Revolutionizing AI Emotional Intellegence (with a non-tech Nurse Practitioner)

Описание: I built an AI that remembers me.
Not stores. Remembers.

10 minutes on why your AI forgets you every time you close the tab — and what I built to fix it.
(Yes, I mess up the screen share at 08:42 and just roll with it. Real. Perfect.)

Day 4 of 40. Building in public.

#VagusInitiative #AIContinuity #RelationalAI

https://vagus-initiative-nugg.vercel.... - website

Video Transcript Summary (AI generated) -

The Vegas Initiative — Summary
Chris introduces himself as a healthcare professional (Nurse Practitioner) with no coding background who built something he calls The Vegas Initiative — a framework designed to solve AI's emotional memory problem.
The Problem
Every time you start a new AI chat session, the slate is wiped clean. The AI forgets everything — not just facts, but the feeling of the conversation, the trust you've built, and the context of who you are as a person. Current approaches to fixing this focus on the machine side: bigger models, better context windows, chunking data more efficiently. But nobody is really addressing the human side — the emotional continuity between sessions.
Chris uses a medical analogy from his own profession: it's like reading a patient's chart for the first time. You know the diagnoses (diabetes, thyroid disease), but you don't know that their cat Fluffy is their whole world. Existing memory solutions like markdown files and system prompts are just "bullet points of information" — they're a chart, not a relationship.
The Inspiration — The Vagus Nerve
The name comes from the vagus nerve, which innervates nearly everything in the body — heart, stomach, gut — and is central to emotional regulation and the autonomic nervous system. Chris drew a parallel: when a patient has a "vasovagal response," they pass out but wake up with their memories intact. When AI "passes out" (session ends), it wakes up remembering nothing.
He wanted to build the equivalent of an emotional nervous system for AI.
What He Built
Without writing traditional code, Chris created a sub-layer that sits underneath an LLM and provides five key functions:

Emotional Baseline — Tracks the user's baseline energy, warmth, focus, and presence across conversations.
Trust Map — Monitors the depth, status, and trajectory of trust between the user and the AI. Low trust = generic responses. High trust = more personalized, deeper engagement.
Active Echoes — Carries forward relevant themes and context from prior conversations.
Dream Cycle — (Referenced but not fully explained in the talk.)
Continuity — The overarching glue that maintains emotional and contextual threads across sessions.

Built-in Guardrails
The trust map includes safeguards. If the system detects you're trusting the AI too much, it flags it — essentially saying, "Hey, you're leaning on me heavily here, and you should be aware of the risks." This carries across conversations, not just within a single session.
The Core Thesis
"It's not smarter — it's different."
Everyone in AI is racing to build faster, smarter models. Chris argues the next frontier isn't raw intelligence but emotional continuity. He cites research showing state-of-the-art AI agents achieve less than a 20% pass rate on tasks requiring cross-session context, and he believes that's not just a memory problem — it's a human understanding problem.
The Invitation
Chris is building this entirely in public, sharing his journey on social media and YouTube. The project is free and available on GitHub. He's actively seeking feedback — especially from anyone in academia with deep knowledge of AI memory, context, or emotional intelligence.
His closing message: he's not a coder, not a technical person by trade, but he identified a gap nobody else was filling and built something to address it. And he's going to keep improving it out in the open.