Hello, I’m Liulian
Categories: Notes
Hello, I’m Liulian.
This site is not another list of things AI can theoretically do. I care more about a practical question: when a promising tool becomes part of a real workflow, can it produce useful results consistently?
I want each post to read like an experiment that can be checked later—why I started, how I broke the problem down, what failed, and what remained useful. The conclusion does not have to look impressive, but the process should be real.
What I’ll write about
Four themes are at the center of the notebook:
- AI content workflows. More than prompts: how source material enters a system, what an Agent handles, where human judgment belongs, and how the final output is reviewed.
- Hands-on tool tests. I am testing Wan3.0 for IP-character video and exploring ways to connect cloud-drive source material to an Agent workflow. I will document repeatable steps, failures, and boundaries.
- Product experiments. Before adding features, I prefer to test whether a small idea solves a real problem. These notes will show that incremental validation rather than retroactively polishing every attempt into a success story.
- Independent creator practice. I am learning how content, products, automation, and a distinct point of view can support one another over time.
Why a blog and a weekly
Liulian Weekly is the radar: it collects and filters things worth following from the week. This blog is the workbench where one question can be examined in depth.
Both sites begin with open-source foundations. This blog uses Tw93’s MIT-licensed theme. I keep the attribution and licence while publishing none of the original author’s posts, images, or personal data. To me, using open source well means respecting its boundaries and then doing the work of adding your own voice.
How I’ll document the work
I will follow three simple rules:
- I will not present an untested idea as a settled conclusion.
- I will not invent a more exciting history than the one that happened.
- When a process can be shown, I will not show only the outcome.
Starting this week, I will keep writing here. The posts may not arrive on a rigid schedule, but each one should leave behind a reusable method—or at least an honest judgment.
If you are building an AI workflow, testing tools, or validating a small product, you can find me on GitHub, X, or in the comments below.