Cognition AI unveils Devin, the world’s first AI software engineer that can write entire software projects from scratch
The world of coding is experiencing a seismic shift with the advent of artificial intelligence (AI) reaching unprecedented capabilities. It’s been over a decade since Marc Andreessen’s influential essay “Why software is eating the world,” and now, it seems software engineers are on the brink of being overtaken by AI, particularly with the emergence of large language models like ChatGPT, which are pushing the boundaries of what was once deemed impossible.
On Tuesday, an AI startup called Cognition AI introduced Devin, an autonomous AI agent that “can plan and execute complex coding tasks, fix bugs, and collaborate with users in real-time.” Devin is being called the world’s inaugural AI software engineer because of its ability to independently build entire software projects from scratch, guided only by simple text prompts.
Unlike existing coding assistants that suggest code or accomplish minor tasks, Devin can plan and execute complex coding projects with hundreds of steps autonomously. Devin can navigate through intricate coding tasks comprising hundreds of steps, all while learning, troubleshooting bugs, and engaging in real-time collaboration with users.
In a demo, Devin was able to build entire websites and applications in under 10 minutes and even complete genuine freelance assignments posted on platforms like Upwork, all without human intervention. Furthermore, on a coding benchmark, Devin achieved an impressive 13.86% resolution rate for real-world GitHub issues, significantly surpassing the previous state-of-the-art benchmark of 1.96%.
Notably, Devin has also excelled on the SWE-Bench coding benchmark, demonstrating its capability to tackle practical engineering interviews at top AI firms and execute real jobs on Upwork with finesse.
In a post on X (formerly Twitter), Cognition AI founder and CEO Scott Wu said:
“Devin is the new state-of-the-art on the SWE-Bench coding benchmark, has successfully passed practical engineering interviews from leading AI companies, and has even completed real jobs on Upwork. Devin is an autonomous agent that solves engineering tasks through the use of its own shell, code editor, and web browser.”
Cognition AI’s unveiling of Devin signifies more than just an advancement in coding assistance; it heralds the dawn of an era where AI could serve as genuine software engineers. If Devin lives up to its hype, it could offer a glimpse into a future where anyone, regardless of coding proficiency, can leverage AI to materialize concepts effortlessly.
However, reactions to this development are mixed. While some are thrilled about the prospect of heightened productivity, others express concerns about the potential impact on human software engineers, highlighting the nuanced discussions surrounding the integration of AI in the field.