Meet the Future of Software Engineering: Introducing Devin - The First of its Kind AI Software Engineer

Welcome to the dawn of a new revolution - Devin, the world's first completely autonomous AI software engineer. Devin is not just a tool, but a teammate, ready to work alongside you to build and execute complex tasks or to take on tasks independently.

What Is Devin?

Devin is a groundbreaking development in the world of software engineering. It represents a significant leap forward in AI's capability to undertake complex tasks that require long-term reasoning and planning.

From executing thousands of decisions to learning from mistakes and recalling relevant context at every step, Devin's capabilities extend far beyond those of any existing AI tools or software.

Devin's Tools

To make Devin as effective and versatile as possible, it has been equipped with all the tools a human software engineer would need. This includes a shell, code editor, and browser, all within a sandboxed compute environment.

But Devin is not just a tool, it is a collaborator. It provides real-time updates on its progress, actively accepting feedback and working with you on design choices.

What Can Devin Do?

Let's take a brief look at some of Devin's capabilities:

1. Learning new technologies: After reading a blog post, Devin can run ControlNet on Modal to create images with hidden messages.
2. Building and deploying apps from start to finish: Devin can create an interactive website that simulates the Game of Life. It can add features requested by the user and deploy the app to Netlify.
3. Finding and fixing bugs autonomously: Devin can help maintain and debug open source programming books.
4. Training and fine-tuning its own AI models: Given only a link to a research repository on GitHub, Devin can set up fine-tuning for a large language model.
5. Addressing bugs and feature requests in open source repositories: With just a link to a GitHub issue, Devin can do all the setup and context gathering that is needed.
6. Contributing to mature production repositories: As part of the SWE-bench benchmark, Devin solved a bug with logarithm calculations in the sympy Python algebra system. Devin set up the code environment, reproduced the bug, and coded and tested the fix on its own.
7. Completing real jobs on Upwork: Devin can write and debug code to run a computer vision model, sample the resulting data, and compile a report at the end.

How Does Devin Perform?

Devin's performance was evaluated on SWE-bench, a challenging benchmark that requires agents to resolve real-world GitHub issues found in open source projects like Django and scikit-learn.

Devin successfully resolved 13.86% of the issues end-to-end, significantly surpassing the previous state-of-the-art of 1.96%. Even when given the exact files to edit, the best previous models could only resolve 4.80% of issues.

About Cognition

Cognition is an applied AI lab dedicated to reasoning. The team is building AI teammates with capabilities far beyond today's existing AI tools. By solving reasoning, they hope to unlock new possibilities in a range of disciplines.

Cognition is well-funded, having raised a $21 million Series A led by Founders Fund. They have also received support from industry leaders such as Patrick and John Collison, Elad Gil, Sarah Guo, Chris Re, Eric Glyman, Karim Atiyeh, Erik Bernhardsson, Tony Xu, Fred Ehrsam, and others.

Hire Devin

Devin is currently in early access as capacity is being ramped up. To start using Devin for engineering work, reach out to Cognition or get in touch at [email protected].