Digital Humans · research notes in public
Less scrolling. Better human possibilities.
The system can do the first pass. You keep the final say.
I am a rising senior studying whether a private AI concierge could understand what matters to you, do the exhausting first pass with other systems, and return only a few human possibilities worth your time. The point is not to automate the relationship. It is to make more room for the right one.
I am not presenting a finished company, a launched dating product, or a proven matching result. I am building the research habit around a hard product question: can an AI reduce the exhausting parts of connection without reducing human choice?
My stance: the AI can listen, coach, filter, and explain. It cannot impersonate a partner, make the final decision, or quietly turn private insecurities into a product advantage. Consent, inspection, deletion, authenticity, and a clear no are core features.
The best version would help someone understand what they want, show up as themselves, and meet a real person with less noise around the connection. This is still the work before the launch. The hard part starts after people trust you with the truth about themselves. I am still proving I can build something worthy of that trust.
Why this question
Connection is a product problem before it is an AI problem
I am Chinese American and grew up in central Indiana (Greenfield, Indiana) around a family restaurant in New Palestine. A system either survives contact with a real person or it does not. That lesson follows me into Serve IT handoffs and AI research: build for the person standing there, not for the screenshot. The restaurant floor taught me respect before I had a research vocabulary for it, and it still makes me ask whether a product is reducing work or merely moving it onto someone else.
Full hire narrative lives on IU Pages and LinkedIn. Ships live on jadexzhao. This door keeps the thinking I am still earning.
Essays on this door
The idea, the evidence, and the boundaries
- Essay The concierge question Private context, agent-to-agent first passes, and how to make matching less exhausting without taking the choice away.
- Essay Places and evidence What public Digital Human examples can teach, what they cannot prove, and why a face is not the same thing as trust.
- Essay · 中文 网页与社交 · 2027 Limits first, then the handoff checklist, then why web and social still matter and which tool classes to use.
- Profile About Who I am when I am not writing the demo story.
How I work
A small standard for a consequential product
- Start with the person Name the ordinary moment Who is asking, what do they need, and what would make the interaction worth returning to?
- Make the system legible Disclosure belongs in the experience People should know what they are using, where an answer came from, and when a human can take over.
- Design for the day after A handoff is part of the product The work is not finished when the page ships. Someone still has to update it, answer for it, and improve it.
Want to think next?
The next version starts with a better question
Start with the concierge question. If something is wrong or thin, email me. I would rather revise in public than protect a polished story. I am still proving I can do the hard part after launch day: protect the user, maintain the system, and make the next decision better.