Devin, the world’s first AI software engineer, announced that it can write, code, and create using a single prompt

Cognition, a startup from the US, introduced world’s first AI software engineer Devin AI, a unique artificial intelligence tool that revolutionizes how code is handled—debugged, written, and deployed. According to Cognition, backed by the Founders Fund, Devin AI stands out as it can transform a basic command into a fully operational website or software program.

Does Devin AI succeed in engineering interviews?

Today, we’re thrilled to present Devin, the pioneer AI software engineer. Devin excels on the SWE-Bench coding benchmark, has aced hands-on engineering interviews at major AI firms, and has accomplished tasks on Upwork. Operating independently, Devin tackles engineering challenges using its own shell, code editor, and web browser,” shared Cognition on Twitter, also known as X.

Devin AI’s performance on the SWE-bench coding test

The company mentioned that Devin AI participates in the SWE-bench coding benchmark. This benchmark consists of 2,294 real software engineering problems sourced from GitHub issues and their associated pull requests.

What can Devin AI do?

Devin AI not only assists with coding tasks but also handles the entire software development and release process, a capability beyond what large language models like Google’s Gemini or OpenAI’s ChatGPT can offer.

What sets Devin apart is its remarkable ability to anticipate and plan intricate tasks. It can make numerous decisions, learn from errors, and improve continuously. Moreover, it comes equipped with essential tools like a code editor and browser, readily accessible.

Devin is recognized as the most advanced solution for assessing software engineering tasks, based on the SWE-bench coding benchmark. It outperformed other solutions significantly when tested against a standard set of software engineering problems. In practical engineering interviews held by leading AI companies, this AI tool demonstrated strong performance. These interviews likely included tasks relevant to AI and software engineering, and Devin met the challenges admirably.

So, what can Devin do? Essentially, almost anything you ask of it. From mastering new technologies, creating and launching apps from scratch, to tracking down and resolving tricky bugs in code, Devin handles it all. It’s even capable of training its own AI models and addressing issues in open-source projects.

How does Devin AI function?

Devin operates independently using its own code editor, command line, and browser. It systematically implements software by breaking tasks into manageable steps and can handle intricate engineering projects involving hundreds to thousands of decisions.

You might also be interested in – Tata Motors inks deal with Tamil Nadu government to establish ₹9,000-crore vehicle manufacturing plant