Based on some recent studies, AI chatbots have displayed unbelievable potential by setting up and running an imaginary software company – ChatDev. The chatbots have developed software from scratch within seven minutes while keeping costs under a dollar. This extraordinary achievement has been made possible by powerful AI technology such as OpenAI‘s ChatGPT. This experiment now opens new doors within software development. In this article, we delve into the findings of this study. We also explore how AI chatbots achieved this feat and its implications for the tech industry.

The Experiment: Developing ChatDev an imaginary software development company

Researchers from Brown University and other Chinese universities tested if AI chatbots, specifically the version of ChatGPT’s 3.5 model, could complete a software development process without any human intervention. Researchers set up ChatDev – an imaginary software development company inspired by the well-known waterfall software development model. The company was then broken down into four key stages – designing, coding, testing, and documenting. From there, researchers allocated AI bots definite roles by providing each bot crucial details that defined selected roles and tasks, termination criteria, communication procedures, and constraints.

Each bot was assigned to its respective stage once the AI bots were allocated their roles. The CEO and CTO of ChatDev, for example, worked in the designing stage, and the programmer and art designer worked in the coding stage.

From CEO to CTO- Allocating Roles at Each Stage

During each stage,  minimal human input was given to complete the explicit parts of the software development procedure. From determining the programming language to identifying the bugs in the code — till the software was complete. The experiment was run across various software scenarios. It was used a series of analyses to check the time ChatDev took to complete each procedure and the cost associated with each task.

Researchers, for example, tasked ChatDev to design a basic Gomoku game, a strategy board game also known as “Five in a Row.”

At the designing stage, the CEO asked the CTO to “suggest a tangible programming language according to a new user’s demand. “The CTO’s answer was Python. The CEO then said, “Great!” and explained that the software language’s readability and simplicity make it a popular choice for most developers.” ChatDev then went to the coding stage, with the CTO asking the programmer to write a file.

After allocating ChatDev 70 tasks, the AI-powered company was able to finish the entire software development cycle within seven minutes at a cost of less than a dollar. The time also includes detecting and troubleshooting the possible vulnerabilities through its memory as well as self-reflection abilities. The study states close to 86.66% of the software systems did execute flawlessly.

The researchers state that “our experimental results display the cost-effectiveness and efficiency of the automated software development process driven by ChatDev. “

Efficiency and Cost-Effectiveness of ChatDev

The experiment involves multiple software scenarios and analyses to measure the time and cost of development. ChatDev has completed the whole software development process within seven minutes at an amazing cost of less than one dollar on average. Thanks to the bots’ “memory” and “self-reflection” capabilities, this remarkable feat was achieved while identifying and resolving potential vulnerabilities. The study reported an impressive 86.66% accuracy rate, with most generated software systems executing flawlessly. This efficiency and cost-effectiveness demonstrated by ChatDev have vast implications for the software development industry.

Implications and Real-World Applications

This study points to the extraordinary potential of generative AI technologies, such as ChatGPT, in executing explicit job functions. Since its introduction, various industries have adopted ChatGPT which enhances productivity and saves time.

Even coders have adopted generative AI tools. For example, Daniel Dippold, a Berlin coder, used ChatGPT to develop a program to help him locate an apartment. Even tech giants like Amazon leverage ChatGPT for software development, displaying its wide range of applications.

Challenges and Future Projections

Though this study is irrefutably groundbreaking, it is essential to acknowledge that the technology has its own limitations. Researchers have detected biases and errors in the language models that could possibly affect software creation. However, researchers are hopeful about the influence of these findings on junior engineers and programmers within the real world. It is a well-accepted fact that such challenges will find a resolution as AI evolves.


AI chatbots running a software company successfully and creating software in a matter of minutes is a revolution. It ushers in a new era in software development. The cost-effectiveness and efficiency of this process are redesigning the tech industry. It is an exciting time for engineers, coders, and AI enthusiasts. While there are multiple challenges that we need to overcome, the capability for AI-driven development is boundless. The technology promised to revolutionize the way we develop software in the fu

Leave a Reply