I finally went down the rabbit hole.
After resisting for months, I subscribed to Claude Max 20x.
After resisting for months, I subscribed to Claude Max 20x.
S$300/month.
The most expensive personal software subscription I’ve ever paid for.
After 20 years as a full-stack software builder, here’s my honest take:
It’s worth every dollar.
Here’s what changed:
Before:
→ Write ticket in Jira/Asana
→ Schedule meeting to explain the issue
→ Wait for engineer to pick it up
→ Review code
→ Request changes
→ Wait again
→ Repeat
Now:
→ Describe what I need
→ Get working code in minutes (sometimes hours, still very fast)
→ Test it
→ Review it (occasionally)
→ Point out issues
→ Watch it fix itself while I’m testing more
It’s like having 5 software engineers who:
-
Never need context-switching time
-
Don’t require standups to understand the problem
-
Fix bugs the moment you spot them
Yes, you still need to QA.
Yes, you still need to review the code (occasionally).
But the feedback loop went from days → minutes.
For solo builders and small teams, this is a game-changer.
The future of building software isn’t about writing more code.
It’s about QA and reviewing faster.
The only thing I can’t automate now is UX/UI testing.
What’s your experience with AI coding tools? Worth the investment?
#GenAI #SoftwareEngineering #Productivity #ClaudeCode
Enjoyed this? Subscribe for more.
Practical insights on AI, growth, and independent learning. No spam.
More in Vibe Coding
I was doing vibe coding and saw AI generated this code.
Notice anything?
No, vibe coding does not create tech debt.
Bad coders do.
I finally concede that AI is smarter than me.
For 2 years, I held onto reasons like “AI can't solve my kid's homework” or “It can't play tic-tac-toe” to believe I was still smarter.
"Wait, you used to do THAT manually?"
I've been building software for 19 years. Something that would take me 2 months now takes 1 week.
Does Qwen 3.5 live up to the hype?
I tested 9 local LLMs on a Claude Code skill I actually use every day. Not a coding benchmark. A real multi-step agentic task described in natural language a...
I caught Cursor trying to be lazy.
The AI agent couldn’t solve the typing error, so it cast the variable to 'any' to suppress the error, just like a sloppy software engineer would.
I was doing vibe coding and saw AI generated this code.
Notice anything?
I finally concede that AI is smarter than me.
For 2 years, I held onto reasons like “AI can't solve my kid's homework” or “It can't play tic-tac-toe” to believe I was still smarter.
Does Qwen 3.5 live up to the hype?
I tested 9 local LLMs on a Claude Code skill I actually use every day. Not a coding benchmark. A real multi-step agentic task described in natural language a...
No, vibe coding does not create tech debt.
Bad coders do.
"Wait, you used to do THAT manually?"
I've been building software for 19 years. Something that would take me 2 months now takes 1 week.
I caught Cursor trying to be lazy.
The AI agent couldn’t solve the typing error, so it cast the variable to 'any' to suppress the error, just like a sloppy software engineer would.