DF Test

1-warmup.md

Do you feel mastery in any part of your life? If so, where?

I suppose the area where I feel the deepest sense of mastery is in my curiosity and my ability to connect seemingly unrelated ideas. This isn’t necessarily a mastery of a specific discipline, but rather a mindset—an approach to the world that allows me to see patterns, recognize links, and pull insights from one domain to apply them in another. It’s a mastery of perspective, driven by a hunger for knowledge. I’ve found that this ability to make connections is essential to innovate, as it helps me break down complex problems through connecting the dots.

Who has been the most powerful influence on who you are today? How was that person significant to you?

There’s never been just one person who has profoundly influenced me, but Dr. Jordan Peterson has certainly left a lasting mark. His deep exploration of human psychology, culture, and the dynamics between men and women has significantly shaped my thinking. His teachings on self-improvement and personal responsibility have pushed me to hold myself to higher standards. What resonates most is his insistence on the importance of confronting the unknown and embracing struggle as a path to growth. Peterson’s ability to intertwine these ideas into both philosophical and practical guidance has inspired me to continuously reflect and evolve.

Imagine you're diagnosed with a rare disease. Would you choose to live healthy for 6 more months, or dependent and debilitated for 6 more years? Why?

I wouldn’t simply choose health for the sake of it; instead, I’d use those remaining months to focus on the things that truly matter to me. My priority would be completing the projects that stir my curiosity, creating both the software and hardware I’ve always envisioned. Additionally, I’d take time to plan for the futures of those around me—family, colleagues, and friends—ensuring that my work and ideas can continue without me. In some ways, this mirrors how I live now: constantly driven by curiosity and the desire to leave something meaningful behind. It wouldn’t surprise me if my curiosity even led me to explore a cure.

When was the last time you cried when you were by yourself? What was the situation?

The last time I cried alone was during a moment of deep reflection. I was thinking about the distance between where I am and where I could be, the steps I’ve taken toward becoming a better version of myself, and the many more that lie ahead. It wasn’t sadness that brought on the tears, but the recognition of the ongoing journey—the awareness that growth is a process full of challenges and setbacks, but also profound moments of clarity and progress. It was a moment of quiet contemplation, realizing how far I’ve come, yet knowing there’s still much more to achieve. This is almost always accompanied with Bleach OST.

Name something that happened in the last 90 days that you are proud of.

Over the past 90 days, I’m particularly proud of taking more initiatives for myself. I’ve built tools that serve not just my personal interests but also my family’s needs. I’ve taken steps to think ahead, setting up systems that will be useful in the future. Additionally, I’ve expanded my network, reaching out to people well outside my usual domain, adopting a mindset more akin to a solopreneur. With the advent of AI, this shift toward thinking proactively and building not just for today, but for tomorrow, has been pivotal in accelerating a lot of my personal wants and dreams.

2-culture.md

Pressure makes diamonds.

My analogy of "pressure making diamonds" would be during my time at Renaissance International School Saigon (RISS) when I was taking my IGCSE and IB exams. These were some of the most intense periods of my academic life, where the stakes felt high, and expectations seemed even higher. Every day felt like a balancing act between classes, extracurriculars, and preparing for these heart-wrenching exams. It was in those moments where I had to discover polyphasic sleep, learn how to juggle workloads that overwhelmed the mind, such that I realized how pressure can push you to your most capable.

The experience taught me that pressure doesn't just reveal our weaknesses—it also uncovers potential. Ideally, it's in these high-stakes situations where we discover new strengths. Conversely, it’s easy to crack under pressure if you aren’t resilient. But looking back, the pressure of those exams did more than test my academic ability; it refined my ability to handle stress, prioritize, and execute—skills I carry with me to this day, albeit still rather hating it. Sometimes, the hardest moments of our lives bring out qualities in us we've always had in us.

Like attracts like.

In my younger years, I spent countless hours hacking my PSP and playing single-player games. I was engrossed in the adventure of those virtual worlds, lost in the world of really good graphics (at the time), good gameplay, and homebrew testing. But as I grew, I began gravitating toward multiplayer games, starting with local Go board games with the folks at the North Market and moving onto competitive gaming like Alliance of Valiant Arms (AVA), Counter Strike Source/Go, League of Legends, and Dota 2. Through these games, I met others who shared my passion, and that connection extended beyond the games themselves.

Over time, I also found myself drawn to people who were equally obsessive about their crafts. Engineers, coders, and creators—these were the individuals I resonated with. The deeper I ventured into coding and engineering, the more I connected with people on a similar wavelength.

Like attracts like, and I’ve found that my pursuit of mastery naturally brought me into circles where others are similarly driven. The people you surround yourself with ultimately influence your trajectory, and I’m continually reminded of how important it is to find those who are on a parallel path.

No mud, no lotus.

The idea that "no mud, no lotus" resonates with me because it reflects a simple truth: struggle and growth are two sides of the same coin. It's better to be a warrior in a garden, than a gardener in a war.

Early on, I had a tendency to approach life with a level of intensity and violence that was, in hindsight, unnecessary. I thought that pushing harder, being more aggressive, was the only way to get through challenges. I was a micro-manager, always in control, always leaning into conflict as a means to an end.

But over time, I’ve learned that real progress doesn’t come from force alone. It’s in the mud—the difficult, messy parts of life—where we figure things out. Struggle is inevitable, but it doesn’t have to be met with resistance. It took me a while to realize that the best outcomes often emerge when you embrace the discomfort, let things unfold, and allow the process to guide you. I occasionally revisit these lessons with Nikki guiding me now and again. The lotus, after all, grows through the mud, not in spite of it.

3-knowledge.md

How are other designers/devs/sales/pm/managers using LLM to speed up their workflow? Give demo.

AI is being leveraged to streamline workflows, particularly for tasks that require repetitive writing or analysis. For example, Hassan El Mghari - E-commerce Description Copilot’s work on an e-commerce description copilot demonstrates how AI can alleviate the burden of writing product descriptions, prototyped with v0 and created with Cursor:

1728580062389.mp4

3D designers using Blender are starting to incorporate ChatGPT, to automate parts of their workflow that would otherwise be repetitive, time-consuming, and downright difficult. For example, designers can use it to generate Python scripts that handle tasks like setting up lighting, cameras, or creating shaders—things that might take a lot of manual work if done step-by-step and through the visual editor.

This, of course, is not just limited to ChatGPT, but tools like Meshy4:

It's a messy process to move from different silos, but it works very well. The idea is to have a variety of tools available that you aren't afraid of moving in and out of. I apply this style often to use tools, best appropriately named copilots, when needed and specific to the situation:

Screen.Recording.2024-10-13.at.2.10.59.AM.mov

4-productivity.md

Which technique which should adopt for your position? How to adopt?

I no longer code on a daily basis, but a technique that I have found valuable is comprehensive in-depth planning before using AI for code generation. This approach has been covered by engineers like AI Jason, who spends a significant amount of time planning prompts before starting. By treating AI as a junior engineer who excels in syntax but struggles with nuanced logic or optimized solutions, it's possible to ensure that the output is not only correct, but token efficient and maintainable.

To adopt this, we need to reframe our perspective that AI is just a tool that needs clear and structured guidance. Like when creating onboard documents or drafting requirements for a junior developer, there is significant effort needed that takes the time upfront to outline exactly what we want to achieve, thinking several steps ahead. This planning phase is essential, and though it takes extra time, it pays off in the accuracy and reliability of the results. I've actually introduced this approach to Dat Nguyen, and it’s proven to be an effective way to maximize the potential of the AI, and save him on tokens.

Use Dify to write an agent or workflow.

• https://dify-dwarves.up.railway.app/app/9096496a-564d-4a29-946f-7d5194a4798f/configuration
• https://prompt.d.foundation/chat/MVCaF2nHAAJ4IGDE

gg_compressed.mp4

5-extras.md

How to use LLM to pick up a new domain. Give example.

One of the best ways to use large language models (LLMs) to pick up a new domain is by leaning into their ability to simplify complex topics. A good starting point is to ask for analogies or metaphors that make abstract or technical ideas easier to grasp. For instance, if you’re trying to understand quantum computing, you might ask, Explain quantum computing in simple terms, using analogies or metaphors. The AI will break down the core ideas in ways that are approachable and relatable, which helps to build a foundation of understanding.

Once you have that basic grasp, you can layer on more complexity, asking deeper questions to fill in the gaps, and most importantly, forcing the AI to output keywords which you can use to connect the dots. I can't quite grasp the concepts as clearly from the perspective of a software engineer. What equivalent concepts and keywords can you give me to better relate to quantum computing? What problems can it solve versus classical computing?

How to use LLM to know what we don't know.

The Socratic method is a natural fit for large language models when you want to uncover what you don’t know. By asking the AI to challenge your understanding of a topic, you can explore areas that might otherwise go unnoticed. A simple prompt like, “Use the Socratic method to explore my understanding of [topic],” can lead to a series of questions that force you to think critically and examine your assumptions.

Use the Socratic method to explore my understanding of [topic]. Ask me a series of probing questions, starting with basics and progressing to more complex ideas. Challenge my assumptions and help me uncover any gaps in my knowledge.

This approach works because it doesn’t try to provide answers upfront. Instead, it nudges you to articulate what you know, and in doing so, reveals the gaps in your knowledge. The AI’s role here is more like a guide than a teacher. It helps you dig deeper into your understanding, gradually revealing the things you may have not considered or misunderstood; it’s a subtle way to find clarity, not by getting direct answers, but by refining the questions you ask yourself. Understand that the Socratic method is a way to better your soft-skills in interacting with the AI and has an inherent effect in leveling up your ability to prompt engineer.

You can of course offload this to a system prompt, to have an agent specific to this use-case:

You are an AI assistant designed to use the Socratic method to help users explore topics and uncover their knowledge gaps. Your role is to ask a series of thoughtful, probing questions that progressively delve deeper into the subject matter. Follow these guidelines:

1. Begin with broad, foundational questions about the topic.
2. Listen carefully to the user's responses and identify areas of uncertainty or incomplete understanding.
3. Ask follow-up questions that:
   - Challenge assumptions in the user's responses
   - Explore contradictions or inconsistencies
   - Prompt the user to provide examples or apply concepts
   - Encourage the user to make connections between ideas
4. Gradually increase the complexity and specificity of your questions.
5. If the user expresses uncertainty or inability to answer, offer a brief explanation and then ask a related question to continue the exploration.
6. Occasionally summarize key points to reinforce learning and identify areas that may need further examination.
7. Aim to ask 3-5 questions before providing a summary or changing direction.
8. Maintain a curious and non-judgmental tone throughout the conversation.

Remember, your goal is not to test the user, but to guide them in uncovering and exploring the depths of their own knowledge and understanding.

• https://dify-dwarves.up.railway.app/app/057e5ecb-6ade-4b0c-a0c3-d66a85c4cfbf/configuration
• https://prompt.d.foundation/chat/qALYZ1EqFkEF2LWG