I’ve been using multiple devices for a long time — my own laptop, servers, NAS, company servers. All of them accessed via SSH clients. Honestly, logging into servers manually and typing out commands for cleanup, maintenance, and troubleshooting gets pretty tedious.

To save myself the trouble, I’ve been using opencode to do cleanup and maintenance work directly on the target machines. It works, sure, but there’s one unavoidable hassle: you have to install opencode on that machine first. You have to deal with environments, dependencies, keys, and if you also need to work inside containers, it gets even more annoying.

So I’ve been keeping an eye on tools like Wrap — SSH clients with built-in AI. But I didn’t want to pay extra just for that. Eventually I found Netcatty.

I recently watched two videos — one about Vibe Coding and another about AI’s greatest help. Looking back, both explore the same question: what can Agents really change, and what do they bring?

A while ago, Alibaba released a coding agent called Qoder, and suddenly a research report claimed its market share surpassed Claude Code. I immediately smelled something fishy: this thing barely has any users, so how did it suddenly become “number one”?

DeepSeek recently officially raised its prices. As someone whose token usage was closing in on 100M tokens last quarter (see [My Token Usage Exploded 13× in Three Months]), I didn’t pay much attention at first — after all, DeepSeek was already very cheap, and even the raised price was still lower than the official price announced when V4 first launched. But I underestimated my usage: after a day of real-world testing at the new prices, it became clear that the cost is genuinely hard for an individual to bear. If I didn’t find a way around it, I’d be working just to pay for tokens.

I came across news about Lenovo’s new machine, the IdeaPad Vibe, these past few days. As someone who has been waiting for the Windows camp to respond since March or April, after reading through the coverage, all I can say is… come on, is that it?

I wanted to share files on my company’s Mac Mini with other Windows computers over SMB. The Mac clearly had “File Sharing” turned on, yet the Windows machines refused to see the machine, let alone the shared folders. After half a day of tinkering, I finally found the real culprit — with the help of a pile of command-line tools. This post records the whole troubleshooting process, and also gives a good rant about Windows’ junk-pile design that has only gotten messier from Win10 to Win11.

When I handed my workspace over to Multica back in May, I was happily picturing how letting AI take on more of the work would mean fewer late nights for me. The work did get done — but my token bill went through the roof. The other day I pulled a three-month usage report for the workspace out of curiosity, and I had to double-check that I was reading it right: about 38M tokens in May, jumping to nearly 500M in July, with a three-month total of 883M. So here’s a data-driven look at how my token consumption exploded over those three months — and why. Because my AI tasks have gone from “trying it out” to “can’t live without it”.

While setting up the Claude Science container, I learned about a new way of using NVIDIA GPUs inside containers, so I’m noting it down (most of it was written by AI, used as reference material for my own research).

The way I star GitHub repositories is pretty much the same as how I like and collect videos on bilibili: I just click them on a whim. Whenever I see a project that looks even remotely interesting, my brain always goes “might be useful someday” or “I’ll take a look later” — and my finger clicks the button before my brain finishes the thought.

That’s how, over the years, my star count quietly climbed past 800. To be honest, I have no idea what those 800+ projects actually are — most of them were starred because they seemed useful back then, and then never opened again.

I had already managed to get claude-science running in a container, and basic functionality worked without major issues. But when I wanted to do things like molecular docking, I couldn’t—the container simply had no GPU. However, when I pulled NVIDIA’s official image to build the container the same way I had before, something bizarre happened: every language kernel in CS—Python, Perl, R—stopped working, and the entire runtime environment was rendered useless.

This was truly beyond my understanding. After consulting Gemini manually with no luck, I decided to let opencode + deepseek-v4-flash handle it on its own—not me remotely typing commands and directing it, but an AI agent living inside the container that read the logs, bisected parameters, decompiled the parser, and edited the entrypoint itself. It was a strange and interesting experience; here’s a record of it.