For many developers, tvOS presents a unique challenge. From design concepts and hardware limitations to debugging, it stands apart. Ronnie W. shares her experiences, highlighting the ups and downs of developing for Apple TV. fatbobman.com/en/posts/survi…

順手用 Gemini + Notion MCP 做了一個輕鬆的命令； 這樣發版本的話，完全不影響我的傳統習慣。真舒服。 gemini --yolo "1. 讀取 notion page「eisonAI 更新日誌」；只需要回報「最新版本」的更新內容，不附帶其他版本；2. 把「最新版本」內容，翻譯到自然的英語，然後寫入

Codex (and similar AI CLIs) are terrible at writing 「Apple AppIntents」. They often generate code that doesn’t even compile. This is not a prompt problem. It needs a domain-specific Agent Skill. But instead, people build “skills” for: commit messages, renaming variables,

有趣，不錯。本地可用。性能優越。 huggingface.co/inferencerlabs…

I’m somewhat disappointed with Lightpanda. 1. Seeing the “headless browser” label, I expected something on the level of headless Chrome. 2. It's written in Zig, which makes it very portable. 3. I even had Codex port it to tvOS, and it runs. 4. But after testing, I found that

I’ve been using this small script for ~2 years to send text input into the tvOS Simulator (via Raycast + AppleScript keystrokes). It’s been a reliable helper for testing flows without an easy on-screen keyboard path. Sharing it in case it helps other tvOS devs.

--enable-videotoolbox --enable-hwaccel=hevc_videotoolbox --enable-encoder=h264_videotoolbox this is not easy. github.com/qoli/FFmpegKit Real-time transcoding solution for black screen when playing H265 with AVPlayer

MLX-Video vs Unreal Egine 5 🤣🔥 Who is wining? Vote below 👇🏽 > purpvfx (Animator, VFX & Graphic Designer) > Prince Canuma (MLX King)

現在就是一個偉大的 AI 時代。 那些面向開發者的維護 app 翻譯字段的工具。竟然被 Gemini CLi 一句話指令搞定了。

youtu.be/O5PI868ApCI?si…

到底誰家開發播放器是這樣開發的？

Ronnie W. thanks for trying us! rn we offer JS execution without graphical rendering. iFrame support is on the roadmap would love to hear from anyone who has feedback on how to improve Lightpanda. DM me if you're up for chatting

Created another MLX-video movie with LTX and scenario prompted by / with ollama launch (glm-4.7) about a few people (inspired by Prince Canuma and Ivan Fioravanti ᯅ ) that lead the local AI gen revolution. Have fun. Took about 2 hours on my M3 Max 128

zai-org/glm-4.7-flash 真的有實力作為本地 ai model。然後，clawdbot 感覺不如 opencode.

ChatGPT 最愛就是事後諸葛亮。

1. 我切換到了 openai/gpt-oss-120b ; 2. Clawdbot 調用瀏覽器很垃圾，結果我要用 Codex 修理它這個能力； 3. 我認為，這種黑盒 Bot 還是太難用了，其實能力上和 Codex / Claude Codex 這些 AI CLI 是差不多的，就是一個入口的問題而已。 4. 垃圾 LM Studio，我懷疑它有一個內部 5 分鐘 就

其實，只需要有 CDP 協議，CC 就是你最節省 Token 的 AI Agent。 我發佈的任務是這個： 你的任務是這樣的：透過控制我的 ARC 瀏覽器（ronnie-Mac-Studio.local:9222）引導我完成 gmail-fetcher.js 缺少的資料。

雖然它是一個舊項目， 但是，我覺得比 Clawdbot 要省 token。 Clawdbot 啟動對話就需要 7k ～ 11k token。 但是 Gemini CLI 就很省。 圖一（大蝦）為 Clawdbot； 圖二（超級大頭）為 Gemini CLI。 github.com/qoli/Gemini-CL… 這是一個以 Gemini CLI YOLO 模式和 telegram