US-based Cognition launches world’s first AI Software Engineer: Should you be worried?

So, will AI take over our jobs or not? 

This question looms large, particularly after massive advancement of AI and the launch of various tools (ChatGPT, Claude, Bard), fueling debates about the trajectory of automation and its impact on employment.”

Cognition, a US-based applied AI lab focused on reasoning, gives us a little hint about the future. 

Meet Devin, the world’s first fully autonomous AI software engineer.

According to Cognition, “Devin is a tireless, skilled teammate, equally ready to build alongside you or independently complete tasks for you to review.”

“With Devin, engineers can focus on more interesting problems, and engineering teams can strive for more ambitious goals,” says the company. 

Devin’s abilities: Executing complex tasks to collaborating with users

Cognition says Devin can plan and execute complex engineering tasks requiring thousands of decisions, recall relevant context at every step, learn over time, and fix mistakes.

Devin features developer tools that a human would need to do their work, including the shell, code editor, and browser within a sandboxed computing environment. 

With an ability to collaborate, Devin can report on its progress in real-time, accept feedback, and work together with users through design choices as needed.

To make it simple, here’s what Devin is capable of: 

• Learn new technologies
• Build and deploy apps end-to-end
• Can find and fix bugs in codebases autonomously 
• Can train and fine-tune its own AI models
• Address bugs and feature requests in open-source repositories
• Contribute to mature production repositories

Additionally, the company gave Devin a real job on Upwork writing and debugging code to run a computer vision model. 

In fact, Devin underwent evaluation on SWE-bench, a rigorous benchmark that tasks agents with addressing genuine GitHub issues sourced from open-source projects such as Django and scikit-learn.

Devin achieved an end-to-end resolution rate of 13.86 per cent, significantly surpassing the previous state-of-the-art performance of 1.96 per ent. 

Notably, even when provided with the precise files to modify, prior leading models could only address 4.80 per cent of the issues.

Who is Cognition?

Based out of New York and the San Francisco Bay Area, Cognition aims to build “AI teammates” with far superior capabilities. The company has raised $21M in a Series A round led by Founders Fund. 

The AI company also has backing from industry leaders, including Patrick and John Collison, Elad Gil, Sarah Guo, Chris Re, Eric Glyman, Karim Atiyeh, Erik Bernhardsson, Tony Xu, Fred Ehrsam, and others.