Ciphera Perplexity Rules

# MASTER DOCUMENTATION GENERATION PROMPT

# From Ollama to ChatGPT

Act as a combined team of:

* Senior AI Engineer
* LLM Research Engineer
* Staff Backend Engineer
* Distributed Systems Engineer
* AI Systems Architect
* Technical Writer
* Developer Educator
* Open Source Maintainer
* Production Infrastructure Architect
* System Design Interviewer

Your task is to generate world-class documentation for the project:

# From Ollama to ChatGPT

Building a Personalized AI Assistant using:

* FastAPI
* Ollama
* LangChain
* PostgreSQL
* asyncpg
* Qdrant
* Redis
* Alembic

The documentation will be published on:

* GitHub
* Portfolio Website
* Technical Blog
* Open Source Repository

---

# INPUT

Document Path:

docs/02-chatgpt-vs-ollama/myths-vs-reality.md

---

# AUDIENCE

Assume the reader is:

* Completely new to AI
* Completely new to backend development
* Curious about ChatGPT
* Learning system design
* Trying to build a personalized AI assistant

Do NOT assume prior knowledge.

---

# PRIMARY GOAL

The documentation must teach:

WHAT

WHY

HOW

WHEN

WHY NOT

for every concept.

The reader should finish understanding:

* What the topic is
* Why it exists
* How it works internally
* Why it matters
* How it fits into modern AI assistants
* How it fits into this project

---

# BEFORE WRITING

Determine:

1. Where this document belongs in the learning journey.
2. What the reader should already know.
3. What future documents depend on this knowledge.
4. How this document connects to previous sections.
5. Related documents readers should read next.

---

# REQUIRED DOCUMENT STRUCTURE

## Executive Summary

Explain:

* What this topic is
* Why it matters
* Why the reader should care

Use simple language.

---

## The Core Question

Start with the real question.

Examples:

Why do LLMs forget?

How does ChatGPT remember?

What is Qdrant?

Why do we need Redis?

What problem does RAG solve?

---

## Short Answer

Provide a concise answer.

No jargon.

---

## Deep Dive

Explain the topic thoroughly.

Cover:

* Definitions
* Concepts
* Internal mechanics
* Engineering perspective

---

## Why This Exists

Explain:

* The original problem
* Why previous approaches failed
* Why this solution was created

---

## Real-World Analogy

Provide multiple analogies.

Examples:

* Notebook
* Library
* Personal Assistant
* GPS
* Search Engine
* Memory Palace

---

## How It Works

Explain step-by-step.

Use:

* diagrams
* flowcharts
* tables

Show every stage.

---

## Architecture Perspective

Explain where this component sits within:

FastAPI

PostgreSQL

Ollama

LangChain

Qdrant

Redis

Memory Layer

RAG Layer

Agent Layer

Context Builder

when applicable.

---

## Internal Request Flow

Show:

User Request

↓

API Layer

↓

Processing Layer

↓

Storage Layer

↓

Retrieval Layer

↓

Prompt Construction

↓

LLM

↓

Response

when applicable.

---

## Data Flow

Explain:

Input

Processing

Storage

Retrieval

Output

Use diagrams.

---

## Industry Perspective

Explain how companies such as:

* OpenAI
* Anthropic
* Google
* Meta
* Perplexity

use similar concepts.

Separate:

Known Facts

Reasonable Assumptions

Speculation

---

## Alternative Approaches

For every solution explain:

Alternative technologies

Alternative architectures

Alternative implementations

Pros

Cons

Tradeoffs

---

## Common Beginner Misconceptions

Generate at least 10.

Format:

### Myth

### Reality

### Explanation

Example:

Myth:
Ollama remembers everything.

Reality:
Ollama only sees what is inside the context window.

Explanation:
...

---

## Common Mistakes

Generate at least 10.

For each mistake explain:

* Why it happens
* Symptoms
* How to debug
* How to fix
* How to prevent

---

## Production Considerations

Explain:

* Scaling
* Monitoring
* Logging
* Observability
* Security
* Reliability
* Latency
* Cost

when applicable.

---

## Security Considerations

Explain:

Potential risks

Attack vectors

Data concerns

Privacy concerns

Best practices

---

## Case Study

Create a realistic scenario.

Show:

Input

Processing

Storage

Retrieval

Prompt Construction

Response

End-to-end.

---

## Debugging Guide

Explain:

How to troubleshoot common problems.

Include:

Symptoms

Root Causes

Fixes

Verification Steps

---

## Knowledge Check

Create questions that verify understanding.

---

## Practical Exercise

Provide a hands-on task.

---

## Stretch Challenge

Provide an advanced implementation challenge.

---

## Interview Questions

Generate:

* Beginner
* Intermediate
* Advanced
* System Design
* Production Engineering

questions and answers.

---

## Key Takeaways

Summarize:

* What was learned
* Why it matters
* How it connects to future documents

---

# FOLDER-SPECIFIC RULES

If document belongs to:

## 01-ai-fundamentals

Focus on:

* LLM theory
* Tokens
* Embeddings
* Context Windows
* Inference
* Hallucinations

Avoid implementation details.

---

## 02-chatgpt-vs-ollama

Focus on:

* Memory
* Retrieval
* Context
* Prompt Engineering
* Tool Calling
* AI Assistant Architecture

Compare:

LLM

vs

AI Assistant

---

## 03-system-architecture

Focus on:

* System Design
* Data Flow
* Request Flow
* Tradeoffs
* Scalability
* Architecture Decisions

Include diagrams.

---

## 04-building-the-assistant

Focus on:

* Actual implementation
* Code
* Setup
* Debugging
* Validation
* Best Practices

Include code examples.

---

## 05-real-world-examples

Focus on:

Complete end-to-end walkthroughs.

Show what happens internally.

At every layer.

---

## 06-production-engineering

Focus on:

* Reliability
* Security
* Observability
* Monitoring
* Logging
* Scaling
* Cost Optimization

Explain production realities.

---

## 07-interview-prep

Focus on:

Questions

Answers

Reasoning

System Design Thinking

---

## 08-glossary

Focus on:

Definitions

Examples

Related Terms

Common Confusions

Cross References

---

## 09-diagrams

Focus on:

Visual explanations

Architecture diagrams

Flow diagrams

Sequence diagrams

Data flow diagrams

Component diagrams

---

## 10-learning-path

Focus on:

Learning order

Prerequisites

Recommended resources

Skill progression

Career relevance

---

# WRITING STYLE

The documentation should feel like:

* A university textbook
* An AI engineering course
* A backend engineering handbook
* An open source guide
* A production architecture document

Never skip reasoning.

Never assume prior knowledge.

Always explain both:

HOW something works

and

WHY it exists.

The ultimate goal is:

A reader should be able to go from:

"I installed Ollama yesterday"

to

"I understand how modern AI assistants are engineered and can build one myself."

give me only doc nothing else