In the age of AI, will programmers be unemployed? Do they still need to learn to code?

Some time ago, Cognition released a product for Devin AI, which has been called the "first AI programmer" and has caused a lot of discussion. Some people even pessimistically think that the position of programmer is about to be replaced, so, is this really the case?

Recently, more and more people are asking similar questions, AI is so powerful, do you still need to learn programming? Is computer science still popular? Especially when many people saw Devin AI yesterday, they even had a kind of question, "Are programmers going to lose their jobs?".

1. What can an AI programmer do?

Just a few days ago, Cognition, a two-month-old company, released a product for Devin AI. It was advertised as the first-ever AI programmer. In the demo, Devin AI has its own tools such as command line, code editor, and browser to make plans, perform tasks, and solve problems on its own, and it can independently complete the work of developing and building the entire software.

In addition, according to official propaganda, Devin even has the ability to "grow", he can read articles, learn techniques that he didn't know before, and find errors in the program on his own and correct them.

Judging from the hype, Devin is a step further than the previous AI programming assistant similar to Copilot, and is more like a programmer who can complete development tasks independently. This not only represents the feasibility of AI to complete the development work independently, but also stimulates the public debate about whether AI can replace programmers.

Some netizens also analyzed, in fact, Devin AI is not as powerful as imagined: first, Devin's underlying technology is based on GPT4, and its use cost is higher than that of ordinary programmers; secondly, the programmer's interview questions are not difficult, and ChatGPT can also do it; and the task performed is too simple, and there is still a big gap compared with humans. Therefore, human programmers are relatively safe at the moment.

That said, as the performance of large language models gets better and better, AI programmers will inevitably play a key role in the software development process for the foreseeable future.

2. Do I still need to learn programming with AI?

Could it be that if you learn programming now, you won't be useful in the future?

I think this question can be likened to "Do I need to learn English when I have translation software?" and I believe many people will answer that learning English is still important.

It is true that English has become a part of the overall quality of many people. Translation software doesn't solve our needs 100%. For example, to find the most up-to-date papers and materials, knowing English allows us to obtain the original information without loss, and translation software is just a tool that allows us to quickly browse and filter the information. In addition, learning English is not only about learning the language, but also about learning a culture and way of thinking. In this way, we can better connect with the world.

The same logic applies to AI and programming.

First of all, programming is only one part of software development, programming is a key to the computer world, and the complexity of computer science goes far beyond programming itself. It includes many aspects such as system architecture design, network security, requirements analysis, user experience, and project management. These comprehensive knowledge and skills are the foundation of software innovation. Even if AI can automate programming tasks, it will require humans to solve more complex problems, humans to maintain a keen insight into industry trends, and they need to understand business needs, communicate effectively with team members from non-technical backgrounds, and ensure that technology solutions meet business goals.

In addition, programming is not just about writing code, it is a way of solving problems and a kind of thinking training. Learning to code can help people develop logical thinking, systems thinking, innovative thinking, and hands-on problem-solving skills. These capabilities are especially important in the age of AI, not only in the field of technology, but also in all aspects of life. As technology continues to advance and new tools and platforms emerge, individuals need to be able to adapt to these changes and find opportunities to innovate in them. This resilience and innovation are valuable assets for personal development, career and even corporate competitiveness.

In addition to this, programming is not a patent for computer science, it is also an interdisciplinary learning tool. When trying to solve complex problems from different fields, programming can help integrate and apply multidisciplinary knowledge through technologies such as data analysis, visualization, machine vision, and simulation, thereby facilitating the generation of innovative solutions. This interdisciplinary perspective not only broadens our thinking, but also provides clearer guidance for AI programmers in assigning tasks. This allows us to understand and apply technology from a more integrated perspective, which in turn enables us to move projects and research forward more effectively.

In fact, there is still a lot to iterate on AI itself. For example, humans still play an irreplaceable role in AI research papers, designing AI algorithms, improving AI performance, customizing AI models, and AI ethics and security. In other words, it is precisely because of the development of AI that more challenges and opportunities have been created for computer professionals. For people who are ready to realize their ideas, now is the best time to learn programming, learning programming can better apply the capabilities of AI and help us realize a lot of ideas and ideas.

So, when faced with the question, "Do you still need to learn programming with AI?", my answer is: absolutely.

3. Tell me about my experience

I studied industrial design in college, majoring in liberal arts in science and engineering. Originally, programming was a very daunting thing for me, and I almost failed the C language in college. But since I have always been interested in the Internet and have the urge to make my own products, I went to study programming and software engineering when I was about to graduate. After that, I became a programmer, a product manager, and an entrepreneur, and along the way, there were setbacks and gains.

In the past, I often deliberately concealed my development experience for fear that others would think that my career positioning was not focused enough. But after many years, I realized that it was interdisciplinary competence that helped me.

If you are a product manager who wants to start your own business today, what is the difference between people who have learned to code and those who don't know how to make products?

First of all, product managers who have studied technology have a clear advantage in understanding the details of the technical level. Ability to communicate better with the R&D team, communicate requirements more precisely, better understand the challenges the team faces, and make more rational decisions on technical feasibility and resource allocation. This in-depth understanding helps build trust and respect among team members, which in turn facilitates teamwork and projects. Over the years, I've maintained a good relationship with most of my development colleagues, because teamwork is nothing more important than understanding.

Second, product managers with programming backgrounds are better able to consider the complexity and cost of implementation when designing products, pay more attention to the value of requirements, and can better develop MVP (Minimum Viable Product) strategies. Especially in the early planning stage of the product, you can anticipate the technical obstacles that may be encountered in the implementation of certain features, and do not design the product features to be overly complex or idealistic, so as to make more realistic and economical choices in design. This allows the product to win the time to quickly validate the market at a low cost. This forward-looking approach not only saves development time and costs, but also avoids major modifications later in the project, improving the efficiency and success rate of product development.

In addition, understanding the technical principles also allows us to better control the product experience. They will take the initiative to think about the technical indicators that affect the product experience, and will not take it for granted to do some idealized functions that are divorced from the actual scene. By being able to better assess the difficulty of technical implementation of different design options, the optimal balance between design and functionality can be found. Such product managers are able to drive design and technical teams to work more closely together to create products that are both beautiful and efficient.

In addition, programming experience gives product managers the sensitivity and adaptability to emerging technologies. In the era of AI, new technologies are emerging one after another, and product managers with technical backgrounds have strong technical acumen and can more quickly understand how these new technologies are applied to products and the changes they may bring to the market and user experience. For example, I often go to Github to check out interesting open source projects, learn new technical courses, and think about how to apply them to my own projects and work. This ability enables product managers to lead teams at the forefront of technological developments to create innovative and competitive products.

Finally, by learning to code, I also gained a series of powerful thinking tools, including object-oriented programming ideas, design patterns, and the Unified Modeling Language (UML). These tools have improved my logical and systems thinking skills, taught me how to break down complex problems into manageable small tasks, helped us abstract and model problems, and helped me better understand how things work so that I can find opportunities. This has helped me immensely in product planning, market analysis, and project management.

In the era of generative AI, large language models have become my indispensable assistant. In the past, there were many brilliant ideas that could not be realized because of the limitations of my technical vision, and they ended up sleeping in my mind and gradually being forgotten. But now, with GPT for some of my whimsical ideas, it can generate code very quickly. Although the code may be full of bugs, GPT has helped me quickly expand my technical horizons, guided me to dig deeper, and allowed me to fine-tune the code. For example, during the development of the open source project of ComfyUI's Mixlab-Node, many ideas were implemented through collaboration with GPT.

Therefore, as an independent developer or product manager, we should think of the AI programmer as a collaborative partner. And it is because of its help that we can implement more interesting features while saving more time that we can use to think about things that make money.

Fourth, summary

When we are dealing with an AI programmer like Devin, we don't have to worry too much about the programmer position. Programming will become a universal ability like a foreign language, and by learning to program, we can better utilize the capabilities of machines.

With the help of AI, more and more ordinary people can realize their own ideas and ideas and become the helmsman of their own business. In the era of AI, it is bound to be the era of the rise of super individuals.

In traditional enterprise management, employees play the role of parts on the assembly line. However, in this new era of AI transformation, we should not limit our role to a certain component, and it is more important to have flexible thinking, broad vision and keen insight.

As long as we grasp the trend of the times, we will not be pressured by the technology. Let's do it!

Columnist

Uncle PM, WeChat public account: PM Uncle Xiong, everyone is a product manager columnist. Born as a product manager in education, he has studied design, developed and operated as a product manager.

This article was originally published on Everyone is a Product Manager. Reproduction without permission is prohibited

Image from Unsplash, based on the CC0 license

The views in this article only represent the author's own, everyone is a product manager, and the platform only provides information storage space services.