What is Devin
Devin is the world’s first fully autonomous robot launched by artificial intelligence startup CognitionAI software engineer agenthas strong programming and software development capabilities, and can assist in multiple aspects or complete software development tasks completely independently. In the SWE-bench benchmark test, Devin’s performance in solving real-world problems far exceeds that of AI models such as GPT-4 and Claude 2.
Although Devin’s development company Cognition has only been officially established for two months, the team members have rich experience in cutting-edge work in AI and have multiple International Informatics Olympiad (IOI) gold medals. They have already received investment from Peter Thiel’s Founders Fund. of US$21 million in Series A financing.
Devin’s main features
- Learn new technologies independently: Devin is able to expand its skill set by reading documentation and code to learn technologies with which it is unfamiliar.
- End-to-end build and deployment procedures: Devin is able to understand the entire software development process, from front-end design to back-end deployment, and even getting the application live. This means it can build a website, game, or other software project from scratch and handle the associated workflow.
- Find and fix bugs independently: Devin has excellent debugging capabilities and can find and fix errors in the code. Even problems that the developers themselves have not noticed can be found and solved by it.
- Train and fine-tune AI models: Devin can not only handle regular programming tasks, but also help train and fine-tune other AI models, showing deep application capabilities in the field of artificial intelligence.
- Fix open source library: Devin is able to understand and solve problems in the open source community, such as fixing known bugs or implementing new feature requests.
- Contribute to mature production libraries: Devin is able to contribute to already mature production libraries, such as fixing known bugs or adding new features.
Devin’s performance comparison
In the SWE-bench benchmark, which requires agents to solve actual GitHub issues found in open source projects such as Django and scikit-learn, Devin was able to correctly handle 13.86% of the issues. This score is significantly higher than the previous technical level of 1.96%, showing Devin’s huge advantage in understanding and solving practical programming problems.
Comparing other AI models: Devin’s performance far exceeds that of other well-known AI models, such as GPT-4 and Claude 2, which generally have lower accuracy in the same test.
How to use Devin
Devin is currently still in closed beta, please visitCoginition’s official websiteFor more information, users who want to experience it first can fill inDevin’s internal beta application form。
Source link
