🔧 Tools 📰 Blog 🥽 XR Hub 🧪 Experiments
Live Experiment Series · Started June 2026

Can AI actually
make you money
on the side?

Prabhu Kumar Dasari
Prabhu Kumar Dasari
Senior AI Developer · Testing so you don't have to

Every week there's a new article claiming AI can replace your income, build your business, or grow your audience. They all say the same things. None of them actually show you the numbers.

This series is different. I run real experiments with real AI tools, document every step, and publish the actual results — including when it fails. No affiliate incentive to hype anything up. No theory. Just what happened.

Each experiment starts with a hypothesis, tracks daily progress, and ends with an honest verdict: Worth it. Conditional. Not worth it. The abandoned ones are usually the most useful reads.

How every experiment works
🎯
Clear HypothesisOne testable question. Can AI do X? Does Y actually work?
📅
Fixed Timeline7 days, 30 days, or a defined sprint — not open-ended.
📊
Real MetricsActual numbers — reach, revenue, cost, hours. No vague claims.
🧾
Full Cost BreakdownEvery tool cost, subscription, and credit is accounted for.
📸
Screenshot ProofAnalytics, dashboards, and outputs — not just descriptions.
⚠️
Honest VerdictIf it flopped, I say so. Abandoned experiments stay published.
5
Experiments PlannedInstagram · YouTube · SaaS · Digital Products · Newsletter
1
In ProgressInstagram AI posts — 7-day live test
£0
Starting BudgetExperiment 01 starts on free tiers only
0
Paid PlacementsNo sponsors. No affiliate deals on experiment tools.
All Experiments

The full experiment log

Every experiment is documented from day one. 🟡 In Progress means you can follow along live. 🔒 Coming Soon means it's planned and scheduled.

Experiment 01
📸

I used AI to run my Instagram for 7 days

No human creativity. Every caption, image, hashtag, and post schedule generated entirely by AI tools. Here's what the algorithm did with it — and whether it's worth trying.

🟡 In Progress ChatGPT · Midjourney · Canva AI · Buffer →
Experiment 02
🎬

Can AI run a faceless YouTube channel?

5 Shorts in 2 weeks, zero original footage, no on-camera presence. Script to upload using VidGen Studio, ElevenLabs, and RunwayML. Tracking views, subscribers, and retention.

🔒 Coming Soon VidGen Studio · ElevenLabs · RunwayML
Experiment 03
🛠️

Building a micro SaaS with AI only — 30-day sprint

Idea to launched product in 30 days. Budget cap: $50. Using Antigravity to build, Claude for copy, Stripe for payments, Vercel to deploy. Tracking every cost and every conversion.

🔒 Coming Soon Antigravity · Claude · Stripe · Vercel
Experiment 04
🛒

Can AI generate sellable digital products?

10 products listed on Gumroad in 30 days — Notion templates, prompt packs, social media kits — all generated by AI. Tracking listings, sales, and which product type actually converts.

🔒 Coming Soon Midjourney · Canva AI · ChatGPT · Gumroad
Experiment 05
📨

AI newsletter — zero manual writing for 30 days

Research, writing, design, and scheduling all handled by AI. Tracking subscriber growth, open rate, click rate, and whether readers can even tell the difference from a human-written one.

🔒 Coming Soon Perplexity · Claude · Beehiiv · Canva AI

How to read the verdicts

✅ Worth It The experiment produced results that justify the time and cost. I'd recommend trying it.
⚠️ Conditional Works under specific conditions — niche, budget, or skill level matters. Read the caveats.
❌ Not Worth It The hype didn't match reality. Here's exactly why it failed and what I'd do differently.

Why I'm doing this

"The internet is full of people telling you AI will change your life. Almost none of them tested it."

I've been building AI systems professionally since 2013 — XR training platforms, enterprise automation, computer vision pipelines. I know what AI can actually do at a technical level. And I've watched the content world fill up with people confidently describing tools they've never seriously used.

These experiments are my answer to that. I pick a hypothesis that matters to real people — can I actually make money from this, save time, replace a workflow — and I run it properly. Fixed timeline, tracked metrics, published results regardless of outcome.

If an experiment fails, I leave it up. The failure report is often the most useful thing I publish. What broke, why it broke, and what you'd need to make it work is more valuable than another success story.

No sponsors involved in any experiment. No affiliate deals on the specific tools being tested. The only incentive is an honest result.

Tools used across all experiments

Every tool linked to its full review — pricing, pros, cons, and real-world examples.

🤖
ChatGPT
Captions · Scripts
 🧠
Claude
Copy · Strategy
 🎨
Midjourney
Images · Assets
 ✏️
Canva AI
Design · Templates
 🎙️
ElevenLabs
Voiceovers
 🎥
RunwayML
Video · B-roll
 🚀
Antigravity
Build · Deploy
 🔎
Perplexity
Research

Follow the experiments live

Get notified when a new experiment launches or a verdict drops. No AI-generated filler — just the experiment updates.