Cognition Labs Unveils “Devin”, An AI That Could Potentially Give Software Engineers A Run For Their Money

Software engineers should probably be "cautious" as an AI startup has created an autonomous agent known as Devin that can replace conventional programmers.

When we talk about AI, we "literally" mean a technology that can co-exist with humans to aid them in specific tasks in unimaginable ways. With the influx of generative AI into mainstream media, followed by several similar developments, we are now at a point where AI has reached a state of maturity, and this is indeed evident in a new development by the startup Cognition Labs, which has created an "AI chatbot," the Devin, explicitly designed to solve complex coding problems, with the likes of clearing code marathons and successfully passing through a "sophisticated" coding interviews as well.

Today we're excited to introduce Devin, the first AI software engineer.

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

— Cognition (@cognition_labs) March 12, 2024

For a bit of background context on Cognition Labs, the firm has recently received massive attention from investors, with the likes of Founders Fund and several reputable investors pouring decent sums of money into the startup's venture. Now, the so-called AI "software engineer" Devin is indeed a marvel by the firm, as instead of focusing on auto-completing pieces of codes, Devin is designed with a "dynamic" way of interpreting data, and the firm claims that their model has seen a huge breakthrough, which involves the use of reasoning rather than relying on just conventional data sets.

4/4 We even tried giving Devin real jobs on Upwork and it could do those too! pic.twitter.com/7XUWQLeZm1

— Cognition (@cognition_labs) March 12, 2024

Now, the exciting part comes in Devin's ability to code, and to cut it short, the technology does perform some wonders. Based on Cognition Labs' tests, Devin managed to generate API keys through a self-written code and fine-tuning different LLM models, which is surprising because AI hasn't reached such a "coding" level until now. At SWE-Bench benchmark tests, the AI resolved 13.86% of the issues unassisted, which is almost 10 times higher than what OpenAI's ChatGPT managed to offer. What we see here is something "divine" in the era of the tech industry, and this is just the beginning.

With such AI coders in the market, many people raise the question of whether AI would replace human coders, and this has been a massive topic of debate. However, for me personally, I believe tools like Devin could act as a huge aid for developers since it eradicates the need to be concerned about the more "boring" elements of coding allowing humans to expand creativity towards their work. However, AI could be a reason behind potential layoffs as we move into an era where technology becomes and thinks more "human-like."