In a blog post, Cognition's founder and CEO, Scott Wu, explained that Devin can access common developer tools within a sandboxed compute environment to plan and execute complex engineering tasks. The AI software engineer can handle a range of tasks from deploying and improving apps/websites to finding and fixing bugs in codebases. Despite the impressive capabilities, the core technology behind Devin remains undisclosed. Cognition is currently offering early access to Devin to select users and plans to open up broader access at a later stage.
Key takeaways:
- Cognition, a recently formed AI startup, has announced a fully autonomous AI software engineer called “Devin” which can handle entire development projects end-to-end.
- Devin is capable of handling a range of tasks including common engineering projects like deploying and improving apps/websites end-to-end and finding and fixing bugs in codebases to more complex things like setting up fine-tuning for a large language model.
- In the SWE-bench test, Devin was able to correctly resolve 13.86% of the cases end-to-end – without any assistance from humans, outperforming other AI models.
- Cognition has not shared how exactly it has achieved this feat and whether it is using its own proprietary model or that from a third party, but it does note that the work is the result of its “advances in long-term reasoning and planning.”