Summary of Key Points
AI is revolutionizing the production and development (P&D) landscape of the software industry: The traditional division of labor, which relied on a sequence of "product manager → UI designer → developer → tester," has been disrupted by AI, significantly reducing the cost of various tasks such as prototyping, coding, and testing. The boundaries between job roles are becoming blurred or even disappearing, leading to a decrease in demand for low-skilled positions (such as simple UI design, testing, and entry-level development). In the future, P&D teams will consist of humans working alongside AI, with a single individual capable of handling entire modules. However, this shift also brings the risk of an influx of low-quality software, shifting the industry's value from technical skills to the ability to make informed decisions about what not to develop and the power to help users filter through the plethora of options.
1. AI Makes Cross-Functional Roles the Norm, and Low-Skilled Positions Are at Risk
In the past, software companies had a detailed division of labor because each step was time-consuming and costly: product managers needed to clearly define requirements, UI designers had to create attractive user interfaces, developers had to write code, and testers had to ensure there were no errors before release. With AI, these tasks can now be automated—product managers can use AI to generate high-fidelity prototypes without tedious styling adjustments, programmers can quickly write code (using millions of tokens in a day), and testing cases can be pre-processed by AI.
As a result, the barriers between job roles are collapsing. There will be fewer UI designers, as product managers can create preliminary versions using AI; there will be fewer testers, as developers can test their own code with AI assistance; and those who only know how to write simple functions or create basic prototypes will face reduced demand, as these skills are no longer as valuable due to their widespread availability.
2. Future P&D Teams: One Person + AI Handles an Entire Module, Reducing Communication Costs Dramatically
Previously, developing a single module required multiple steps involving product managers, UI designers, developers, and testers, with frequent handovers, explanations, and rework. In the AI era, one person can take on all these responsibilities—from requirement analysis to prototype design, code implementation, testing, and even post-release data analysis. The key is to manage AI effectively—informing it about the business context, reviewing its outputs, and determining whether they address user needs.
This approach eliminates much of the time spent on coordination, greatly improving efficiency. For example, what used to take a week for a small feature can now be completed in a day, without waiting for others to complete their tasks.
3. Software Is Now Easier to Create, but Could Lead to More Low-Quality Products
While the cost of developing software has decreased significantly, AI has also lowered the barrier to entry, allowing anyone to create a working product in just a few days. This potentially leads to an overabundance of low-quality software. The reason is that lower execution costs mean people may overestimate their abilities; what used to be rejected as poor ideas due to time-consuming development processes can now be quickly implemented.
4. In the AI Era, Skills Are Less Valuable than "Judgment" and "Trust"
AI can automate repetitive tasks, but it cannot replace critical decision-making skills and the trust that users place in products. The most valuable talents in the future will be those who:
1. Can Make Informed Decisions: Understand whether a feature is useful, whether a requirement is genuine, and whether users are likely to buy it. This requires industry knowledge, customer understanding, and an understanding of human behavior (e.g., realizing that users often request faster loading speeds out of boredom).
2. Have Influence: Help users filter through the sea of software by providing trusted recommendations. Users will trust tools recommended by industry experts or brands they know to be reliable.
Conclusion
AI disrupts the execution aspect of software development but does not replace the need for strategic judgment and trust. The true value in the future lies in those who can make informed decisions and build trust with users. This means a shift from focusing on replicable skills (such as prototyping and coding) to developing unique, non-replicable abilities (decision-making and influence). The changing nature of software P&D reflects a redistribution of industry value. Individuals should focus on developing these skills, while companies should avoid wasting resources on ineffective products and prioritize delivering meaningful solutions.