Advertisement

Updated March 17th, 2024 at 14:06 IST

Will Cognition Labs’ Devin AI make software developers’ jobs redundant?

Devin is capable of solving complex software development problems without human intervention and developers from around the world are discussing its potential.

Reported by: Anirudh Trivedi
Scott Wu, Co-Founder and CEO, Cognition Labs
Scott Wu, Co-Founder and CEO, Cognition Labs | Image:Republic Business
Advertisement

Future of software development:  Amid fanfare, the launch of OpenAI’s ChatGPT 3 in 2022, overwhelmed the tech industry and caused a deep curiosity about the road ahead. 

The potential threats posed by the new AI model replacing a huge chunk of the workforce were imminent. Over a year later, ChatGPT has not only become an effective assistant to working professionals but has also generated a new segment of jobs that are driven by AI models. 

Advertisement

The most recent advent in AI is Cognition Lab’s Devin AI, a new AI software developer model that claims to be significantly better than any other similar AI models that we currently have. 

As per Cognition Labs, Devin is capable of solving complex software development problems end-to-end without human intervention and only gets better by learning from their previous experiences. 

Advertisement

Devin’s astuteness and levels of efficiency have become a cause of concern for software developers globally. Devs from around the world are discussing the potential of Devin and whether it could pave the way for a mass software developer job threat in the industry. What are the areas where Devin AI can make devs worried about? Republic Business takes a deep dive. 

What is Devin AI? 

Devin is an innovative AI software developer created by US-based startup Cognition. Unlike other AI code generators, Devin can not only code but also autonomously handle complex engineering jobs with proficiency. 

The key yardstick to judge Devin’s performance can be gauged by the fact that this AI marvel scored 13.86 per cent accuracy on the SWE-Bench benchmark, which tests AI models on various software engineering tasks. This means that out of every 100 issues or problems encountered, Devin successfully resolved nearly 14 of them without any human assistance.

Advertisement

Operating within a secure computing environment similar to human developers, Devin integrates with tools like a code editor and browser. It learns from past experiences, recalls relevant context, and corrects errors independently, which improves with time. 

When comparing it with the previous best AI software developer Claude 2 which achieved only 1.96 per cent accuracy without any help and 4.80 per cent accuracy with some assistance from humans, Devin seems to be a major leap that can ‘actually’ materialise its effect in the industry. 

Advertisement

What makes Devin AI different from ChatGPT? 

Devin AI and ChatGPT serve distinct purposes and are designed for different tasks. Devin specialises in development and is designed to handle complex coding tasks. On the other hand, OpenAI’s ChatGPT is a conversational AI model designed for generating human-like text based on the input it receives and is not specifically trained for software development tasks like Devin. 

Advertisement

Devin is trained by a team of coding geniuses including Scott Wu, his elder brother Neal Wu, and other industry top brains who have previously worked for Google DeepMind, Cursor, and ScaleAI. With tons of relevant programming data and under the watch of ‘legendary grandmasters’ of coding (Scott Wu’s Codeforces rank) Devin is extensively trained to debug code and build applications.

ChatGPT caters to a different purpose of generating natural language responses, engaging in conversations, providing information, and performing various language-based tasks.

Advertisement

Should Devs be worried about Devin?

While Devin's performance on the SWE-Bench benchmark and Cognition Labs’ videos may look like nothing but a threat to developers working in the industry, the actual impact of this emerging intelligence can be very different from what we can imagine. 

Advertisement

Firstly, the US Bureau of Labor Statistics projects a strong growth in software developer employment which means the demand for human developers is only going to grow as AI technology gets more complex and its scale nears its potential. Secondly, Devin's capabilities are more directed towards non-technical founders, democratising software engineering for those who do not understand it. 

The majority of senior developers working in the industry do not code most of the time they are working. With time, software development is becoming more focused towards deobfuscate complex requirements and understanding the problems and finding their effective solutions. 

Advertisement

Even if Devin becomes accessible to all developers around the world, it will only empower them to cut down on the coding work and focus more on problem-solving and architectural design. Devin can become more of a tool that complements the skills of the developers rather than a threat to their profession.

Advertisement

Published March 16th, 2024 at 14:16 IST

Your Voice. Now Direct.

Send us your views, we’ll publish them. This section is moderated.

Advertisement
Advertisement

Trending Quicks

Advertisement
Advertisement
Advertisement
Whatsapp logo