have revolutionized the way in which I program. After I first learned coding back in 2019, I wrote all of the code, character for character. In hindsight, I’m grateful for this experience, as a consequence of the problem-solving mindset coding taught me.
Nonetheless, with AI agents, I consider I’m at the very least 10x more practical as an engineer than I was once. It’s because I’m utilizing AI agents to perform as most of the repetitive, mundane tasks as possible that I used to need to do. Previously, I needed to:
- Manually find the indentation error in Python
- Perform long research by Googling for answers
- Perform large refactors all manually
And an extended list of other tasks I don’t spend much time doing now. Some would say you shouldn’t let AI do all of the be just right for you, since it makes you a worse coder. I’d counter that AI is simply doing the mundane repetitive work, while I can perform essentially the most cognitive-straining tasks, organizing and orchestrating the AI agents.
In this text, I’ll offer you an insight into three of an important techniques I implement to effectively utilize AI agents for coding. I consider these techniques each make me rather a lot more efficient as an engineer, and it also maximizes the potential of AI agents.
I’m at all times in search of ways to be a more practical engineer, so if you’ve more suggestions, I’d greatly appreciate any feedback!
Throughout the article, I’ll confer with each Cursor and Claude Code. I’m not sponsored by any of them, and it’s simply the tools I take advantage of for my agentic coding.
Why you need to use AI agents for coding
I first want to start out off with why you need to use AI agents when coding. The most important argument is you can do more in less time.
AI agents let you do more, in less time
Features that used to take 5 hours to implement can now be realistically implemented and tested in quarter-hour.
Bugs that took you an hour to work out and half-hour to unravel can now simply be solved by your agent in 5 minutes. With the Linear MCP, you possibly can even just copy the difficulty URL and have your agent read the request, find the error, implement an answer, and create a ready-made PR for you.
This is solely amazing. You possibly can read reports on how AI only increases effectiveness in 5% of implementations. Nonetheless, in the event you start using agentic coding tools and the techniques I’m presenting in this text, I’ll guarantee you see a noticeable change in your productivity.
I’m pushing code greater than ever before, and not using a significant increase in bugs. That’s the literal definition of what makes a programmer more practical.
Technique 1: Cursor Planning Mode
The primary and doubtless most vital technique I utilize is the planning mode at any time when implementing a brand new feature. Planning mode is accessible in the entire well-known agentic coding tools, equivalent to Claude Code and Cursor. In Cursor, you possibly can select it by pressing Shift-Tab while prompting your agent.
Planning mode makes it so the agent makes a plan for an implementation, either for a brand new feature, a bug report, or whatever you should perform in your codebase. This plan is made by the agent reading through your request and the code repository so as to determine one of the best approach to unravel your request.
Plan mode then writes the plan to a Markdown file, and might ask you some questions on learn how to implement your request:
- Which language do you wish the implementation in? Python or TypeScript?
- Should the implementation be backwards compatible?
- Monolith architecture, or microservices?
Plan mode is so effective because natural langauge is inherently ambigoutous
That is the rationale we got here up with a programming language: A bit of code is deterministic, and at all times outputs the identical, given the identical input. There isn’t any ambiguity in a chunk of Python code, for instance.
Nonetheless, now we’ve reverted to coding through natural language, since the agents are implementing code as a substitute. Still, we want an approach to eliminate ambiguity, where the plan mode is available in useful.
Technique 2: Cursor Browser Actions
Cursor browser actions is one other very useful technique I’ve began actively utilizing recently. Cursor browser actions allow your Cursor agent to perform actions while working in your implementation. These actions could be:
- Open a URL
- Press a button
- Read console logs
That is an incredible tool to assist the agent solve one-shot problems, all by itself, as a substitute of you having to manually test implementations within the browser and replica console logs containing errors.
As a substitute, you possibly can simply ask Cursor to open the URL, click around, and be sure that there aren’t any issues with the implementation. You essentially make Cursor perform an end-to-end test for you, which is super effective at discovering challenges in your implementation.
It saves a variety of time to prompt Cursor to open the browser, click around, and check the browser logs each time I add a brand new implementation.
Technique 3: Check logs with Claude Code
One other useful technique I utilize heavily is to examine logs with Claude Code or Cursor.
I take advantage of Claude Code to examine logs in the event that they’re unrelated to an implementation I’m working on. For instance, if a document is suddenly stuck within the processing pipeline, without me having modified anything within the relevant code recently.
I take advantage of Cursor typically to examine logs at any time when I’m working on an implementation. I can, for instance, prompt Cursor to send a test event to a Lambda function, and check the CloudWatch logs to be sure that the whole lot was processed as expected.
I previously spent a variety of time within the AWS console, manually navigating to the relevant log group and searching the log group, which takes an extended time. Thus, I began prompting my coding agents to examine the logs for me as a substitute, which saves me 10-60 minutes per day: an incredible efficiency gain.
Having agents check logs is super useful, and has almost no downside. Checking logs is generally a sure bet that you just just need to do. It’s not intellectually difficult or something you should spend time on. Thus, utilizing agents to scan through logs is an excellent priceless use case of coding agents.
Checking logs is an easy and repetitive task: an ideal task to have coding agents perform, when you do more priceless work
If you happen to use agents to scour through your logs, it’s necessary to supply your agent with as much background information:
- What are your log groups called
- What are the table names
- What are the S3 bucket names and prefixes
This protects you a variety of money and time, because your agent doesn’t need to list your entire infrastructure as code and find the relevant service to leaf through. I’ve talked about this idea in my article on effective usage of AGENTS.md.
Conclusion
In this text, I’ve discussed three of the most important techniques I take advantage of to effectively utilize coding agents. I consider that using coding agents and these techniques has made me at the very least 10x more practical as an engineer, from an overall perspective. It’s revolutionized the way in which I work, and saves me incredible amounts of time in my day-to-day work. I consider being effective at utilizing AI tools can be incredibly necessary for the programmers of the long run.
👉 My Free Resources
🚀 10x Your Engineering with LLMs (Free 3-Day Email Course)
📚 Get my free Vision Language Models ebook
💻 My webinar on Vision Language Models
👉 Find me on socials:
🧑💻 Get in contact
✍️ Medium
