What AI tools are you guys using to code?

  • Thread starter Thread starter Deleted member 752298
  • Start date Start date
jakubbanan said:
I used only Cursor PRO - working like diamond doesnt need anything like API with ChatGPT, or Claude - in PRO version you can do anything you want :)

PS. If somebody want Cursor PRO - I have good offer :) Very cheap and working very well :)
Click to expand...
You can't sell here mate. Just make it public.

I am using Claude and Gemini 2.5 Pro.

I tried Cursor as well, but paying for the Pro subscription just to use my own API keys anyway is ridiculous, so I just use Claude API by myself.
 
Advertise on BHW
nakamura said:
You can't sell here mate. Just make it public.

I am using Claude and Gemini 2.5 Pro.

I tried Cursor as well, but paying for the Pro subscription just to use my own API keys anyway is ridiculous, so I just use Claude API by myself.
Click to expand...
Sorry, my bad :(

On Cursor PRO you dont need your own API keys ... you dont need charge your chatgpt api account, or claude to use it. Only one thing that you must is PRO version, and you can code anything you want
 
Claude for coding, V0 for building, Perplexity for research and questions.
 
nakamura said:
You can't sell here mate. Just make it public.

I am using Claude and Gemini 2.5 Pro.

I tried Cursor as well, but paying for the Pro subscription just to use my own API keys anyway is ridiculous, so I just use Claude API by myself.
Click to expand...
I can't send you a dm... Can you send me please DM
 
I use Cursor, Gemini 2.5, DeepSeek, and ChatGPT, in that order. Sometimes I'll run through more than one just to see which outputs the best, or with debugging one may do well while others don't.
 
xReminisce said:
This
I use multi, sometimes 1 will figure out the issue right away, another time it be stuck in loop being completely useless for ages.

Currently have at hand
chatgpt
supergrok
github copilot

considering dropping github to give claude a go since github honestly been the weakest of them all.
Supergrok makes quite alot of mistakes but their reasoning is nice
Click to expand...
What good is a “nice reasoning” if the end result is a mistake?
 
have multiple tools at hand.
Use claude/gemini as the primary, when it stuck after trying multiple times, just head to other tools for inspiartion.
 
Been mostly using ChatGPT-4 and Claude lately both solid for PHP and JS stuff.

Claude feels sharper with JS logic and React patterns, while ChatGPT still wins for debugging messy legacy code.

Also starting to test Cursor and Junie inside PHPStorm. Cursor’s inline tools are fast and helpful when jumping between files. Definitely worth trying if you do plugin dev regularly.
 
I've tried chatgpt, claude and cursor.... i'll rate it as
1 - cursor
2 - claude
3 - chatgpt

it's all quite close
 
Used to use ChatGPT & Claude. They stopped being good, and both confidently give wrong answers every time. Even for basic stuff these days... Doesn't help that I stopped coding in Python/JS and switched to Lua mainly for the kinds of coding projects that I need to get done.

Going back to coding by hand (and actually being able to link up my personal libraries rather than risk ChatGPT training off of the code that I don't Open Source... Yes, some of my projects are obscure enough that this is a genuine concern for me) has actually been kinda nice. Brain feels faster, and surprisingly I feel like the projects get done a lot faster too. Feels like I spent more time debugging AI than I did solving problems with it.

If you haven't coded by hand in a while (or ever lol) then take a break from AIs. Your typing speed will increase, you won't have that brain delay (that delay when you're waiting for copilot to suggest something so you can tab-complete it), and you'll probably surprise yourself with how easy problems can be when you program the right way.

GPT3.5 Completions was peak AI coding BTW, its all been downhill from there ever since all of the AI companies decided the only things worth developing were chat models and not completion models... /sigh, even for writing, the completion models were far superior...
 
GitHub Copilot has definitely been a go to for quick code completion especially in JavaScript. It's really good for boilerplate stuff. For PHP and WordPress plugin development specifically we've found Cody by Sourcegraph surprisingly helpful. It seems to have a better grasp of larger codebases and can sometimes suggest genuine improvements which is awesome. ChatGPT is great for brainstorming and understanding complex concepts though I wouldn't fully trust it to write production code without careful review.
 
Sounds crazy, but I'm actually using Gemini together with GitHub Copilot. I find that Gemini seems to work better than ChatGPT.
 
Been using ChatGPT and Copilot side by side, ChatGPT’s better for explaining and improving code, while Copilot shines in autocomplete. For PHP + WP dev, ChatGPT has been a huge time-saver for debugging and refactoring.
 
nakamura said:
You can't sell here mate. Just make it public.

I am using Claude and Gemini 2.5 Pro.

I tried Cursor as well, but paying for the Pro subscription just to use my own API keys anyway is ridiculous, so I just use Claude API by myself.
Click to expand...

Can't you use the free version of Cursor with your own API keys? You used to be able anyway?
 
Advertise on BHW
You must log in or register to reply here.
Sign up now!

Main Menu

Home Main Forum List Black Hat SEO White Hat SEO BHW Newbie Guide Blogging Black Hat Tools Social Media Downloads

Marketplace

Account Selling Content / Copywriting EBooks / Methods Hosting Misc Proxies For Sale SEO - Link building SEO - Packages Social Media Social Media - Panels Web Design More

Making Money

Affiliate Programs Hire a Freelancer IM Journeys Making Money Pay Per Click Site Flipping Want to Buy

BlackHat World

Dispute Resolution Forum Suggestions & Feedback Introductions Lounge
Back
Top