虎嗅

Burned 200 million tokens; a complete novice used AI to develop a system for monitoring investment opportunities.

原文:烧了 2 个亿token,纯小白用AI搓出一套坑位监测系统

Summary of the Key Points

The core of this sharing is as follows: The author's company faced a serious issue due to a shortage of toilet stalls, which led to lengthy queues. He used an AI code tool called Codex to quickly develop a monitoring system that solved this problem. At the same time, he shared the various challenges he encountered during the development process, particularly those associated with "vibe coding" (writing code without a clear plan or understanding of the underlying principles). Essentially, this story illustrates how technology can be used to address everyday frustrations and highlights the common difficulties non-professionals face in embedded development.

1. The Company's Toilet Stall Crisis: How Painful Is This for Employees?

A lack of toilet stalls is a common issue for many employees, but the situation in the author's company seemed particularly extreme. During peak hours, people would crowd at the toilet doors, with some getting anxious and others occupying stalls for extended periods while using their phones. There was even a chaotic competition where the first person to arrive got to use the stall. This not only wasted time but also affected work morale, considering the physiological need to use the restroom. This seemingly trivial issue of uneven resource distribution directly impacted the employees' daily comfort, which motivated the author to take action.

2. Codex: How Did the AI Tool Help Him Quickly Build the Monitoring System?

Codex is an AI code assistant developed by OpenAI that can automatically generate code based on user instructions. The author didn't have to write the entire code from scratch; he simply needed to describe his requirement in plain language. Codex then provided the necessary code framework. For example, it helped the author write the code to read the status of the toilet doors and display the availability of each stall on a screen. This significantly reduced the development time and effort, allowing the author to turn his idea into a functional system even without being a professional embedded engineer.

3. The Challenges of Embedded Development: Why Is "Vibe Coding" So Troublesome?

Embedded development involves writing software for hardware devices such as sensors, displays, and microcontrollers. The author referred to the process as "vibe coding," meaning writing code without a systematic approach. Since he had no formal training in embedded development, he had to rely on trial and error to overcome various issues:

  • Hardware Connection Problems: Incorrect connections between sensors and microcontrollers could cause the system to fail, making it difficult to identify the error.
  • Debugging Nightmare: The software worked fine on the computer, but failed on the hardware, resulting in unexpected behavior or inconsistent sensor data.
  • Compatibility Issues: Different brands of sensors and displays required frequent code adjustments to ensure proper functionality.

These challenges were particularly frustrating, highlighting the complexity of embedded development, which requires a deep understanding of both software and hardware.

4. The Effect of the System: No More Queues for Using the Toilet?

The system works by installing door magnets at each stall to detect if the door is closed. The sensors send this information to a microcontroller, which then updates the display in a public area of the company. Employees can check the display before using the restroom to see which stalls are available, eliminating the need for waiting in line. As a result, queue times were reduced, saving time and reducing conflicts over stall usage.

5. The Implication for Everyone: Ordinary People Can Use Technology to Solve Everyday Problems

The most interesting aspect of this story is that technology is no longer limited to professionals. In the past, embedded development required extensive knowledge of hardware and complex coding. However, with tools like Codex, anyone with an idea can use AI to create practical solutions. For instance, you could use similar techniques to create smart home devices or pet feeders. This story shows that even small, everyday problems can be improved with technology, and the value of such solutions is not limited to large-scale projects.

In summary, this sharing not only addresses the practical challenges faced by employees but also showcases the power of technology to simplify everyday life. It demonstrates how AI tools can make embedded development more accessible to non-professionals, highlighting the potential of technology to enhance our lives.