Building with Claude

Building with Claude: Finding Passive Income Without Losing Myself

June 20, 2026 · 12 min read

The problem I started with

I was drowning in possibilities and paralyzed by them.

I’m a web developer working remotely for a UK company, earning $800/month. That’s good money in Bangladesh, but not enough. I wanted passive income. I wanted to write poetry. I wanted to do research. I wanted to build with AI. I wanted to build a personal brand. I wanted funded scholarships. I wanted to publish papers.

I wanted everything.

And every idea I found online assumed I was in the US, competing on price, with 40+ hours/week of free time. None of it fit my life.

So I asked Claude: “What actually works for my specific situation?”

What followed was one of the most honest conversations I’ve had about money, identity, and what’s actually realistic.

What I got wrong (the journey)

Wrong #1: Thinking I had to choose

I came in thinking: “Should I do newsletters OR consulting OR teaching OR poetry OR research?”

Claude kept saying: “You don’t have to choose. But you have to sequence.”

I resisted this for a while. I wanted to do everything at once. I thought if I worked hard enough, I could write blog posts AND create courses AND write poetry AND research scholarship opportunities.

What I learned: Trying to do everything at once while working full-time is how you burn out and do nothing well.

Wrong #2: Thinking bigger income = better

I was fixated on $2000–5000/month from consulting. Seemed like the obvious path.

But Claude asked: “Are you trying to escape your job or build freedom?”

Turns out, they’re different things.

I don’t hate my job. I make $800/month. If I add $300/month passive income, I’m not rich, but I have OPTIONS. I can say no to bad projects. I can take a month off. I can try weird things.

$300/month passive > $3000/month trading time for money.

I was chasing the wrong number.

Wrong #3: Thinking passive income means no work

I thought “passive” meant I write something once and money appears forever.

Nope. It means I do the work once, then it compounds over time. But I still have to be consistent.

A newsletter sponsor deal needs 500+ subscribers first. That takes 3–4 months of 2–3 posts/week. That’s not passive, that’s building. Once it’s built, THEN it becomes passive.

I confused “eventual passive” with “immediate passive.”

Wrong #4: Thinking content had to be optimized

I was overthinking the format. Blog + Twitter + LinkedIn + Email + TikTok + Instagram.

Claude said: “Write a blog post. Everything else comes from that.”

Simple. But I was making it complicated.

Wrong #5: Thinking I needed to escape being a “dev guy”

I was worried about getting “caged” as a developer.

But Claude pointed out: “You’re not choosing between dev OR poet. You’re using dev income to fund poet + researcher work.”

Dev is the vehicle, not the cage.

What Claude got wrong

Wrong #1: Initial freelancing assumptions

Claude initially suggested $2000–5000/month consulting as the “fast path.”

That’s true for US developers. Not true for Bangladesh developers competing on Upwork. Claude wasn’t contextualizing the currency difference properly at first.

It took me explicitly saying “$800/month is my salary” for Claude to recalibrate.

Lesson: Context matters. Even good advice can be wrong for your situation.

Wrong #2: Underestimating passive income difficulty

Claude kept saying “newsletter sponsors by month 3–4” like it was guaranteed.

It’s possible, but not guaranteed. It requires consistency, good content, and luck. Claude presented it more confidently than was warranted.

What Claude could have done: “This is possible, but depends on execution. Here’s the realistic range and failure modes.”

Wrong #3: Not asking about my actual constraints earlier

Claude went through several long strategies before asking:

What’s my actual salary? ($800)
What’s my actual goal? ($200–300)
What’s my timezone/location? (Bangladesh)
What’s my actual free time? (3–4 hours/day)

All of this would have made Claude’s advice better from the start.

Lesson: Better questions earlier = better answers.

What we actually solved

By the end of the conversation, I had:

A realistic income goal: $300/month passive in 6 months (not $5000).
A sequenced path: Newsletter + digital products + poetry (not everything at once).
A time-bound plan:

Months 1–3: Write 2–3 posts/week, build the newsletter (3–4 hrs/day).
Months 4–6: First sponsor inquiries, product launches.
Month 6+: Sustainable $300–500/month passive.

Reusable prompts: To turn any day’s work into blog posts.
Permission to keep my job: It’s security, not a cage.
Understanding my actual edge: Dev + poet + researcher is rare, not limiting.

The real insight

The breakthrough came when Claude said:

“Right now, being a ‘dev guy’ for 3 months is how you buy yourself freedom. Because then you can do poetry full-time. Because then research becomes real.”

I was thinking: “How do I not be just a developer?”

Should have been thinking: “How do I use development income to fund my actual interests?”

Different question. Better answer.

The system we built

At the end, we built a simple prompt system for turning any workday into publishable content.

Three prompts:

Mid-work reflection (use 2–3× during the day)
End-of-day markdown post (publish to the blog)
Social posts from the blog (create later)

Takes ~20 minutes/day. Turns work into content. Content builds audience. Audience → sponsors.

And importantly: the blog post stays forever. Social posts come and go. But the blog post becomes SEO, becomes reference, becomes portfolio.

What happens next

In 3 months, I should have:

300–500 newsletter subscribers
24–36 blog posts
1 digital product ready
12 poems published
First sponsor inquiries maybe arriving

In 6 months:

$300–500/month passive
Time freed up for actual research
Poetry that’s been tested publicly
A real body of work

And I didn’t have to choose between code and poetry. I just had to sequence them.

The prompts (copy-paste ready)

Prompt 1: Mid-work reflection

Use this 2–3× throughout your workday when you finish something:

REFLECTION CHECKPOINT

We just finished working on something.

Please help me reflect on this work:

1. What category of problem did we solve? (be general: "form validation", "API integration", etc.)
2. What approach did I suggest and WHY did I suggest it?
3. What did I get wrong or assume incorrectly?
4. What did you get wrong or suggest poorly?
5. What's the core lesson we learned?

Write your reflection as a detailed section (400-600 words).
Be conversational and honest about mistakes.
Explain the HOW and WHY, not just the WHAT.
Include code snippets if relevant (but generalized).
Show the actual thinking process.

IMPORTANT: Don't mention any client names, company details, site names, or identifying information.
Keep everything generalized and broadly applicable.

Prompt 2: New day start (optional)

Use this at the start of your workday if you want:

I'm starting a new work day today.

Throughout the day I'll ask you to reflect on what we accomplish.

Key principles for today:
- Focus on METHODOLOGY, not project specifics
- Generalize all details (never mention client names, site names, company details)
- Document the PROCESS and LEARNING, not the project
- Be detailed about HOW and WHY
- Be honest about mistakes and confusion

Acknowledge you understand these guidelines.

Prompt 3: End-of-day — create markdown

Use this at the end of work to create your blog post:

CREATE BLOG POST FROM TODAY'S WORK

Look through our entire conversation history from today.

IF reflection sections exist:
- Use those reflections to create the blog post
- Combine all reflections into one coherent narrative

IF NO reflection sections exist:
- Analyze our conversation naturally
- Extract the topics, problems, solutions, and lessons from what we discussed
- Figure out what we worked on, what went wrong, what went right
- Come up with the main themes and insights yourself
- Don't ask me for anything - just create the blog post from the conversation

Either way, create a markdown blog post:

IMPORTANT: Write this blog post entirely from Mohammed's (the user's) first-person perspective.
- Use "I" when referring to Mohammed's actions, thoughts, mistakes
- Use "Claude" or "the AI" when referring to what the AI did
- The story is told from Mohammed's experience and viewpoint
- Mohammed is the narrator, not Claude

---
title: Building with Claude, [Today's Date]: [Main Insight]
date: [Today's Date in YYYY-MM-DD format]
---

## Opening Hook
2-3 sentences introducing the problem category (generalized, no project details).
Write from Mohammed's perspective: "Today I debugged...", "I worked on..."

## The Problem We Were Solving
Explain what kind of problem this was and why it matters.
From Mohammed's viewpoint: "I was solving...", "The challenge I faced was..."
Generalized language only - no client/company/site names.

## What I Tried First (The Approach)
Detailed explanation of Mohammed's initial thinking.
Why this approach was chosen.
Mohammed's mental model and assumptions.
Use "I thought...", "I decided...", "My approach was..."
Include relevant code snippets (generalized).

## What I Got Wrong
Mohammed's specific mistakes.
Explain the false assumption Mohammed had.
Why Mohammed believed it at the time.
What changed Mohammed's perspective.
Use "I assumed...", "I was wrong about...", "I didn't consider..."

## What Claude Got Wrong
Specific details about what Claude (the AI) did wrong.
Why Claude's approach seemed plausible but was wrong.
What data Claude was missing.
How Claude's mistake derailed the investigation.
Write this as: "Claude suggested...", "The AI missed...", "Claude's approach failed because..."

## What We Actually Solved
Detailed explanation of the real solution that Mohammed discovered.
Why it worked.
How it differed from Mohammed's initial approach.
The thinking process that led Mohammed to it.
Use "I realized...", "I discovered...", "The actual fix was..."
Include code if relevant (generalized, safe to share).

## Key Lessons Learned
3-5 actionable insights that apply broadly.
Lessons Mohammed learned from this experience.
Explain the WHY behind each lesson.
Show how these lessons apply beyond this specific work.
Use "I learned...", "This taught me...", "The key insight is..."

---

CRITICAL REQUIREMENTS:
- 1200-1800 words total
- Conversational tone
- Detailed HOW and WHY explanations
- Honest about confusion and mistakes
- Show the actual problem-solving journey
- NO client names, company details, site URLs
- NO identifying business information
- Code examples are generic and safe to share
- Focus on METHODOLOGY and LEARNING
- WRITTEN ENTIRELY FROM MOHAMMED'S FIRST-PERSON PERSPECTIVE
- Mohammed is the narrator and protagonist
- Claude is the supporting character in this story

FILE SAVING:
Save this as a markdown file with the filename: building-with-claude-[TODAY'S-DATE]-[SHORT-TITLE].md

Use YYYY-MM-DD format for date. Use 1-3 word title, lowercase, hyphens only.

Output the complete markdown content, ready to save.
Include the frontmatter (title and date).
No additional text before or after.
The file should be ready to copy directly to the blog folder.

Use this AFTER publishing to create social content:

CREATE SOCIAL POSTS AND SAVE TO FILE

I'm giving you my published blog post. Create social media posts and save them as a markdown file.

MY BLOG POST:
[PASTE YOUR PUBLISHED BLOG POST HERE]

---

IMPORTANT: Write all social posts entirely from Mohammed's (the user's) first-person perspective.
- Use "I" when referring to Mohammed's actions, thoughts, discoveries, mistakes
- Use "Claude" when referring to what the AI did
- Mohammed is the narrator and protagonist
- The story is told from Mohammed's experience and viewpoint

---

INSTRUCTIONS:

Create social posts in markdown format and save as a file.

Format the output like this:

---
title: Social Posts - [Blog Post Title]
date: [Today's Date]
---

## Twitter Thread

Tweet 1:
[Opening hook from Mohammed's POV about what I discovered/learned today]
[Example: "Today I learned something about X that changes how I think about Y"]

Tweet 2:
[What I got wrong - my false assumption]
[Example: "I was wrong about... I assumed..."]

Tweet 3:
[What Claude got wrong - the AI's mistake]
[Example: "Claude suggested... but missed that..."]

Tweet 4:
[The actual solution I discovered - the methodology]
[Example: "Here's what actually worked..."]

Tweet 5:
[Key lesson 1 I learned - broadly applicable]
[Example: "I learned that..."]

Tweet 6:
[Key lesson 2 I learned - broadly applicable]
[Example: "This taught me..."]

Tweet 7:
[CTA with blog link]
[Example: "Read my full reflection: [link]"]

---

## LinkedIn Post

[200-300 word professional summary from my perspective]
[Start with what I worked on: "Today I debugged...", "I discovered...", "I learned..."]
[What I got wrong and what I learned]
[How Claude helped but also where it failed]
[Key insights I gained]
[Why this matters for developers/builders]
[Generalized language throughout - no client info]
[Include blog link]

Write as a first-person reflection: "I've learned...", "My approach was...", "This experience taught me..."

---

## Reddit Post (Optional)

[Story format - what happened, what I learned]
[Conversational tone from my experience]
[Humble and honest about my mistakes]
[Generalized, no client info]
[Link to full post]

Use "I" throughout: "I was working on...", "I thought...", "I discovered...", "I realized..."

---

TONE REQUIREMENTS:
- Honest about my mistakes
- Focus on methodology and learning
- Broadly applicable (not project-specific)
- Invites critical thinking
- Safe to share publicly (no client exposure)
- Quotable tweets
- Engaging LinkedIn post
- Conversational Reddit post
- ALWAYS from my (Mohammed's) first-person perspective

---

FILE SAVING:
Save this as a markdown file with the filename: social-posts-[TODAY'S-DATE]-[BLOG-TITLE].md

Examples:
- social-posts-2026-06-25-certificates.md
- social-posts-2026-06-26-methodology.md

Use YYYY-MM-DD format for date.

Output the complete markdown content, ready to save.
Include the frontmatter (title and date).
No additional text before or after.
The file should be ready to copy directly to your notes or posting system.

The meta realization

I just wrote a post about how to turn work into blog posts by asking Claude to help me turn this conversation into a blog post.

So I’m using the system to write about the system.

Which means this post becomes:

Content for my site
An example for others
Proof that the system works
Fodder for tweets/LinkedIn posts
A portfolio piece

One conversation. One blog post. Infinite uses.

That’s the whole thing.

*This post was created by reflecting on a day-long conversation with Claude about building passive income, finding identity as a developer + poet + researcher, and creating a sustainable content system. It’s the first real test of the prompt system described above.*