Thoughts
Ephemeral Software
Traditional software caps out at some point because it has to serve an audience. Agents remove that constraint, and the software they build has no reason to outlive the task. On what becomes disposable, what has to stay durable, and why design systems matter more when every interface is temporary.
→
Minus One to Zero
Before an idea becomes a company, the work is reducing forty live questions to three and finding a reason to commit. Minus one is where taste, real stakes, and cheap rejection matter most, and where killing bad ideas quickly can also kill slow good ones. As building gets cheaper, choosing what deserves to exist becomes the expensive work.
→
Designing Design
I truly believe Design is one and it is a transferable way of thinking for making complex systems understandable and trustworthy. It's an exploration of how experience, honest failure handling, and reusable systems compound judgment and taste. Especially as AI makes production and distribution more accessible, deciding what should exist becomes the harder task.
→
Virtue Ethics
AI systems reflect the character of their makers. Virtue Ethics reminds us that technology is never just technical. AI systems inherit the moral orientation of their builders, and only people who cultivate fairness, humility, and courage can build products that avoid harm.
→
Expert Loop
The future of AI depends on Expert-in-the-Loop design, where domain expertise is required to direct how systems reason and decide. Trust comes not from flattening complexity, but from centering those who bear the consequences.
→
Prompt Designing
Prompt engineering tweaks inputs to get better answers. Prompt designing moves beyond tuning answers to shaping how AI participates in human work. By building roles and structure, it treats prompts as design surfaces for safety, accountability, and trust.
→
Design in VC
I did not plan to end up in venture capital. In fact, no one can plan for it, not unless you are rich enough to start your own fund. Breaking into VC isn’t something you can simply walk into. You get there by knowing someone who knows someone who are the right person with significant access and capital.
→
Ephemeral Software
Traditional software caps out at some point because it has to serve an audience. Agents remove that constraint, and the software they build has no reason to outlive the task. On what becomes disposable, what has to stay durable, and why design systems matter more when every interface is temporary.
→
Minus One to Zero
Before an idea becomes a company, the work is reducing forty live questions to three and finding a reason to commit. Minus one is where taste, real stakes, and cheap rejection matter most, and where killing bad ideas quickly can also kill slow good ones. As building gets cheaper, choosing what deserves to exist becomes the expensive work.
→
Designing Design
I truly believe Design is one and it is a transferable way of thinking for making complex systems understandable and trustworthy. It's an exploration of how experience, honest failure handling, and reusable systems compound judgment and taste. Especially as AI makes production and distribution more accessible, deciding what should exist becomes the harder task.
→
Virtue Ethics
AI systems reflect the character of their makers. Virtue Ethics reminds us that technology is never just technical. AI systems inherit the moral orientation of their builders, and only people who cultivate fairness, humility, and courage can build products that avoid harm.
→
Expert Loop
The future of AI depends on Expert-in-the-Loop design, where domain expertise is required to direct how systems reason and decide. Trust comes not from flattening complexity, but from centering those who bear the consequences.
→
Prompt Designing
Prompt engineering tweaks inputs to get better answers. Prompt designing moves beyond tuning answers to shaping how AI participates in human work. By building roles and structure, it treats prompts as design surfaces for safety, accountability, and trust.
→
Design in VC
I did not plan to end up in venture capital. In fact, no one can plan for it, not unless you are rich enough to start your own fund. Breaking into VC isn’t something you can simply walk into. You get there by knowing someone who knows someone who are the right person with significant access and capital.
→