You're wasting your time coding with AI.

@julianisdoing

4.8K followers · usually 871 views a post

Views

4.7K

Likes

104

Comments

23

Shares

10

SoundTikTok sound · original sound - julianisdoing

Promoted product

Cursorcursor.com

Built to make you extraordinarily productive, Cursor is the best AI coding agent.

67 videos47 creators

Why it worked

The video provides practical, actionable advice by recommending specific tools (AI models) for a popular developer tool (Cursor), addressing a common pain point (AI inconsistency) and offering a clear solution with a tiered approach.

Summary

The creator explains that using the wrong AI model for coding tasks leads to wasted time and inconsistency. They recommend a stack of three AI models for the Cursor IDE: Gemini 2.5 Pro for long context, GPT-5 High for complex problems, and Claude 4 Sonnet 1m as a daily driver, emphasizing the importance of choosing the right tool for the job.

Structure

1Introduction to the problem of AI inconsistency in coding
2Explanation of why inconsistency occurs (wrong model for the job)
3Recommendation of three AI models for Cursor: Gemini 2.5 Pro, GPT-5 High, and Claude 4 Sonnet 1m

Original caption

Here are my top 3 models on cursor

More from @julianisdoing

Product placement

The video discusses and lists three AI models (Gemini 2.5 Pro, GPT-5 High, and Claude 4 Sonnet 1m) that can be used within the Cursor IDE, positioning Cursor as the platform for utilizing these models.

Call to action

Don't stick to one model. Start with Cloud as your default and use GT5 if you're truly stuck.

On-screen text

YOU'RE WASTING YOUR AI NEXT IT'S MAKING DUMB MISTAKES THAT INCONSISTENCY IS TIME WASTER RANDOM USING THE WRONG JOB Hundreds of hours I've learned THREE USE FIRST GEMINI 2.5 PRO 3 best models on Cursor: 1. Gemini 2.5 Pro 2. 3. THIS IS MY LONG CONTEXT KING MASSIVE FILES TO THIS IS THE IT'S AN ABSOLUTE WORKHORSE AND BARELY EVEN CLOSE TO ITS CONTEXT LIMITS SECOND GPT5 HIGH 3 best models on Cursor: 1. Gemini 2.5 Pro 2. GPT 5 High 3. HEAVY DUTY REASONER EXPENSIVE SO I REALLY ONLY USE IT FOR COMPLEX WHERE I REALLY TO THINK AND THIRD CLAUDE 3 best models on Cursor: 1. Gemini 2.5 Pro 2. GPT 5 High 3. Claude 4 Sonnet 1m THIS IS MY DAILY DRIVER PERFECT BALANCE OF SPEED AND INTELLIGENCE GET 90% OF MY WORK DONE SO THAT'S MY USING CURSOR ONE MODEL AS YOUR DEFAULT AND USE GPT5 STUCK YOUR TOOLS AND START WINNING

Transcript

You're wasting your time coding with AI. One minute your AI is a genius and the next it's making dumb mistakes. That inconsistency is the single biggest time wasting. But it's not random. It's because you're using the wrong model for the job. After hundreds of hours, I've learned you only need three. Here's what I use. First, Gemini 2.5 Pro. This is my long context game. If I have massive files to read through, this is the one I use. It's an absolute workhorse and barely hallucinates even close to its context limits. Second, GPT-5 high. This is my heavy duty reviewer. It's slow and expensive, so I really only use it for complex problems where I really need the AI to think. And third, Cloud 4 Sonic 10. This is my daily driver. It has the perfect balance of speed and intelligence. And I probably get 90% of my work done like this. So that's my stack when I'm using cursor. Don't stick to one model. Start with Cloud as your default and use GT5 if you're truly stuck. This is how you stop fighting your tools and start winning. See you tomorrow.