In a recent YouTube video, Santiago, a software engineer and course creator with many years of experience, takes a closer look at Devin, an AI system that has been causing quite a stir in the tech community. In this article, I will make a review of his video, as his opinion and experience is perfect to serve as valuable source.

Created by Cognition Labs, Devin is being heralded as the first AI software engineer, with the potential to replace human developers. The video, which you can watch here, offers a captivating insight into the current state of AI in software development and its implications for the future of the industry.

Initial skepticism

Known for his expertise and contributions to the field, Santiago initially approached Devin with skepticism. However, he was granted unrestricted access to the tool to assess its performance on five diverse software projects, ranging from a digit classification app using machine learning to a backend API and even a simple game. The objective was to evaluate Devin’s ability to handle various tasks and determine whether it could genuinely replace human developers.

Devin’s successes

Throughout the video, Santiago guides viewers through each project, highlighting Devin’s achievements and shortcomings. Remarkably, Devin demonstrated the ability to autonomously build functional software for most of the projects with minimal human intervention. The AI system only required occasional guidance when faced with obstacles and needed API keys to be provided. In one particularly impressive instance, Devin independently identified the need to upgrade the code in an older project from TensorFlow 1.0 to 2.0, showcasing its capability to adapt to evolving technologies.

Limitations and challenges

Despite its successes, Santiago also sheds light on several significant limitations in Devin’s current form. The AI-generated code often contained bloated and unused functions, and in some cases, core functionality was compromised when Devin attempted to make changes. The system could be slow, grapple with seemingly minor UI adjustments, and occasionally pursue unproductive paths. In the final project, which involved creating a command-line app using OpenAI, Devin struggled to deliver a fully functional solution.

The potential and future of AI in software development

Notwithstanding these limitations, Santiago emphasizes the immense potential of AI systems like Devin. While it may not be ready to completely replace human developers, its ability to work on multiple projects concurrently could drastically increase the amount of software being developed. As the technology continues to advance, AI-assisted coding could revolutionize the industry.

Santiago predicts that AI will drive a surge in demand for software by enabling the creation of more sophisticated and efficient automation. However, this shift will also raise the bar for human developers. As routine tasks become increasingly automated, developers will need to focus on refining their advanced engineering, architecture, and problem-solving skills to remain competitive in the job market.


Devin, although still in its early stages, represents a significant milestone in the realm of AI-augmented software development. As these technologies continue to progress, it will be essential for developers to stay informed about the latest advancements and continuously adapt their skill sets.

Santiago’s video serves as an insightful exploration of the future of coding and the role AI will play in shaping the industry. It encourages developers to embrace continuous learning and to view AI as a powerful tool rather than a threat. While Devin may not be ready to take over just yet, it offers an exciting glimpse into the possibilities that lie ahead.

To gain a more comprehensive understanding of Devin and Santiago’s firsthand experience testing its capabilities, be sure to watch the full video. It’s a must-watch for anyone curious about the cutting edge of AI in software development and its potential impact on the field.

Categorized in:

MLOps, Programming,

Last Update: 03/05/2024