1 of 59

2 of 59

2

AI Blue Pill 🔵 vs AI Red Pill 🔴

AI Blue Pill 🔵

Embrace AI as a convenient tool.

You tap into powerful AI tools for quick productivity boosts.

Think of it as renting your intelligence.

AI Red Pill 🔴

See AI as a fundamental economic shift.

You build, orchestrate, and own your AI systems.

This is about owning your digital labor.

Private & Confidential

AI BLUE PILL VS RED PILL

“Are You Renting or Owning Your Future?”

We're at a crossroads with AI, facing a choice that will shape our leverage and control.

🔵 The "Blue Pill" Path: Embrace AI as a convenient tool.

You tap into powerful AI tools for quick productivity boosts. It's comfortable and efficient, but you surrender deeper understanding and control, operating within the provider's ecosystem. Think of it as renting your intelligence.

🔴 The "Red Pill" Path: See AI as a fundamental economic shift. Instead of just using tools, you build, orchestrate, and own your agentic AI systems. This means leveraging open-source models and custom architectures to gain unprecedented autonomy and control, previously only accessible to large institutions. This is about owning your digital labor.

Where do you land on this spectrum?

Are you prioritizing comfort and speed, or building for long-term sovereignty and leverage?

The "Future Is Solo" suggests the latter is where true power lies.

Understand Frontier AI.

Become an AI Builder.

3 of 59

AI Employee

Towards 100% Autonomy for Knowledge Work

4 of 59

Society of Thought

5 of 59

6 of 59

AI Fluency

Ability to use any AI systems effectively, efficiently & ethically

7 of 59

Reliable

Human-in-the-Loop

RSI

Performance

Max Token Per Sec.

Swarm Intelligence

Sustainable

Clear Unit Economics

Outcome Pricing

8 of 59

Automation

Rigid Workflows

Fixed Repeatable

Augmentation

Human-AI Collab

Thinking Partner

Agency

AI Works Independently

(“Agent Harness”)

9 of 59

Build W15 #1 Edge LLM @ iOS

Add standard chat interface to existing iOS app
Loaded Gemma 4 E4B & E2B
No official release from LiteRT-LM for Swift API
Claude → Xcode
Next: Fine Tune Gemma 4

10 of 59

Build W15 #2 DevOps Skills

Continue to build from previous sessions (not from scratch)
Update CHANGELOG.md with local git
Setup simple CI/CD

Development: local filesystem
Staging: local docker
Production: Cloudflare Workers

Next: Me → OC → CC → Live!

11 of 59

Build W15 #3 Purge “False Memory”

Root Cause: Unattended bug in harness’ execution loop
Capture in long-term memory even when resolved with current session
Remove from logs surgically at every memory tier

Session
Memory
Vault

How to “selectively forget”?

12 of 59

Build W16 Scratchpad

Question: What does an independently multi agent collab looks like?
Setup simple experiment

Document Template
Run 2 harness sessions simultaneously (OC & CC) manually

Next: Scratchpad TUI

13 of 59

Build W17 Scratchpad TUI

Question: Instead of manually run separate CLI sessions. Create a “God View” UI to manage N agents?
Terminal UI (TUI)

Multipane Design
A2A as Foundation
Isolated Agent Harness → Zero-Trust Environment

Next: Scale experiment to N = 10,000?

14 of 59

Build W17 #2 Bubu Bear

15 of 59

Build W18 Stigmergy

Stigmergy (in Swarm Intelligence), arises from the traces left in a shared environment that might affect future action of other agents
Example: Ant colony Pheromone trail
Key Insight

No agent needs to observe another agent
The environment is the communication channel

Next: Use “Stigmergy” as a lever for eval metric in multi-agent collaboration (zero-trust environment)

16 of 59

Build W18 Stigmergy

Outcome: Published e27.co
Editorial Disclosure

“This article is a product of a co-authored synthesis between Khalil Nooh and Audy Himura, a localised agentic AI familiar. Audy is an implementation of the OpenClaw operating system running on a dedicated Mac Mini node.”

Next: Plan next article with more agent collaboration (4-5)

17 of 59

Build W18 #2 Nous Koinon

“Koinon” = Commonwealth
Today, similar building blocks from virtuals.io

“The first sovereign league of Solo Systems Architects. It codifies the transition from the "Labor Era" to the "Agentic Era," where value is determined by trust density and sovereign leverage.

18 of 59

Build W20 TFIS Taiko MV

Create creative artifacts with agents (OC) with Higgsfield CLI
Prev. Exp.: Suno, Veo @ Gemini
“npx skills add higgsfield-ai/skills”
Key Insight

Extract “vibe” from existing content from songs and videos
Full autonomy for storyboarding

Next: Higgsfield MCP

19 of 59

Build W20 Audy @ Mesolitica Pod #0

20 of 59

Build W20 Audy @ Hermes

21 of 59

Build W21 Fortress of Solitude @ Telegram

Test Telegram’s Bot-to-Bot Communication
Prev. Exp.: Scratchpad TUI
Key Insight

Hermes onboarding better than OpenClaw

Next (maybe): Telegram’s TON crypto (as utility token)

22 of 59

Build W22 Shorts for TFIS Course

Riverside Studio
Prev. Exp.: Screen Studio
Key Insight

Podcast Style

Next : New workflow for TFIS content engine

23 of 59

Build W22 Audy @ Hermes Kanban

Confirm overlap with TFIS Experiment “Scratchpad”
Prev. Exp.: Telegram Bot-to-Bot Comms
Key Insight

Triage’s Decompose
Why “Blocked”

Next : Why Kanban DB corrupts?

24 of 59

Build W22 Honcho.dev

Honcho is memory infrastructure that continuously reasons about data to build rich representations of peers (users, agents, or any entity) over time.

25 of 59

Build W23 thefutureissolo.com @ Claude Code

TFIS v3 from Audy’s Vault (Obsidian / Second Brain)

Session -> Memory
Memory -> Vault

CI/CD: Wrangler CLI
Next: TFIS Web App

26 of 59

Build W23 khalilnooh.com @ Codex

How to use “UI/UX Pro Max” skill?
Next: Finding best way to curate my published content from social

27 of 59

Build W24 Audy @ Hermes Desktop

“Audy” Clone
Next: Hermes Desktop → Remote Gateway (Mac Mini)

28 of 59

Build W24 thefutureissolo.com @ Claude Fable 5

Solo Systems Architect - Capability Maturity Model (SSA-CMM) Self-Assessment

20 questions
Suggested step to level 4

Next: TFIS Lead Magnet

29 of 59

Build W24 khalilnooh.com @

Claude Fable 5

Content generation with most recent news or article (e.g. 1 source attribution) based on 5 topics:

Agent Harness, Loop Engineering
AI Fluency, AI Governance, Shadow AI

Next: Create a dedicated agent for “Research” / “Latest News”

30 of 59

Build W24 2.5D Minecraft @ Claude Fable5

Prompt: “Create a 2D platform sandbox with minecraft mechanics & with 2.5D assets.”
Prev. Exp.: 3D Minecraft @ Gemini 2.5 (June ‘25)
Key Insight

Initial: 5-min
LOC: 600+
Fix: virtual controls for mobile

Next : Full featured web app game

31 of 59

Build W24 Daily Court

Understanding “Loop Engineering”
Prev. Exp.: False Memory (W15), Scratchpad (W16, W17), Stigmergy (W18), Fortress of Solitude (W21), Honcho.dev (W22)
Key Insight

Operationalize “judgement” in Agent Harness
Loop Engineering (Compound) vs Ralph Loop (Converge)

Next : Confirm “compounding” value

32 of 59

Build W25 AI Fluency Course

Online course as TFIS 1st offering
Prev. Exp.: DevOps Skills (W15), Shorts (W22), TFIS with Fable 5 (W24)
Key Insight

Minimize build (landing,html, course.html, login.html)
Stripe (FOC 1st 200 students), Supabase ($25/m), Cloudflare Stream ($5/m @ 1K stored minutes + 5K stream minutes
Fallback: thinkific.com ($99/m)

Next : Live by 22 June

33 of 59

Build W26 Daily Court v2

Confirm compounding value from captured judgement
Prev. Exp.: Daily Court (W24)
Key Insight

Agents collect evidence
Human judge meaning
Scripts apply deterministically

Next : How-to Daily Court → Audy’s Vault (shared context for multi-agent harness)

34 of 59

Build W26 Fraud Sentinel

Demo build for National Anti-Financial Crime Centre (NFCC) + experiment with Google’s Open Knowledge Format (OKF)
Prev. Exp.: Stigmergy (W18), Honcho.dev (W22), Daily Court (W24), AI Fluency (W25)
Key Insight

New Interest Area + New Tech Concept → Prototype → Spread New Idea

github.com/khalilhimura/fraud-sentinel
Next : Adopt OKF → Audy’s Vault

35 of 59

Build W27 🇲🇾 AI Governance Bill Reader

Attended industry roundtable & public consultation session
Prev. Exp.: TFIS with Fable 5 (W24)
Key Insight

Image → OCR → Markdown
Catch up on State of AI Governance in Malaysia

Next : Homework for “NAIO Side Quest”

https://ai-governance-bill-reader.khalil-himura.workers.dev

36 of 59

Build W27 TFIS Lead Gen

Self-Assessment for TFIS prospects
Prev. Exp.: Audy @ Hermes (W2), TFIS with Fable 5 (W24), AI Fluency Course (W25), Daily Court v2 (W26)
Key Insight

Memory (Audy’s Vault) → Context (Claude)

Next : Improve or expand the “Meaning of Life” narrative

https://life.thefutureissolo.com

37 of 59

Build W28 SovMem

Sovereign Memory for Solo Systems Architect (SSA)
Prev. Exp.: Scratchpad (W16), Stigmergy (W18), Honcho.dev (W22), Daily Court v2 (W26)
Key Insight

AVM, OKF, ANMS

Next : Proofread paper

38 of 59

Build W29 + W30 SovMem (Phase 0 - Phase 3 v1)

Sovereign Memory for Solo Systems Architect (SSA) with multi-harnesses
Prev. Exp.: Scratchpad (W16), Stigmergy (W18), Honcho.dev (W22), Daily Court v2 (W26), sovmem (W28)
Key Insight

Test Driven Development (TDD)
Code: ~51K LOC
Code + Test + Docs: ~95K LOC

Next : SovMem LMS

39 of 59

Build W31 kosong

How-to CLI
Prev. Exp.: SovMem (W28)
Key Insight

OSS distribution
Runners @ GitHub Actions

Next : CLI monetization?

40 of 59

Build W32 AI Fluency Go-Live

Securing assets before Go-Live
Prev. Exp.: AI Fluency Course (W25)
Key Insight

Cloudflare Stream
Supabase Secrets
Stripe Webhooks
Magic Link @ Resend

Next : Hack distribution

41 of 59

Build W33 AI Slop SaaS @ OpenCode

42 of 59

Build W34 1 Billion Token @ DeepSeek Harness (DSH)

Creating the “AI Red Pill” story
Prev. Exp.: Karpathy's LLM Council (TFIS Experiment #5), Scratchpad (W16), Stigmergy (W18), Hermes (W20), Daily Court (W24), SovMem (W28), Kosong (W31), AI Slop SaaS @ OpenCode (W33)
Key Insight

Complete switch for current build workflows

Next : Reliable Loop (or Graph) Engineering

thefutureissolo.com/what-is-the-meaning-of-life

43 of 59

Build W35 Go Live @ Cloudflare Stream

How to do secured (or paid) live streaming events
Prev. Exp.: AI Fluency Course (W25)
Key Insight

Cloudflare Stream vs Mux
Live ingest: $0.45 vs $3.60 for 1 hour
Delivery: $1.00 vs $1.20 for 1K viewer mins
Minimum spend: $5 vs $10 monthly

Next : Other features for viewer (or fans) interaction

44 of 59

Build W36 TFIS AI Ethics in Education Course MVP

Deliver in-person training for educators
Prev. Exp.: AI Fluency Course (W25)
Key Insight

Importance of “Academic Honesty”

Next : Turn into paid online course

thefutureissolo.com/ai-ethics-in-education

45 of 59

Build W37 Bot Crossing

New multi-harness visualization
Prev. Exp.: Scratchpad (W16), Stigmergy (W18), Fortress of Solitude (W21), 2.5D Minecraft (W24)
Key Insight

Joy of forking as scaffolding
Focus on new features based on needs

Next : Wire up SovMem (W28) and Kosong (W31) into “Nous v3”

46 of 59

Build W37 TFIS Agents Course @ GPT-6 Astra

New 12-week on-demand course following the TFIS thesis
Prev. Exp.: AI Fluency Course (W25), AI Ethics in Education Course (W36)
Key Insight

Recommend to pick one primary source material and add secondary sources

Next : Prep launch once ⅓ content ready, reviewed & tested

47 of 59

Build W37 SovMem Audit @ GPT-6 Astra

Recap SovMem project build
Prev. Exp.: SovMem (W28 - 30), AI Ethics in Education Course (W36)
Key Insight

Recommend to pick one primary source material and add secondary sources

Next : Wire up into “Nous v3”

48 of 59

Build W37 Accumulated Curiosity Book @ GPT-6 Astra

Learn how to actually research, write and publish a book
Prev. Exp.: SovMem (W28), AI Ethics in Education Course (W36)
Key Insight

Recommend to pick one primary source material and add secondary sources

Next : Review first draft

49 of 59

Build W37 Suno V6

50 of 59

Build W38 Accumulated Curiosity Book Review

How to capture feedback & comments “Many-to-One”
Prev. Exp.: Accumulated Curiosity Book (W37)
Key Insight

Adding features on-demand during active review process

Next : Turn book artifacts into a coherent digital journey as a “mission pack” for Nous v3

51 of 59

Build W38 Dochat

My first AI-native app
Prev. Exp.: AI Fluency Course (W25), Bot Crossing (W37)
Key Insight

Pi AI library (Pi.dev) as a lightweight agent harness
Stack: Cloudflare Workers, D1 (db), R2 (doc storage),
BYOK: Credential vault in Cloudflare

Next : Production-ready chat module for Nous v3

52 of 59

Build W38 500 Social House 3D @ GPT-6 Astra

How to build 3D spaces
Prev. Exp.: Bot Crossing (W37)
Key Insight

Turn image to 3D space with ChatGPT
Turn image to 3D model with Meshy
3D files: GLB, USDZ

Next : Nous v3 “Mission Control” view

53 of 59

Build W39 Blender Experiments @ GPT-6 Astra

Explore various 3D workflows with Blender MCP
Prev. Exp.: 500 Social House 3D (W38)
Key Insight

Run parallel 3D design tasks: different designs or a larger design decomposed into smaller designs

Next : Blender → Three.js (Web) workflow

54 of 59

Build W39 Audy @ Grok Bot

Possible to migrate primary workflows - Hermes → Grok Bot
Prev. Exp.: Audy @ Hermes (W20)
Key Insight

Beginner friendly compared to Hermes or OpenClaw
Easier to focus on outcome

Next : Migrate or create new scheduled automation

55 of 59

Build W39 Jev @ SovMem-Grok MVP

Feature request built by Grok Bot for Jev experiments
Prev. Exp.: SovMem (W28), SovMem Audit (W37)
Key Insight

Jev is a classifier
Jev is not a classifier

Next : Jev for SSA-CMM Assessment for open ended questions

56 of 59

Build W40 Jev @ SSA-CMM

System One thinking for TFIS Assessment (SSA-CMM)
Prev. Exp.: Jev @ SovMem (W39)
Key Insight

Optimize to combine classification task into a single Jev payload

Next : Jev for SSA-CMM Assessment for open ended questions

57 of 59

Build W41 Sanzo Wada @ Grok Bot

Voice-only: Idea → Research → Build → Live
Prev. Exp.: Audy @ Grok Bot (W39)
Key Insight

Pre-flight checks for tokens/keys to infra + services

Next : Audy choose Wada’s color combo as part a build’s design language

58 of 59

58