Zhengzhou

Codex mobile is here — and it’s blowing up!

I’d recommend a code agent product I’ve been using recently: OpenClacky. I topped up $20 and have been using the DeepSeek V4 Flash model. So far, it’s been surprisingly cheap — I haven’t even used up $1 a day.

One issue I’ve run into though: when the agent is actively working in the terminal, if I try to enter a new task at the same time, the whole thing just freezes and stops responding for quite a while. So now I usually wait for the current task to finish before sending the next instruction.

Huashu Design , Say one sentence to your agent — Claude Code, Cursor, Codex, OpenClaw, Hermes all work.

3 to 30 minutes — you ship a product launch animation, a clickable App prototype, an editable PPT deck, a print-grade infographic.

Not "decent for AI" quality — it looks like a real design team made it. Give the skill your brand assets (logo, colors, UI screenshots) and it reads your brand's voice; give it nothing and the built-in 20 design vocabularies still keep you out of AI slop territory.

teco-lab.com, this website for learning spoken English through English vlogs is really well done. A lifetime membership only costs RMB 39.9. Since I watch English vlogs quite often, I had actually thought about building a similar English learning website myself. But now it seems I don’t need to anymore.

heroui.com - Beautiful, fast and modern React UI library. (Previously NextUI)

fumadocs.dev is a beautiful and flexible documentation framework for any React.js project.

A few days ago, I tried translating a PDF ebook using Ollama with a local Qwen 3.5:4b model. Translation speed wasn't an issue, but the final layout turned out poorly. My workflow was to first convert the PDF to Markdown, translate it, and then generate a new PDF. However, the bilingual vertical layout (Chinese-English) in the resulting PDF was hard to read and very uncomfortable.

I looked into this problem and realized PDF text layout is indeed difficult to handle. Different books have different content and require different presentation styles. Optimizing and adjusting the layout for each book costs too much effort.

Finally, I found that HTML is the most suitable choice—styling is easy and results in a clean layout, and my goal is simply to translate and read comfortably. With this approach, I'll continue to improve it later and use it to help me read English books.

Today I saw Vercel's Lab open-source portless, which allows accessing local projects without using port numbers and eliminates the need to worry about port conflicts.

- "dev": "next dev" # http://localhost:3000

+ "dev": "portless myapp next dev" # https://myapp.localhost

In the early stages of MVP, the priority should be on making rational decisions in business logic and focusing on running through core processes, with monolithic applications and local development being prioritized. UI, interaction, permissions, deployment, etc., should be addressed in subsequent phases.

It's crucial to avoid wasting energy on non-core aspects, reduce cognitive load, and quickly validate feasibility.

Matt Pocock Skills is truly impressive!

AI is no longer just working in isolation, crafting a plan and then spewing out a solution without engaging in the process. Instead, it starts by raising questions based on my product ideas, addressing each point one by one. If there are doubts, it provides feedback, then proceeds to analyze and confirm.

The rhythm is perfectly balanced—it draws me into the product design process. This sense of collaboration is fantastic. It doesn’t just hand me a complete solution outright; it avoids wasting tokens by working in isolation.

Everyone can learn to cook, yet restaurants and fast-food chains haven't disappeared.
Ingredients and recipes are abundant, even with step-by-step tutorial videos available, but many people still choose to order takeout.
This indicates that demands are diverse, and most things cannot remain scarce indefinitely.
What's truly scarce is exceptional taste, quality, dining experience, and service.

If the Apple EarPods could be redesigned with a magnetic connector, it would be perfect.

I ordered a Hazzys short-sleeve shirt and a pair of trousers on Douyin. After receiving them and trying them on, they fit very well.

I’ve bought this brand before—aside from the relatively high price, I’ve been quite satisfied with it overall.

You don’t need many clothes; one or two sets of formal wear are enough.

Threads has a great product feel—clean and minimal. It looks simple, yet feels very comfortable to use. It’s exactly the kind of style I like.

Its interaction design is also highly worth studying and learning from.

https://www.threads.com/

“MacBook Pro 16-inch in Space Gray, powered by the M2 Max chip (12-core CPU + 38-core GPU), with 96GB RAM and 8TB storage. US version, minor cosmetic wear, battery health at 100% with only 6 cycles, AppleCare+ valid until July 2026. Price: ¥21,588.”

Given the configuration and the price, it feels like a very solid deal.

My current primary machine is a 2023 Mac mini with 32GB RAM and 1TB storage. It was sufficient at first, but with the increasing use of AI-assisted development tools, memory has gradually become a bottleneck.

I also have a MacBook Air (24GB + 1TB), mainly for mobile work, though in reality it doesn’t get used that often.

So the situation now is: I have multiple machines, but none that truly serve as a high-performance, all-in-one solution. Taking everything into account, I decided to upgrade to a machine with significantly higher memory that can also handle mobile scenarios effectively.

I reached out to Yige (a fairly reliable second-hand electronics seller), and once they confirmed availability, I placed the order right away. After the new machine arrives, I plan to sell my current two devices back to her to offset part of the cost.

Most AI coding tools now support multi-agent workflows, meaning you can build different features in parallel. But in practice—even on small personal projects—I haven’t found this approach particularly helpful. The codebase grows rapidly, and beyond potential conflicts, the biggest downside is that it just feels exhausting.

With a more traditional way of coding, moving from thinking to implementation, it’s much easier to get into a flow state. Ironically, actually writing code is what helps me relax.

For a simple feature, the real challenge is making it intuitive and well-designed. That’s where meaningful collaboration with AI should happen. Your thinking process needs to work alongside AI to produce code that’s actually useful and sustainable over time.

Back in August 2025, after getting the Feiniu × Beelink co-branded NAS, I gradually picked up two ZhiTai 7100 4TB SSDs. They felt pretty expensive at the time, but I went for them anyway for the silent setup.

The experience has been great so far—no issues at all.

What I didn’t expect is how much prices have gone up now. They’ve gotten so expensive that I can’t afford them anymore. The AI boom hasn’t just soaked up capital—it’s also driven up the cost of memory and hardware across the board.

It’s kind of crazy. This level of hype and price inflation feels similar to what we saw during the real estate boom.