Jump to content
Main menu
Main menu
move to sidebar
hide
Navigation
Main page
Recent changes
Random page
freem
Search
Search
Appearance
Create account
Log in
Personal tools
Create account
Log in
Pages for logged out editors
learn more
Contributions
Talk
Editing
Openai/692ad5f8-1fbc-8002-bc75-245a4b2194e9
(section)
Add languages
Page
Discussion
English
Read
Edit
Edit source
View history
Tools
Tools
move to sidebar
hide
Actions
Read
Edit
Edit source
View history
General
What links here
Related changes
Special pages
Page information
Appearance
move to sidebar
hide
Warning:
You are not logged in. Your IP address will be publicly visible if you make any edits. If you
log in
or
create an account
, your edits will be attributed to your username, along with other benefits.
Anti-spam check. Do
not
fill this in!
===== I started using AI to code in August 2025 and it’s honestly one of the most empowering and crazy things I’ve experienced. ===== When I was about 16 (like 12 years ago) I messed around with HTML, CSS, and a little PHP. I even made some money online as a kid on forums like DigitalPoint. I learned some fundamentals and got comfortable with MySQL, but life went a different direction and I didn’t stay with coding. When I tried getting back into it more recently, it felt like everything got way more complex and I didn’t even know where to start. Then I heard about “vibe coding.” I paid for Cursor and started throwing ideas at it just to see what would happen, and it actually made some pretty cool stuff. That turned into me building little tools for my salvage yard job to make my life (and coworkers’ lives) easier and cut down on paper. At first my process was rough. I’d have the AI build something, upload it with FTP to test, find errors, paste the errors back, and do that over and over. I do understand just enough to follow what the AI is doing, and I read fast, so a lot of the time I could catch when it was hallucinating or stuck in a weird loop. But I knew there had to be a better workflow. Learning Docker changed everything. Once I got my site running locally for testing, while still having staging/production online, it got way easier to move fast without breaking everything. Later I realized Cursor wasn’t the best fit for me, so I switched to Claude. I learned how to run Claude in the VS Code terminal using Git Bash and that was a huge upgrade. I used Sonnet almost all the time because it matched my pace. Opus was powerful but it could create spec docs so long that I’d get overwhelmed. Over time I learned SSH too, and had the AI help me build simple scripts to deploy to staging and another to deploy to production. Now I can build locally, test on staging, and deploy without digging through folders and doing everything manually. That workflow has been awesome. Since August I’ve built 14 tools for my salvage yard job. Five of them are actually used in daily operations now by more than 10 users. I know the codebase isn’t perfect and there’s definitely spaghetti code in places, but seeing people use things I made in real life is still kind of unreal. Now with Opus 4.5 and MCP dev tools, it feels like what used to take me weeks can happen in days. I rebuilt a tool that I previously struggled with (multiple attempts, lots of frustration) and this time it got rebuilt and working in under a 5 hour session. That genuinely shocked me. My wife gave me an idea that was supposed to be a simple Christmas list web app. Just a clean way to list people, add gift ideas, compare prices at different stores, and decide later. Of course it turned into a multi-tenant gift planner with stages, budgets, tracking, and a bunch of extra features. I also learned a new approach that helped a lot. Instead of just asking the AI to start coding, I spent time talking through the full concept first. Workflow, login/security, sessions/cookies, UX, edge cases, all of it. Then I had it write a spec doc in sections with an index, so it could delegate pieces to sub-agents and build it more cleanly instead of one giant spaghetti blob. It built the whole thing in one shot, but at first it didn’t work and there were errors everywhere. The difference was that with MCP dev tools, it could actually verify stuff, find patterns, assign fixes, and bring it into shape without me guessing. After a couple more hours it was working, and I’ve been improving it since. This AI stuff is honestly the wildest tech shift I’ve seen in my life. I’m kind of addicted to building now. And I can’t even imagine what someone who really understands software can do with these tools if they use them at full power. I’m really excited to see what’s next.
Summary:
Please note that all contributions to freem are considered to be released under the Creative Commons Attribution-ShareAlike 4.0 (see
Freem:Copyrights
for details). If you do not want your writing to be edited mercilessly and redistributed at will, then do not submit it here.
You are also promising us that you wrote this yourself, or copied it from a public domain or similar free resource.
Do not submit copyrighted work without permission!
Cancel
Editing help
(opens in new window)