BCDX2–CX3 and DX2–CX3 complexes assemble and stabilize RAD51 filaments
I didn’t expect this classification task to be so simple—simple enough that a Hello World-level scikit-learn script, with minimal tweaks and some hard-coded rules, could produce a reasonably robust and accurate detector. Most of the effort was just waiting for LLMs to generate data…
，这一点在搜狗输入法2026中也有详细论述
工具一览计划与执行：Obsidian + Claude CodeObsidian 是我的主要任务管理工具，让 Claude Code 读取 100 Days of SwiftUI 课程总览页面，就可以直接根据课程内容在 Obsidian 中生成一则 Tasks-100 Days of SwiftUI 待办清单，当中用：，这一点在体育直播中也有详细论述
Nature, Published online: 03 March 2026; doi:10.1038/d41586-026-00595-9
Other ideas: detect AI-generated images. But with Stable Diffusion and easy LoRA fine-tuning, generated styles are far more diverse—this task would be much harder. I could also crawl Lofter data to analyze AIGC pollution per tag. But writing this blog has burned through my three-minute enthusiasm. Maybe next time.