<div id="readability-page-1" class="page"><div><div><a rel="noopener follow" href="/?source=post_page-----4071ac6e4245--------------------------------"><div aria-hidden="false"><p><img alt="Beejei" src="https://miro.medium.com/v2/resize:fill:88:88/0*uqVHtuzJjYbj_uOG.jpeg" width="44" height="44" loading="lazy" data-testid="authorPhoto"></p></div></a></div><figure><figcaption>Seamless Collaboration — Dall-E</figcaption></figure><h2 id="abfe">Eliminating the Understanding Gap Between Development Teams</h2><p id="231b">When developing a service targeted at general users, it’s common knowledge that there are typically Frontend and Backend developers involved. While there are many other categories, focusing on these two can illustrate the point. These roles are often at odds due to their separation of concerns. Even when addressing the same value, such as “efficiency,” their strategies can vastly differ. For a Frontend developer, efficiency means smooth loading and high responsiveness in a web browser. For a Backend developer, efficiency pertains to the server resource usage and cost-effectiveness.</p><p id="4da9">These technology-driven areas can be approached in numerous ways, depending on the perspective. However, from a business standpoint, trying to maximize usability and profitability within a reasonable cost often results in vague directives like “cost-effective efficiency.” This makes it challenging to pinpoint specific methodologies, often leading developers to default to familiar technology stacks.</p><p id="a95e">Using LLMs can simplify these dilemmas. By providing appropriate conditions and requesting code, the output can serve as a guideline. The key is to create prompts that require minimal technical detail while yielding satisfactory results. This approach ensures that any developer can easily align with the output, fostering better collaboration.</p><h2 id="167f">Eliminating the Gap in Code Reviews</h2><p id="7bfe">The second point is about eliminating the gap in traditional code reviews. While code reviews help developers produce higher quality code over time, they can be a source of significant stress. The original author often prefers to stick to their approach, even if there are minor issues. Simple, agreed-upon changes can be made easily, but complex situations make code reviews difficult and time-consuming. This leads to high costs and delays in project completion. Culturally, code reviews can be an opportunity to build camaraderie, but they are still challenging and heavily reliant on individuals’ soft skills.</p><p id="f9ac">By using LLMs to generate code, everyone can take on the role of a code reviewer. Instead of traditional code review processes, developers can focus on refining the prompts and checking the output. This iterative process allows the team to align more closely with each other’s expectations. As the team reviews and adjusts the prompts together, they maintain a similar vision for the final product. Over time, this repeated collaboration fosters stronger bonds and more effective teamwork.</p><h2 id="99c0">Bridging the Gap Between Planning and Execution</h2><p id="b791">The third point is about closing the gap between planning and execution. Planning is often established by upper management based on the company’s grand vision. However, these plans frequently overlook the day-to-day tasks that need to be done. As plans trickle down to individual teams, they have to contend with ongoing tasks, backlogs, communication with other teams, and internal coordination. This results in increasingly rough plans with declining achievement rates.</p><p id="d3da">Using LLMs can help reduce the time needed to reach planned goals. This allows for more detailed planning that aligns with the company’s aspirations. Employees can more easily align with the company’s direction, resulting in a more dynamic organization. With LLMs streamlining the execution of tasks, plans can be more specific and ambitious, fostering a more cohesive and agile workforce.</p></div></div>

Stories by Beejei Kim on Medium

Programming in the LLM Era #2 — How
Eliminating the Understanding Gap Between Development Teams
When developing a service targeted at general users, it’s common knowledge that there are typically Frontend and Backend developers involved. While there are many other categories, focusing on these two can illustrate the point. These roles are often at odds due to their separation of concerns. Even when addressing the same value, such as “efficiency,” their strategies can vastly differ. For a Frontend developer, efficiency means smooth loading and high responsiveness in a web browser. For a Backend developer, efficiency pertains to the server resource usage and cost-effectiveness.
These technology-driven areas can be approached in numerous ways, depending on the perspective. However, from a business standpoint, trying to maximize usability and profitability within a reasonable cost often results in vague directives like “cost-effective efficiency.” This makes it challenging to pinpoint specific methodologies, often leading developers to default to familiar technology stacks.
Using LLMs can simplify these dilemmas. By providing appropriate conditions and requesting code, the output can serve as a guideline. The key is to create prompts that require minimal technical detail while yielding satisfactory results. This approach ensures that any developer can easily align with the output, fostering better collaboration.
Eliminating the Gap in Code Reviews
The second point is about eliminating the gap in traditional code reviews. While code reviews help developers produce higher quality code over time, they can be a source of significant stress. The original author often prefers to stick to their approach, even if there are minor issues. Simple, agreed-upon changes can be made easily, but complex situations make code reviews difficult and time-consuming. This leads to high costs and delays in project completion. Culturally, code reviews can be an opportunity to build camaraderie, but they are still challenging and heavily reliant on individuals’ soft skills.
By using LLMs to generate code, everyone can take on the role of a code reviewer. Instead of traditional code review processes, developers can focus on refining the prompts and checking the output. This iterative process allows the team to align more closely with each other’s expectations. As the team reviews and adjusts the prompts together, they maintain a similar vision for the final product. Over time, this repeated collaboration fosters stronger bonds and more effective teamwork.
Bridging the Gap Between Planning and Execution
The third point is about closing the gap between planning and execution. Planning is often established by upper management based on the company’s grand vision. However, these plans frequently overlook the day-to-day tasks that need to be done. As plans trickle down to individual teams, they have to contend with ongoing tasks, backlogs, communication with other teams, and internal coordination. This results in increasingly rough plans with declining achievement rates.
Using LLMs can help reduce the time needed to reach planned goals. This allows for more detailed planning that aligns with the company’s aspirations. Employees can more easily align with the company’s direction, resulting in a more dynamic organization. With LLMs streamlining the execution of tasks, plans can be more specific and ambitious, fostering a more cohesive and agile workforce.

Programming in the LLM Era #2 — How

댓글