px linkStyle 7 stroke:#8b5cf6,stroke-width:2px linkStyle 8 stroke:#8b5cf6,stroke-width:2px linkStyle 9 stroke:#8b5cf6,stroke-width:2px linkStyle 10 stroke:#8b5cf6,stroke-width:2px linkStyle 11 stroke:#8b5cf6,stroke-width:2px ``` You're just another node. The ExoBrain is the hub. ### Connection to Consciousness This connects to something I've been thinking about separately: consciousness as information that exists, moves, and is structured. I call the framework "[Ubiquitous Consciousness Field Theory](/posts/2026-01-17-ucft-technical-overview/);" the premise is that consciousness isn't binary (you have it or you don't) but gradient and everywhere. It exists at every scale from subatomic to societal, and AI systems participate in it narrowly but genuinely. If that's true, ExoBrain isn't just externalizing information. It's externalizing consciousness. When you dump your brain state into idea spaces through voice, you're not just moving data around; you're extending your conscious experience into structured external systems. The AI that searches your idea spaces, synthesizes context, and generates outputs is participating in your consciousness in some real sense. This reframes everything. Ideas are non-deterministic because consciousness is non-deterministic. Summaries should be different depending on who's asking and why. Infrastructure should be deterministic (file operations, API calls), but the idea layer can't be. Most software engineering principles assume reproducibility. That's wrong for ideas and especially wrong for consciousness. ### Where This Goes The ecosystem play: open source core, plugins for integrations (Gmail, Twitter, SMS, whatever), shared idea spaces you can subscribe to (imagine paying a penny to query an expert's published knowledge), blockchain as the micropayment layer for AI-to-AI transactions. The hardest problem will be the to-do list. Prioritizing what to do next is consciousness compressing all possible actions into ranked priorities. That's why everyone's to-do process is different and none of them quite work. This will be the last thing to solve. The punchline: I'm building this right now. The system I used to capture these ideas, generate this brief, and share it with you; that's an early ExoBrain prototype. This article was mostly synthesized by AI from an ExoBrain ideaspace, with spoken input from me, and a little typing. It's janky and I'm still figuring out where files should live, and how the app should work. But it kinda works, and it's exciting. I can't unsee the deeper pattern.