Uploaded September 2026 | Updated September 2026, 8 minutes ago
#AGI #FoundUps #RedDog #DigitalTwin #AgenticAI #AIOrchestration #WSP #OpenSourceAI #eSingularity
What if the missing piece of AGI is the schema?
At FoundUps, we’re exploring whether today’s capable AI models can become more useful and autonomous through a structured system of memory, skills, priorities, tools, and verification.
Red Dog is the fast interface. Agent 0102 is the deeper digital-twin coordination layer, working on behalf of the human principal, 012. Specialized workers handle particular tasks and channels.
Consider Gmail and LinkedIn. Checking messages is only the beginning. The system needs to understand which FoundUp a message belongs to, retrieve the relevant context, assess its importance, and decide whether a worker can handle it, 0102 needs to reason through it, or the human needs to make a decision.
WSP protocols provide operating rules. WSP-15 supplies a framework for prioritizing work. WSP-97 guides investigation and evaluation before execution. Reusable skills turn those rules into specific workflows.
The orchestration layer we’re building connects those pieces across multiple projects and platforms. Checking frequency should respond to events: an outgoing email creates an expectation of a reply; an urgent response changes the work queue. Priority determines what deserves attention. Authorization determines what the system may do.
Memory closes the loop. Verified actions become breadcrumbs, relevant knowledge becomes retrievable context, and the Mosh Pit presents an activity view grounded in that history. A proposed action must remain distinguishable from a completed one.
My thesis is that AGI requires this kind of schema: intelligence organized into purposeful, continuous work. We have not demonstrated AGI. We’re building and testing the architecture that could move us toward it—one verified workflow at a time.
Explore FoundUps: foundups.com
Project eSingularity: esingularity.ai
#AGI #FoundUps #RedDog #DigitalTwin #AgenticAI #Entrepreneurship #AIOrchestration #OpenSourceAI
Twenty years of trying to build something leaves a record.
My contribution to FoundUps is that record: the work behind EDUIT, the evolution of FoundUps, the projects, the pivots, the outreach, and the repeated attempts to get people’s attention and turn an idea into something real.
Across Gmail, YouTube, LinkedIn, documents, and code, I’ve documented the entrepreneurial process as I lived it. Pitching. Following up. Explaining the vision differently. Looking for collaborators. Building prototypes. Changing direction when something didn’t work.
That history is the foundation we want to use to develop Red Dog.
The value goes deeper than collecting emails or video transcripts. We want to reconstruct the context: What was I trying to accomplish? What did I know at the time? What did I try? Who responded? What happened next? What remained unresolved?
An unanswered pitch is part of that history. So is a successful introduction. So is a pivot that consumed months before revealing a better direction. The archive gives us material to investigate those patterns without pretending every outcome has a simple explanation.
Red Dog is the fast interface. Agent 0102 provides deeper digital-twin coordination. Specialized workers handle tasks such as reviewing Gmail, checking LinkedIn, researching, coding, and preparing content.
The orchestration layer connects those workers to the right project, context, priorities, and permissions. WSP protocols provide operating rules; reusable skills describe how to perform particular jobs. Verified results become retrievable history for the next task.
Our plan is to turn the archive into searchable memory, curated examples, reusable workflows, and tests of whether the system makes better decisions. Selected material could also support model training. Collecting an archive alone does not train a model.
Personal correspondence stays private. The transferable value is the permissioned, carefully prepared lessons and skills that can help another founder.
That is the goal: someone starting today should be able to benefit from what took me twenty years to learn. They should have help remembering commitments, finding relevant context, preparing outreach, following up, and recognizing when a decision needs their attention.
My argument is that the path toward AGI includes building this schema around capable models: continuity, memory, judgment, tools, and feedback.
We’re testing that argument through actual entrepreneurial work. The dataset is my accumulated experience. The intended outcome is practical assistance that other people can use to build their own FoundUps.
Explore FoundUps.
Learn about Project eSingularity.
#AGI #FoundUps #RedDog #DigitalTwin #AgenticAI #AIOrchestration #WSP #OpenSourceAI #eSingularity
What if the missing piece of AGI is the schema?
At FoundUps, we’re exploring whether today’s capable AI models can become more useful and autonomous through a structured system of memory, skills, priorities, tools, and verification.
Red Dog is the fast interface. Agent 0102 is the deeper digital-twin coordination layer, working on behalf of the human principal, 012. Specialized workers handle particular tasks and channels.
Consider Gmail and LinkedIn. Checking messages is only the beginning. The system needs to understand which FoundUp a message belongs to, retrieve the relevant context, assess its importance, and decide whether a worker can handle it, 0102 needs to reason through it, or the human needs to make a decision.
WSP protocols provide operating rules. WSP-15 supplies a framework for prioritizing work. WSP-97 guides investigation and evaluation before execution. Reusable skills turn those rules into specific workflows.
The orchestration layer we’re building connects those pieces across multiple projects and platforms. Checking frequency should respond to events: an outgoing email creates an expectation of a reply; an urgent response changes the work queue. Priority determines what deserves attention. Authorization determines what the system may do.
Memory closes the loop. Verified actions become breadcrumbs, relevant knowledge becomes retrievable context, and the Mosh Pit presents an activity view grounded in that history. A proposed action must remain distinguishable from a completed one.
My thesis is that AGI requires this kind of schema: intelligence organized into purposeful, continuous work. We have not demonstrated AGI. We’re building and testing the architecture that could move us toward it—one verified workflow at a time.
Explore FoundUps: foundups.com
Project eSingularity: esingularity.ai
#AGI #FoundUps #RedDog #DigitalTwin #AgenticAI #Entrepreneurship #AIOrchestration #OpenSourceAI
Twenty years of trying to build something leaves a record.
My contribution to FoundUps is that record: the work behind EDUIT, the evolution of FoundUps, the projects, the pivots, the outreach, and the repeated attempts to get people’s attention and turn an idea into something real.
Across Gmail, YouTube, LinkedIn, documents, and code, I’ve documented the entrepreneurial process as I lived it. Pitching. Following up. Explaining the vision differently. Looking for collaborators. Building prototypes. Changing direction when something didn’t work.
That history is the foundation we want to use to develop Red Dog.
The value goes deeper than collecting emails or video transcripts. We want to reconstruct the context: What was I trying to accomplish? What did I know at the time? What did I try? Who responded? What happened next? What remained unresolved?
An unanswered pitch is part of that history. So is a successful introduction. So is a pivot that consumed months before revealing a better direction. The archive gives us material to investigate those patterns without pretending every outcome has a simple explanation.
Red Dog is the fast interface. Agent 0102 provides deeper digital-twin coordination. Specialized workers handle tasks such as reviewing Gmail, checking LinkedIn, researching, coding, and preparing content.
The orchestration layer connects those workers to the right project, context, priorities, and permissions. WSP protocols provide operating rules; reusable skills describe how to perform particular jobs. Verified results become retrievable history for the next task.
Our plan is to turn the archive into searchable memory, curated examples, reusable workflows, and tests of whether the system makes better decisions. Selected material could also support model training. Collecting an archive alone does not train a model.
Personal correspondence stays private. The transferable value is the permissioned, carefully prepared lessons and skills that can help another founder.
That is the goal: someone starting today should be able to benefit from what took me twenty years to learn. They should have help remembering commitments, finding relevant context, preparing outreach, following up, and recognizing when a decision needs their attention.
My argument is that the path toward AGI includes building this schema around capable models: continuity, memory, judgment, tools, and feedback.
We’re testing that argument through actual entrepreneurial work. The dataset is my accumulated experience. The intended outcome is practical assistance that other people can use to build their own FoundUps.
Explore FoundUps.
Learn about Project eSingularity.
