Clever Programmer
AI coding assistants and app builders with a focus on live workflows and speed.
Nutrition Label
Rafeh Qazi delivers high-energy, live-coding demonstrations of the latest AI development tools. While his willingness to show real-time errors adds authenticity, the analysis often leans on hype and sales funnels rather than deep technical benchmarking.
Strengths
- +
- +
- +
Notes
- !Critiques of trending models often serve as funnels for specific affiliate tools.
- !Technical claims frequently rely on hyperbolic binaries rather than nuanced testing.
Rating Breakdown
Breakdown across the key dimensions we rate. Methodology →
Recent Videos
This is INSANE! You won't believe which LLM model WON 👀
Deepseek vs. O3-mini vs. Qwen 2.5 max vs. Claude!! (Real World Examples)
Cursor Has SHARED AI Knowledge Base 🤯
Deepseek... More like Deep SUCKS. My honest thoughts...
Google Wrote 25% of It's Code with AI
Reacting to Why Are AI Generated Videos So Terrifying?
JavaScript Tutorial 2025 for Beginners - Full Course in 10 Hours
New Claude AI Can Control Your Computer Now!
Major GPT Upgrade | *CANVAS* Crash Course
Your CODE is broken. Until you DO THIS…
10 AI Tools - You Must Know as a Developer in 2024
The resume that got a software engineer a $300,000 job at Google
Stop watching coding tutorials. Do this instead.
Insecurity is stopping you from becoming a better developer
Is Freecodecamp Worth It? My Honest Reaction
Why this rating
Evidence receipts showing why each dimension is rated the way it is.
“I can literally just go here, hit Command K... type in '@' and you can see right here it says 'Docs'.”[0:25] →
Presenter demonstrates the exact UI workflow (shortcuts and menu selection) in real-time, validating the feature's existence.
“We really fucked up. And so I want to share that story with you... this one story will tell you what it takes for a junior developer to act like a senior developer.”[00:05] →
“This is powerful because ChatGPT's knowledge cutoff is [dated]... but with this, you can feed it the latest documentation.”[1:15] →
Correctly identifies the underlying utility of the feature (RAG/context injection) to solve training data staleness.
“Okay, wait... that actually worked perfectly. I can't even lie.”[01:12] →
The creator immediately contradicts the 'Deep SUCKS' title when the first test (Snake game) works flawlessly, revealing the title as clickbait.
“DeepSeek is good for snippets, but for full apps...”[05:45] →
Compares a raw LLM chat interface (DeepSeek) to a specialized AI app-builder (Lovable) without acknowledging the category difference, creating a skewed benchmark to favor the affiliate product.
“So I did a heart surgery on a fucking frog one time... and then I tried this thing on your mother and your mother fucking died... well, it worked on a frog.”[04:05] →