Yes, you heard it right Devin AI: A Devil For Software Engineers. A US-based startup named Cognition launched their AI-powered Tool Called Devin AI. They claim it to be the World’s First AI-powered Software Engineer, which can easily solve engineering tasks with the help of its own shell, code editor, and web browser. Here is what the Startup tweeted about the Tool:
Today we're excited to introduce Devin, the first AI software engineer.
— Cognition (@cognition_labs) March 12, 2024
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… pic.twitter.com/ladBicxEat
What Devin Can Do?
Cognition mentioned that Devin AI has its own shell, code editor, and web browser. In a demonstration shown by the company Devin is using the browser to pull up API (Application Programming Interface) documentation to read and learn how to plug into each of the API’s. An API helps in the communication of two or more computer programs to communicate with each other. When the AI agent throws an error, it automatically adds a debugging print statement to the main code within the code editor interface and re-runs the code.
On their YouTube Channel, they showed a demonstration of various use cases of the tool. This includes building and deploying apps, finding and fixing bugs in codebases, and even fine-tuning other AI models.
Is Devin AI Accurate?
The company said that they tested Devin on an SWE-bench, it is a benchmarking platform that tests agents to resolve real-world issues on open-source projects on GitHub. According to Cognition, Devin correctly resolves13.86 percent of the issues end-to-end. And When It was compared with OpenAI’s ChatGPT 4 it was only able to resolve 1.74 percent issues. The previous best score was of Anthropic’s AI model called Claude 2 which was able to resolve 4.80 percent of the issues found.
The above Report shows the SWE-bench performance Report of Devin As Compared to Other AI Models. Devin has the highest accuracy rate among all of them.
How Much Potential Does Devin Have?
Microsoft has AI-powered tools for developers, including the GitHub Copilot which is mainly a code completion tool. GitHub Copilot is made to assist programmers who require human interference and auto-completes chunks of code into prompts into runnable code entered by the user. But Devin does not require any kind of human interference and complete codes end-to-end on its own.
How to Use Devin AI?
Devin is currently on early access and individuals who wish to use Devin for engineering work can raise a request to the company’s website to get early access to the AI-powered coding agent.
- The JS Developer’s Podcast [EP: 2] Variables and Data Manipulation - October 15, 2024
- YouTube Channels to Learn Coding: Top 9 Picks That Will Make a You Master - October 10, 2024
- The JS Developer’s Podcast [EP: 1] Introduction to JavaScript - September 27, 2024