Momentum Mentor

2-Hour Sessions, Living towards Thriving

Visit Website
October 27, 2025 Built an AI consultant for Life Decisions (Not Therapy, Not Coaching)

The Problem I Was Solving

During a lonely trip to Las Vegas and visiting California, I was dealing with a relationship problem that consumed way too much mental energy. I'm an engineer—I think in systems and frameworks. But when I looked for help, I found:

  • Therapy: Too clinical, focuses on past trauma, takes months

  • Life coaching: Random advice without systematic frameworks, hit-or-miss

  • Friends/family: Well-meaning but inconsistent mental models

I realized: there's no systematic, engineering approach to personal decisions.

The Core Insight

Most people treat life decisions like random events. But activities in relationships, time management, and resource allocation are engineering problems with predictable patterns.

If we can optimize:

  • Supply chains (operations research)

  • Poker strategies (game theory)

  • Common pool resources (Ostrom's principles)

Why can't we apply the same rigor to life?

What I Built

Momentum Mentor - An AI consultant that analyzes client's situation through multiple frameworks simultaneously such as:

  • Personal Resource Management (time, energy, attention as finite resources)

  • Interpersonal Resource Management (relationship dynamics as commons governance)

  • Communication Effectiveness (structured frameworks for hard conversations)

  • Backtesting and Unit test (validate the system designed for the future life is as good as before, but able to handle more scenarios)

Expected outcome

Think of it like training an Reinforcement Learning agent on a human.

Life doesn't become perfect—it still fluctuates. But the trend improves when you apply systematic frameworks consistently.

Life System Performance Over Time

Key insight: We're not chasing happiness. We're engineering systems that produce sustainable, improving outcomes despite daily volatility.

Traditional approaches:

❌ Try to eliminate all bad days (impossible)

❌ React to problems as they come

❌ No systematic method to validate the effectiveness

Momentum Mentor:

✅ Accepts volatility as normal

✅ Optimizes the trend line

✅ Builds self-improving decision systems

✅ Covers multiple aspects in how to use the time and effort for better living

How It Works

  1. You describe your situation (2-hour intensive session, $100)

  2. Claude Sonnet 4.5 analyzes through all frameworks simultaneously (50K+ tokens)

  3. Get integrated analysis with specific, testable recommendations and try to teach you how to use those frameworks

  4. Data deleted after 30 hours (by design—no dependency, you learn the frameworks)

Key Design Decisions

Why 2 Hours Only?

It is a bug caused by reducing maintenance cost.

I deliberately limited sessions to prevent over-reliance. The goal is to teach you systematic thinking, not become a crutch. Think: personal trainer who teaches you to exercise independently, not a companion you can't live without.

Why $100 Flat Rate?

  • Minimal development

  • Works with session based revenue model

  • Lower than life coaching but review decision making process and help client self-evolve with evolutionary algorithm for mental models

Why Delete Data After 30 Hours?

It is a bug caused by minimizing development and maintenance.

Forces you to internalize the frameworks, not just outsource thinking. You describe your problem fresh each session, breaking old narratives.

Technical Stack

  • AI: Claude Sonnet 4.5 (50K cached mega-prompt across 7 frameworks)

  • Frontend: Next.js

  • Hosting: Google Cloud (Firebase)

  • Cost: $0.12~$0.70 per message (before prompt caching)

The "Aha!" Moment

Before testing, I was debating a decision with myself. During testing with this topic the bot revealed:

  1. A blind spot I hadn't considered

  2. A self-consistency issue in my reasoning

This would've taken me a week to figure out on my own. That's when I knew this worked.

Early Results

  • Built in 1 week

  • $0 marketing spend (launching today)

  • The back-end algorithm will try to maintain 50%-70% margin

Challenges

  1. [Vibe coding] Gemini CLI kept reverting my code to use Gemini API instead of Claude API (was annoying)

  2. Rate limits with multi-call approach (solved with mega-prompt)

  3. Explaining to non-engineers why this isn't "just another chatbot"

  4. Vulnerabilities in the business logic implemented by Gemini to be discovered and patched by hand

What's Next

  1. Content marketing on HN/LessWrong (my target audience: systematic thinkers)

  2. Case studies showing framework application

  3. Free "Life Systems Monitor" game (lead generation tool—tracks your "energy efficiency" over 14 days, identifies patterns, offers consultation if you're in the red)

Questions for IH Community

  • Is $100 acceptable?

  • How do I balance between accessibility to potential users that may have limited attention in readings and geeks like me who wants to read all perspectives at once?

Try It

🔗 momentummentor.com

Built by an engineer for systematic thinkers. Not therapy. Not coaching. Engineering consultation for life.

Comment

About

To educate users use time/effort wisely and share their resources wisely with partner base on engineering principles. If we optimize life of individuals, we should free up energy to be mobilized in prosocial activities.