Jules is a new AI coding assistant from Google, built on the Gemini 2.0 model. It is designed to automate coding tasks such as bug fixes and code modifications within GitHub workflows, aiming to help streamline development.

How does Jules compare to GitHub Copilot?

Currently, GitHub Copilot offers a more seamless and integrated experience within the IDE, providing real-time code suggestions and autocompletion. Jules integrates with GitHub but operates as a separate interface and can be slower and less reliable, particularly with complex tasks.

How reliable and fast is Jules?

Jules performs well with simple, repetitive tasks but struggles with complex code changes, sometimes getting stuck in loops and needing supervision. Its speed is slower than competitors, which can impact user experience.

How does OpenAI Codex compare to Jules?

OpenAI’s Codex often excels in generating complex code and understanding advanced instructions compared to Jules, owing to its broader training data and more mature development.

What is the future potential for Jules?

With further refinement, Jules could become a valuable tool for automating repetitive development tasks. Google's integration of Gemini 2.0 and ongoing investment indicates future improvements, but reliability and speed issues need to be addressed for widespread adoption.

Should developers use Jules now?

As of now, developers may find more dependable assistance with established tools like GitHub Copilot or OpenAI Codex. Jules offers innovative features and could be useful for experimentation, but it's still in beta and faces challenges with complex and time-sensitive tasks.

a new contender in ai-powered coding assistants

Key Takeaways

Jules is a new AI coding assistant from Google, built on the Gemini 2.0 model.
Jules excels at simple tasks but struggles with complex ones, often getting stuck in loops.
Jules is slower than established tools like GitHub Copilot.
GitHub Copilot offers a more seamless and integrated experience.
OpenAI’s Codex often excels in generating more complex code snippets.
The competition between Google (Jules) and OpenAI/Microsoft (Codex/Copilot) is a key dynamic in the AI coding assistant space.

First Impressions and Setup

Recently I took the opportunity to explore various AI-powered coding assistants. Google’s recent introduction of Jules, an experimental AI agent built on the Gemini 2.0 model, piqued my interest. Designed to automate tasks like bug fixes and code modifications within GitHub workflows, using simple language and instructions, Jules aims to significantly streamline development processes using Googles vast resources and Agentic AI capabilities.

In this review, I’ll share my firsthand experience with Jules, comparing it to established tools like GitHub Copilot, OpenAI’s Codex and discuss its potential impact on software development, which is being increasingly shaped by trends (like those I speculated on in 4 key tech trends for 2025).

Integrating Jules into my existing GitHub workflow was straightforward. Single sign-on with Github and repository access was intuitive, and within minutes, Jules was ready to assist. Its seamless integration with the existing repositories allowed for a smooth transition, enabling me to dive into some complex tasks without the usual overhead of manual configurations.

Performance and Reliability

Jules demonstrated real promise in handling simple tasks. For instance, it successfully converted a JSON export from a blog into seperate Markdown files without any supervision. This task, which often requires manual effort copy/past was completed in less than 4 minutes - showcasing Jules’s potential in automating routine processes.

However, when faced with more complex scenarios, Jules encountered challenges. When trying to convert the underlying framework for my blog, things hit a snag. Three out of four tasks resulted in the AI getting stuck in loops, necessitating constant input to proceed. One task failed outright, highlighting the current limitations in handling intricate coding problems. This experience aligns with reports from others (via reddit) who have noted similar issues with Jules’s reliability in complex tasks.

Speed and Performance

Another notable drawback was Jules’s speed. It was SLOW. While it’s understandable (sinceJules is in its beta phase) the AI operated sluggishly and it really impacted the experience of the tool - constant refreshes and clear caches to make it catch-up. The performance issues were more pronounced than I anticipated. I couldnt help but think that a fallback to established tools like GitHub Copilot would result in more mileage (depending on your use-case).

Comparison with Other AI Tools

Github CoPilot

GitHub Copilot has been a staple in the developer community, offering real-time code suggestions and autocompletion. In my experience, Copilot has consistently provided accurate and context-aware code snippets, enhancing my overall coding efficiency.

In terms of outcomes for the real-world use case I threw its way, Jules while promising, currently falls short in matching something like Copilot’s capabilities IMO. The suggestions and plans were great and accurate, but its inability to handle complex tasks without supervision was a significant drawback for me. The question remains when it comes to leaning on AI as a coding assistant or pair programmer - how much ground is still left to cover?

In terms of usage, Copilot brought a more seamless and integrated experience within the IDE, providing real-time suggestions as you type. Jules, on the other hand, is driven by instructions and an interface outside of your coding environment - and it often requires explicit prompting, approval and interaction (such as specifying tasks, validating commands or providing context through comments). This difference in UX makes Copilot feel more like a natural extension of the coding process, while Jules can feel more like a separate tool that requires more deliberate engagement.

OpenAI Codex

I also recently tried another tool - OpenAI’s Codex - im my experience, with its broader training data and more mature development it often exceeded that bar in generating more complex code and handling a wider range of high level instruction scenarios compared with Jules.

The competition between Google (Jules) and OpenAI/Microsoft (Codex/Copilot) is a key dynamic in the AI coding assistant space. Google’s entry with Jules signals a strong commitment to this market, and the ongoing advancements in both models will likely shape the future of software development. The race is on to see which company can provide the most powerful and user-friendly AI coding tools.

Potential and Future Outlook

Despite its current limitations, Jules holds promise. Google’s integration of the Gemini 2.0 model suggests a commitment to advancing AI in software development. With further refinement, Jules could become a valuable tool for automating repetitive tasks, allowing developers to focus on more strategic aspects of coding. However, addressing the reliability and performance issues is crucial for Jules to compete effectively in the market and more up-take in the market makes the conversations around ethics of AI code all the more meaningful.

Summary

While Google’s Jules introduces innovative AI-assisted coding features, it currently faces challenges in performance and reliability. For developers seeking a dependable coding assistant, established tools like GitHub Copilot remain the preferred choice. As Jules evolves, it will be interesting to see how it addresses these challenges and integrates into the broader development ecosystem, potentially changing how developers approach coding assistance with AI and large language models.