In recent years, even in the field of system development and app development, opportunities to utilize generative AI have increased.
Whether it's building a code base, investigating the cause of an error, organizing a design proposal, or putting together a document, AI is here to help engineers.
On the other hand, some people are worried that relying too much on AI may reduce the engineers' own technical skills.
This is a very natural question. The more you become accustomed to useful tools, the weaker your judgment and understanding may be. Especially in programming, if you use the code released by AI as it is, there is a danger that development will proceed without understanding why it works and where the problem is.
However, using AI does not in itself lead to a decline in technology.
The important thing is to use AI not as a substitute but as a tool to increase the speed of thinking.
AI is useful, but the final decision must be made by an engineer
AI can produce code, text, and design proposals in a short period of time.
However, what the AI puts out is not always correct. Even if the code seems to work, there may be a security issue. Even if it looks like there is no problem now, it may have a structure that is troublesome for future maintenance.
For example, system development requires the following decisions:
- What kind of design is easy to maintain?
- Is the database structure reasonable?
- Is there a security issue?
- Can you withstand future feature additions?
- Is it difficult to cause trouble during operation?
- Is there a problem with integrating with the existing system?
- Is it easy to understand for users of smartphone apps and web systems?
You can't just decide whether or not you can write code. You need to think about the actual operation, operation, user movement, and even the work of the administrator.
AI is very good as an auxiliary, but it is the engineers who are ultimately responsible for judging.
We do not develop round throws to AI
At Makoto Tejima, AI may be used in development.
However, that doesn't mean leaving the technology to AI.
Engineers design, implement, review, modify, and validate, even with AI. Instead of adopting the AI-issued code as it is, we will review the content, modify it as necessary, and determine whether it fits the overall structure of the system.
Using AI will increase the speed of work, but it will not leave it to AI until quality control.
Rather, it is because of the use of AI that engineers are required to have a higher degree of confirmation.
Basic technical skills are required to see what the AI has come up with and determine whether it is correct. Implementation experience is required to spot mistakes. Assuming post-operational issues requires on-site experience.
In other words, engineers in the AI age need more than just the ability to write code. It is the ability to identify the output of AI, use it appropriately, and finish the necessary parts responsibly by humans.
There are also occasions where AI can be used to improve the quality of development.
AI can also help improve the quality of development when used correctly.
For example, if you compare multiple design proposals before implementation, investigate the cause of the error, or identify the test point of view, AI can be used to broaden the scope of consideration.
AI is also useful for creating documents and organizing specifications.
In development, not only does it take time to write code, but it also takes time to review specifications, research, and organize materials. Streamlining these tasks will allow engineers to spend time on more important designs and quality checks.
The important thing is not to omit the task of thinking by AI.
Reduce the waste of work with AI and focus on what humans need to think about.
How to use and utilize technical skills
Whether or not using AI will reduce your technical skills depends on how you use it.
An easy way to use technology is to use it without understanding the output of AI. When an error occurs, ask the AI, paste the proposed correction as it came out, and when another error occurs, ask the AI again. Repeating this will obscure your overall understanding of the system.
On the other hand, there are ways to use technology.
Let the AI come up with a proposal and see why it is being implemented. Compare with another implementation method. Review in terms of security and maintainability. Align with existing design policies.
In this way, AI is not a deprivation of technical capabilities, but a tool to support development.
It's not about whether or not you use AI that matters to your customers.
When it comes to system development and app development, it's not just whether you're using AI that really matters.
What matters is whether the completed system works safely, is easy to use, is easy to maintain, and is a good fit for the business.
Even if you are using AI, if the design is messy, it will not be a good system. Conversely, even if you are using AI, if the engineers have thoroughly checked, designed, and verified, you can develop efficiently while maintaining quality.
At Makoto Tejima, even when using AI, we value the design power, implementation power, confirmation power, and maintainability that are the essence of development.
AI can speed things up, but it doesn't slow you down.
Rather, it improves the overall quality of development by using AI appropriately and looking more carefully at what humans are responsible for.
Summary
It is believed that AI will become one of the indispensable tools in the future development site.
However, AI-enabled development is quite different from AI-enabled development.
If you leave it to the AI, your technical understanding may be shallow and your ability to respond in case of trouble may be reduced. On the other hand, if engineers understand, review, and make the necessary decisions, AI can help increase development efficiency and quality.
At Makoto Tejima, we use AI as a useful tool without downplaying technical skills or quality control.
If you are considering web system development, smartphone application development, business improvement using AI, etc., we will propose it according to the purpose and operation.
