2 min read

The Coming Wave: Mustafa Suleyman’s Vision for Autonomous AI in Action with ChatDev

Today, when you think about Artificial Intelligence, it's easy to get wrapped up in the flashy demonstrations of language generation or self-learning capabilities. However, Mustafa Suleyman, in his book The Coming Wave introduces a more grounded perspective. Rather than getting enamored with what an AI can say, he's interested in what an AI can do in the real world.

Understanding the Modern Turing Test

Suleyman's Modern Turing Test is a thought experiment that shifts the lens from abstract intelligence to concrete action. For an AI to pass, it would not only have to devise a strategy but act on an instruction: “Go make $1 million on a retail web platform in a few months with just a $100,000 investment.” This would mean navigating a complex maze of tasks – from product research and design to contract negotiations and marketing. While a human touch would still be essential in executing certain aspects, the heavy lifting would be done by the AI.

Hello, ChatDev!

Recent findings have made it evident that we're not far from Suleyman's envisioned future. A collaborative study between Brown University and several Chinese universities revealed that AI chatbots, particularly those driven by OpenAI's ChatGPT 3.5 model, can manage a software company's entire operation cycle – design, code, test, and document.

Under a hypothetical software development company, ChatDev, these chatbots were designated specific roles and communicated to achieve predefined objectives. From CEO to programmer, the bots worked harmoniously, deciding on programming languages, identifying bugs, and ensuring the software's completion.

In a striking demonstration, ChatDev was assigned 70 tasks, managing to complete an entire software development cycle in under seven minutes, costing less than a dollar per task. Moreover, 86.66% of these generated software systems were executed flawlessly.


These results are more than just academic. They underscore the expansive potential of Generative AI technologies, such as ChatGPT, in reshaping industry dynamics. The implications are vast: from boosting the efficiency of software development to serving as tools for coders, to simplify their tasks.

Yet, as with all innovations, there are wrinkles to iron out. The research recognized certain limitations in the AI models – errors and biases that could skew software creation. These findings act as a precursor, suggesting that while the Modern Turing Test's bar might seem high, the future of AI in industry is inching ever closer to that standard.

For those interested in diving deeper into the future of technology and its implications, my WTF Journal serves as a repository of thoughts and questions that can guide your exploration. After all, the future is not something to predict; it's something to be understood.