I plan to capture my thoughts spontaneously as they come , embracing the raw, unedited facets and potential misunderstandings.
Initially, I began a dialogue with GPT about the essence of agile within Japanese warrior culture.
Really? What was I thinking?
Let's kick things off with the core premise of this piece.
GPT: The Unexpected Parallel to Reverse Polish Notation
When I started using GPT, I expected it to streamline my internet research. However, I was struck by a surprise: my interaction with GPT paralleled the reverse Polish notation, adhering to a "Last In, First Out" (LIFO) approach. This process not only emphasized my latest findings but also infused unexpected depth and spontaneity into my content.
Divergent Dialogues and Their Recursive Essence:
Dialogue, especially during divergent activities like brainstorming, often follows recursive patterns. While GPT interactions might appear straightforward, they surprisingly drive swift cycles of idea divergence and convergence. And unlike traditional mind maps, there's the stamina to revisit and refine continuously. This process meshes well with design thinking, maintaining a consistent flow of ideas while subtly guiding them towards convergence.
Like recursive functions in programming, each new inspiration starts the cycle anew. This process can be compared to navigating a maze using depth-first search. I had two challenges: first, I often ran out of energy to explore them all; second, listing these paths in order became tedious. Thankfully, GPT provides a solution to these challenges.
Reference:【Depth-first search - Wikipedia】 https://en.wikipedia.org/wiki/Depth-first_search
Test-Driven Blog Writing with GPT:
Incorporating test-driven development (TDD) principles into blog writing using GPT was an effective way to increase the quality and depth of articles.
To me, writing an English blog using GPT feels akin to coding in 'C language'. When I think of it this way, it actually seems to apply to virtually all projects and, perhaps, even to life itself. The emotions and procedures feel similar, and it might even be perceived as rustic or simplistic.
First, visualize the desired outcome, then pinpoint the steps needed to achieve it. Once that's set, all that remains is to execute. This is often referred to as a "as-is/to-be" analysis.
In programming, we cater to both machines and humans. Similarly, when I write a blog, I'm mindful of an audience that spans both my internal and external worlds. The output from GPT, often dubbed as "GPT language," has been ood balance. This may cause things to get lost in translation and cause some confusion, but such deviations can provide new insights and ultimately benefit the team. I believe that.
The approach comprises the following five steps:
- Setting the Test: Before initiating the creation of the article, clear expectations for the output from GPT are defined. This can be aligned with specific keywords, themes, or answers to particular questions. An essential aspect here is my personal criteria of sensing "beauty", "justice", "value", "essence", and "insights". This transforms me not just into a writer but also a tester.
- Minimal Coding: GPT is inconvenient due to the restriction on the maximum number of characters and the inconvenience of not having any tools to support structured writing, the initial questions or prompts inevitably have to be brief and simple. This coincidentally matches with the TDD principle of writing minimal code.
- Review and Refinement: Upon receiving responses from GPT, a comparison with the set expectations is made. Often, this process brings unexpected analogies and inspirations. These new discoveries act like recursive functions, deepening the thought process, amplifying understanding, and sometimes even initiating a completely new topic. As a result, the article's resolution and insights are enriched.
- Iterative Process: Continuously refine the content to meet the blog's requirements and standards.
- Final Review: A concluding check and adjustments are made to finalize the content.
Much like the structure of a narrative, the initial test set might momentarily be forgotten as the dialogue unfolds. However, as the conversation converges, it resurfaces and gets tied back, serving as a foreshadowed plot point. The mid-stages of this process can sometimes feel naive and foolish. If a colleague were to witness these stages, I would probably be overwhelmed by deep feelings of embarrassment. Yet, for me, peak performance often goes hand in hand with vulnerability. It's akin to the crane from the folk tale "Tsuru no Ongaeshi".
Reference:【Tsuru no Ongaeshi - Wikipedia】 https://en.wikipedia.org/wiki/Tsuru_no_Ongaeshi
The first theme was:
The metaphor that resonates most with the test-driven development I've honed over the years is the imagery of the Satsuma Jigen-ryu. Could it be that the ninjas from Mikawa and the swordsmen from Satsuma have remained in people's memories because they thrived in the chaotic VUCA era? I posit that these traditions formed the groundwork for the Agile culture in Japanese manufacturing in the wake of the post-World War II uncertainties.
Reference:【示現流 - Wikipedia】 https://ja.wikipedia.org/wiki/%E7%A4%BA%E7%8F%BE%E6%B5%81
