📢 早安!Gate 廣場|4/5 熱議:#假期持币指南
🌿 踏青還是盯盤?#假期持币指南 帶你過個“放鬆感”長假!
春光正好,你是選擇在山間深呼吸,還是在 K 線裡找時機?在這個清明假期,曬出你的持幣態度,做個精神飽滿的交易員!
🎁 分享生活/交易感悟,抽 5 位幸運兒瓜分 $1,000 仓位體驗券!
💬 茶餘飯後聊聊:
1️⃣ 假期心態: 你是“關掉通知、徹底失聯”派,還是“每 30 分鐘必刷行情”派?
2️⃣ 懶人秘籍: 假期不想盯盤?分享你的“掛機”策略(定投/網格/理財)。
3️⃣ 四月展望: 假期過後,你最看好哪個幣種“春暖花開”?
分享你的假期姿態 👉 https://www.gate.com/post
📅 4/4 15:00 - 4/6 18:00 (UTC+8)
Using Codex also has new pain points:
Sluggish architecture migration is a common problem for every AI, especially noticeable on GPT 5.4, with various hacky compatibility layers.
Additionally, Codex adds test coverage every time it makes changes, and the test files are bigger than the implementation! While it looks stable, it's really slow.
The result is that refactoring is extremely, extremely slow! Plus, every time I have to say "continue", I've been sitting at my computer for a full 2 days and Codex is still doing the same thing!
The image below shows Claude's sharp criticism of Codex: