简体中文 / English
“We do these things not because they are easy, but because we thought they should be easy.”
I build AI agents, developer tools, and software experiments that tend to begin with the words “How hard could it be?” I prefer practical tools, local-first thinking, and learning by making things slightly more complicated than expected.
| Ref. | Experiment | Field notes |
|---|---|---|
01 |
Stella | Model-agnostic QQ-group AI agent. Local or online LLMs, extensible tools, and cost-conscious design. |
02 |
Caligo | Experimental Rust framework for modern NTQQ protocol clients. |
03 |
Athanor Studio | Structured documents, robust conversion, versioning, and information preservation. |
04 |
Kevin_Kelly-Game | Experimental harness for evolving coding agents through controlled self-rewrite, mutation, and independent evaluation. |
index.status: expanding · Selected public projects from my own repositories, not the full archive.
while tokens_available():
claim()
experiment() # "How hard could it be?"
if rate_limited():
sleep(3600) # Officially classified as rest.The rate limit is just a suggestion, right? (This is a joke. Probably.)
Toolchain / Installed modules
Open to interesting conversations and collaborations involving AI agents, local-first software, developer tooling, or just an idea that sounds a little unreasonable.
Transmission address: 1694717255@qq.com
astral@observatory:~$ exit · Connection closed. Curiosity still running.