A startup, Cognition, recently launched a new product calledDevinofAI Assistant, designed to assist software engineering teams with coding and other development tasks. Unlike existing AI coding assistants, Devin can be programmed to execute end-to-end software projects, including deploying applications, fixing bugs, and learning new technologies, with humans playing a supervisory and mentoring role.
Devin executes multi-step workflows based on user needs while keeping work on track. Engineers can observe its progress in real time and jump to instructions to fix errors as they are discovered. This enables teams to outsource some of their work toAI AssistantYou'll be able to focus on more creative work.
Devin's performance in the SWE bench test
According to the demo, Devin can handle a variety of tasks including deploying websites, debugging code, generating images of hidden information, and training computer vision models. In software engineering benchmarks, it was able to independently complete the 13.86% case, much higher than other large language models.
While core technical details were not disclosed, Cognition says Devin stems from advances in its long-term reasoning and planning research. The tool is currently in internal beta, and interested users can request an early trial. Wider access may be available in the future.
Cognition hints that coding is just the beginning, which means its AI assistant may be rolled out to more areas. The company plans to explore empowering multiple industries by leveraging the benefits of AI's reasoning across domains.
The emergence of Devin brings a new AI collaboration experience to software developers. By supervising AI systems to handle tedious tasks, engineers are able to focus on innovative tasks, which is expected to increase productivity. However, the technology is still in its early days, and its maturity and effectiveness are subject to further industry evaluation.