DEV Community: Harsha Kumar

I found my content formula by accident — here's what 2,420 impressions taught me

Harsha Kumar — Fri, 05 Jun 2026 06:54:45 +0000

For 4 months I posted on LinkedIn consistently.
Journey posts. Vulnerability updates. Real numbers. Honest struggles.
Average impressions: 100-200.
Then 3 days ago I posted something different — a specific AI tool stack for starting a freelancing business. Step by step. Real tools. $0 cost. Specific timeframe.
2,420 impressions.
Nearly 10x my previous best.
Here's what I think happened:
Journey content builds followers. People who already follow you engage with it.
Actionable stack content gets shared. People who don't know you yet share it because it's immediately useful to someone they know.
Shares are the only thing that breaks you out of your existing audience.
The lesson: if you want to grow beyond your current followers, create content worth sharing not just content worth reading.
The difference: shareable content solves a specific problem for a specific person in a specific situation. Readable content is interesting but not immediately actionable.
I have 15 AI business stack playbooks. That's 15 LinkedIn posts in this exact format.
Should have figured this out 3 months ago.
Full stacks at xedge.tech — what business model do you want me to cover next?

The real ROI of building something nobody asked you to build

Harsha Kumar — Thu, 04 Jun 2026 09:07:47 +0000

Four months ago I started building XEdge.
No co-founder asked me to. No investor funded it. No professor assigned it.
Just me, a problem I personally had, and a decision to build the solution.
Here's what that decision has actually returned so far — not in revenue, but in knowledge:
Market research skills
I can now size a market three different ways and explain why the numbers converge or don't. No MBA taught me that. Three months of trying to convince myself XEdge was worth building did.
Distribution understanding
I know which content styles drive traffic and which don't — because I've tested 60+ posts across 5 platforms with real data. Emotional honest content outperforms informative content 4:1 for my audience. That's not theory. That's my actual analytics.
Pricing psychology
I almost launched at ₹199 out of fear. I didn't. The playbooks are $29. Understanding why that matters — and why fear-based pricing destroys perceived value — is worth more than any pricing course.
Resilience that can't be faked
I've had days with 14 users and days with 1. I showed up both days. That's a muscle. It only gets built one way.
Revenue: $0
Users: 500+
Lessons: irreplaceable
The ROI on building something real at 18 with no money isn't financial yet.
But it's compounding faster than anything else I could be doing.
xedge.tech — still going.

What 4 months of building in public actually taught me about content

Harsha Kumar — Wed, 03 Jun 2026 03:21:35 +0000

I've been tracking which posts drive traffic to XEdge for 4 months now.
The pattern is uncomfortably clear:
Emotional honest posts → 14 users in a day
Informative educational posts → 1-3 users
Promotional posts → 0-2 users
Everything I learned about content marketing said educational content builds authority and drives traffic.
My actual data says the opposite — at least at this stage with this audience.
Here's my theory on why:
Information is everywhere. Anyone can write "here are 5 AI tools for productivity." ChatGPT can write that in 10 seconds.
But nobody else is living my specific story. Nobody else is 18, solo, in India, building this specific thing with these specific constraints and these specific results.
That specificity is the only thing I have that's genuinely scarce.
The lesson: at early stage your story is your competitive advantage. Not your knowledge. Not your expertise. Your specific unrepeatable situation.
Tell that. Not the generic version of what you know.
Building xedge.tech — still going.

Why I stopped trying every new AI tool that dropped

Harsha Kumar — Tue, 02 Jun 2026 03:56:40 +0000

For the first two months of building XEdge I felt obligated to try every new AI tool that launched.
New coding assistant dropped? Downloaded it.
New writing tool launched? Tried it.
New productivity AI? Signed up.
I was spending more time evaluating tools than actually building.
The switch that changed everything: I stopped optimizing for having the best tools and started optimizing for knowing my current tools deeply.
My current stack hasn't changed in 6 weeks:
Claude for thinking and writing.
Claude Code for building.
PostHog for analytics.
Brevo for email.
Gumroad for payments.
That's it. Five tools. One product used by 500+ founders.
The best tool isn't the newest one. It's the one you actually know how to use.
If you're still drowning in tabs trying to find the perfect stack — xedge.tech cuts that down to 30 seconds.
What does your actual daily stack look like?

The exact AI stack to start freelancing this week — $0 budget

Harsha Kumar — Mon, 01 Jun 2026 05:27:06 +0000

Most freelancing advice tells you to "find your niche" and "build a portfolio" without telling you exactly how.
Here's the exact AI stack that does both — fast, free, and actually works in 2026.
Tool 1: Perplexity AI — Niche Research
Don't guess what skills are in demand. Ask Perplexity: "What freelance skills are businesses actively hiring for right now and what are they paying?" It pulls live data not outdated blog posts. Takes 20 minutes to find a niche with real demand.
Tool 2: Claude — Portfolio Building
You need samples before you have clients. Use Claude to generate 3 high quality samples in your chosen niche. Give it context about your target client and the specific deliverable. Edit it to sound like you. Most clients can't tell the difference between a polished AI-assisted sample and years of experience.
Tool 3: Apollo.io free tier — Client Finding
Apollo lets you search for decision makers at companies by industry, size, and role. Free tier gives you enough to build your first outreach list. Search for small businesses in your niche who would need your service.
Tool 4: Claude again — Outreach Writing
Paste the prospect's LinkedIn bio and company description into Claude. Ask it to write a personalized cold email that references their specific situation. Not a template — a specific message. Response rates are 3-5x higher than generic outreach.
Tool 5: Notion AI — Client Management
Once you land a client, Notion AI keeps everything organized — project briefs, deadlines, feedback, invoices. Looks professional from day one even if it's just you.
Total cost: $0
Time to first client: 1-2 weeks if you execute
This is one of 15 execution stacks I've built at xedge.tech — specific goal, exact tools, step by step guidance.
What freelance skill are you considering

Why Googling AI tools is costing you more time than it saves

Harsha Kumar — Sat, 30 May 2026 09:27:28 +0000

I asked founders in a Discord server how they find AI tools.
The answer was almost universal: Google first, Twitter trends second.
I get it. It's what I did too before building XEdge.
Here's the problem with both:
Google returns lists. "Top 10 AI tools for productivity." Those lists are SEO optimized not quality optimized. Half the tools are paid placements. The other half haven't been updated in 8 months.
Twitter trends return hype. A tool trends because it's new or has a good marketing team — not because it's the right fit for what you're building.
Neither one answers the actual question you have:
"I'm building X — what specific combination of AI tools gets me there fastest?"
That's a stack question not a list question.
The difference:
A list gives you options.
A stack gives you a workflow.
I built XEdge to answer stack questions — describe your goal, get the exact tools that work together for it, with guidance on how to actually deploy them.
500+ founders use it. Built solo. Zero paid marketing.
If you're still Googling every time you need a new tool — xedge.tech
What's the last AI tool you found through Google that actually stuck?

Why your AI tool collection is slowing you down

Harsha Kumar — Fri, 29 May 2026 07:47:14 +0000

There's a difference between having AI tools and having an AI stack.
A tool does one thing.
A stack does one goal.
When I started building XEdge I had 23 AI tools bookmarked. I was using maybe 4 of them regularly. The rest were either redundant, too complex to integrate, or solving problems I didn't actually have.
The moment I stopped collecting and started stacking — matching specific tools to specific workflows — my output doubled.
Here's the stack I use to run XEdge solo:
Writing and content: Claude — thinking, drafting, strategy
Coding: Claude Code — architecture, debugging, building
Analytics: PostHog — user behavior without sending data to Google
Email: Brevo — automated sequences, free tier covers everything
Payments: Gumroad — zero setup, handles tax automatically
Hosting: Firebase + Cloudflare — free at current scale
Six tools. One person. One product used by 500+ founders.
The goal was never to use the most tools.
It was to use the right ones together.
If you're still collecting instead of stacking — xedge.tech builds your stack based on what you're actually trying to do.
What does your current stack look like?

The most important metric I track isn't users or revenue

Harsha Kumar — Thu, 28 May 2026 07:44:54 +0000

It's whether I showed up today.
Three months building XEdge taught me one uncomfortable thing:
My traffic is almost entirely determined by whether I posted honestly that day.
Post honest emotional content → 14 users
Skip a day → 2 users
That's not an algorithm problem.
That's a consistency problem.
And consistency is the one thing that's entirely in my control regardless of budget, team size, or age.
Building xedge.tech — showing up today.
What's your honest metric today?

14 users in one day with zero budget — the exact post style that worked

Harsha Kumar — Wed, 27 May 2026 06:44:07 +0000

Yesterday I tried something different.
Instead of posting about features or tools I posted about the journey. The real numbers. The honest struggles. The specific feeling of building something alone at 18 with no money.
14 new users. Best single day in 3 months.
The posts that worked all had three things in common:
They led with emotion not information.
They were honest about being in the middle of it not at the end.
They drove to xedge.tech not directly to the product.
The posts that didn't work were promotional. Feature focused. Salesy.
The internet doesn't share what's selling them something.
It shares what makes them feel something.
Building xedge.tech — 500+ users, pre-revenue, still going.
What's your honest number today?

I wrote a startup playbook while building the startup. Here's why that matters.

Harsha Kumar — Tue, 26 May 2026 11:11:55 +0000

Most startup guides have the same problem.
They're written in hindsight by people who've already made it. Which means by the time you read them the author has forgotten what zero actually felt like. The uncertainty. The specific fears. The moments where you genuinely didn't know if any of it was worth it.
I wanted to write something different.
So I documented everything while building XEdge — an AI tool discovery platform I've been building solo at 18 with zero budget for the last 3 months.
The TAM/SAM/SOM chapter was written while I was figuring out XEdge's market.
The pricing chapter was written while I was setting XEdge's prices.
The psychology chapter was written after my analytics hit zero for the first time.
Every framework was tested before it was written down. Every prompt was used before it was shared. Every mistake was made before it was documented.
That's what makes it honest.
It's called the XEdge Startup Execution Playbook. 21 pages. $29.
If you're building something right now — not planning to build, actually building — it's at xedge.tech.
The truth from someone still in the trenches.

6 users in one day with zero marketing budget — what actually worked

Harsha Kumar — Tue, 26 May 2026 07:09:30 +0000

Three months of building XEdge taught me one thing above everything else:
Honest beats promotional. Every single time.
Yesterday I posted a genuine update on Peerlist and Threads — real numbers, real struggles, no hype. 6 new users in one day. My best day in weeks.
Meanwhile my most promotional posts — "check out this platform", "we just added X feature" — consistently get ignored.
The pattern is clear:
People share what makes them feel something. They ignore what's trying to sell them something.
Give people a real story to follow. The traffic follows the story.
Building xedge.tech — 500+ users, still pre-revenue, still going.
What's your honest number today?

I got 500 users with $0 marketing. Here's the exact breakdown.

Harsha Kumar — Mon, 25 May 2026 09:15:04 +0000

Everyone asks about growth hacks.
Nobody asks about the unglamorous stuff that actually worked.
Here's the honest breakdown of every user XEdge got in 3 months:
Discord — ~200 users
Not from posting links. From genuinely answering questions in startup and AI servers for two weeks before mentioning XEdge once. Trust first. Product second.
Direct / word of mouth — ~150 users
People telling other people. This only happens when the product actually does something useful. Build useful first.
Social media — ~100 users
One tweet hit 1K views: "Your competition isn't AI. It's someone using AI better than you." Everything else averaged 100-200 views. One post did 10x the work of everything else combined.
SEO — ~50 users
Early but compounding. Three months of indexed pages starting to show up in searches.
The honest lesson:
There is no hack. There is just showing up in the right rooms, being genuinely useful, and saying something true that makes people stop scrolling.
Building xedge.tech — solo, 18, India, still going.
What's your highest converting channel at early stage?