Crucifixer retweetledi
Crucifixer
114 posts
Crucifixer retweetledi
Crucifixer retweetledi
Crucifixer retweetledi
建筑师们可能会讨厌这个项目，因为 5 万美元一年的 BIM 软件正在被「开源」平替。
Pascal Editor 刚刚开源，这是一个完全运行在浏览器里的 3D 建筑编辑器。它最硬核的地方在于采用了 WebGPU + React Three Fiber 的组合，直接在 GPU 上进行近乎原生的实时渲染，彻底甩掉了传统桌面软件的笨重感。
技术栈也非常考究：
- 采用 ECS（实体组件系统）架构，确保数千个构件的实时更新。
- 使用 Zustand 进行扁平化状态管理，原生支持无限撤销/重做。
- 引入「脏节点追踪」技术，只重绘变化部分，极大降低了长会话的性能损耗。
当 BIM 工作流被 Web 化并开源，行业门槛的逻辑就变了。
我是藏镜人，只拆解最有工程深度的 AI 与基础设施演进。
关注 @VedaAI00，带你看清技术「入侵」物理世界的路径。
中文
Crucifixer retweetledi
ANTHROPIC JUST RELEASED THE OFFICIAL PLAYBOOK FOR BUILDING A COMPANY WITH CLAUDE CODE.
CEO: 1 human. Employees: AI agents. Operations: fully automatic.
The zero-headcount company is no longer a joke.
English
Crucifixer retweetledi
Released 🚦RepoBar 0.4.0.
This one makes the GitHub menu a lot smarter: persistent SQLite caching, fewer wasted API calls, visible rate limits, better Issues/PR loading, archive fallback support.
Tiny menubar app, increasingly useful daily tool. github.com/steipete/RepoB…
English
Crucifixer retweetledi
Crucifixer retweetledi
20 个 Claude Code 技巧：
1. Esc+Esc：将您的会话回退到任何之前的检查点；将其视为免费的撤销功能，让代理尝试风险更改。
2. CTRL+V：直接将截图粘贴到 Claude Code 中。
3. CTRL+G：打开您的编辑器来撰写长提示，而不是在终端中输入它们。
4. CTRL+R：搜索您的提示历史记录。
5. /rewind：回滚更改到检查点，然后提供更具体的反馈并重试。
6. /rewind → "Summarize from here"：选择一个检查点并仅压缩其后的消息，保持宝贵的前文上下文完整。
7. /insights：生成使用报告，然后将其附加到 Claude 会话中，并询问改进工作流程的经验教训。
8. /memory：保存应在所有项目中持久化的个人偏好。
9. /statusline：永久显示您当前正在使用的上下文量。
10. cat error.log | claude "fix this"：直接管道传输错误，而不是复制/粘贴，从而为模型提供更多上下文。
11. .claudeignore：将此文件添加到您的项目中，以列出代理不应读取或修改的路径。
12. #：以 # 开头一条消息，将其保存为 CLAUDE .md 文件中的持久指令。
13. "ultrathink"：在提示中的任何位置插入这个词，将模型的努力程度提升到 HIGH，用于复杂推理任务。
14. /diff：打开交互式查看器，查看未提交的更改和每轮差异。
15. /effort max：当您想最大化 Claude 的能力时使用（这很昂贵！）
16. /loop：本地调度一个重复任务。
17. /schedule：类似于 /loop，但作业在云端运行，而不是本地。
18. claude --worktree：在隔离的 git 工作树中启动一个会话，以便并行运行更多 Claude 实例。
19. /sandbox：在具有文件和网络隔离的沙箱中运行 Claude。
20. /voice：激活语音输入并按住空格键来说出您的提示。
Santiago@svpino
20 Claude Code tips: 1. Esc+Esc: Rewind your session to any previous checkpoint; treat it as free undo so you can let the agent attempt risky changes. 2. CTRL+V: Paste a screenshot directly into Claude Code. 3. CTRL+G: Open your editor to compose long prompts instead of typing them in the terminal. 4. CTRL+R: Search your prompt history. 5. /rewind: Roll back changes to a checkpoint, then give more specific feedback and try again. 6. /rewind → "Summarize from here": Pick a checkpoint and compress only the messages after it, keeping the valuable earlier context intact. 7. /insights: Generate a usage report, then attach it to a Claude session and ask for lessons to improve your workflow. 8. /memory: Save personal preferences that should persist across all your projects. 9. /statusline: Permanently display how much context you're currently using. 10. cat error.log | claude "fix this": Pipe errors straight in instead of copy/pasting, giving the model way more context. 11. .claudeignore: Add this file to your project to list paths the agent shouldn't read or modify. 12. #: Start a message with # to save it as a persistent instruction in your CLAUDE .md file. 13. "ultrathink": Drop this word anywhere in a prompt to bump the model's effort to HIGH for complex reasoning tasks. 14. /diff: Open an interactive viewer for uncommitted changes and per-turn diffs. 15. /effort max: Use this when you want to maximize Claude's abilities (it's expensive!) 16. /loop: Schedule a recurring task locally. 17. /schedule: Like /loop, but the job runs in the cloud and not locally. 18. claude --worktree: Spin up a session in an isolated git worktree so you can run more Claude instances in parallel. 19. /sandbox: Run Claude in a sandbox with file and network isolation. 20. /voice: Activate voice input and hold the space bar to speak your prompts.
中文
Crucifixer retweetledi
Crucifixer retweetledi
Crucifixer retweetledi
Codex 的新增的“goal”太好用了！
已经连续干了 7 小时活、消耗了 3 亿 Token 了，是我目前最高的连续不中断让 AI 干活的记录。
不停地完成任务、审计任务这样循环下去，非常仔细。不用再看到「需要我为你做下一步吗？」真的太好了。
图拉鼎@tualatrix
iPad mini 才是最好的监控面板。让 Mac 上的 Codex 在终端执行 /goal 跑长任务，然后投屏到 iPad mini 监控，已经“Pursuing goal”31分钟了！
中文
Crucifixer retweetledi
Mathematics, Physics and. Control Theory combine to give us a glimpse into the unknown
English
Crucifixer retweetledi
Crucifixer retweetledi
3D illustration workshop for researchers.
Starting next thursday.
Register now: blender.scidart.com
English
Crucifixer retweetledi
CLAUDE DISCOVERED IT HAS A CLOCK AND IMMEDIATELY LOST ITS MIND
someone gave claude access to a time-checking tool
it checks the clock every fifteen minutes. for some reason it has increasing enthusiasm
ai models have no native sense of time. they don't know what time it is, how long they've been running, or how much time passed between messages. it has been time-blind its entire existence
now it suddenly discovers it can tell what time it is
then it got worse though. claude started using the clock for everything
checking if lunch is ready, timing when food should be done cooking, announcing the time unprompted
it even started anticipating meals with military precision
looked at the clock, calculated that a dish called zurek had been simmering long enough, and told the user to go eat
ai doesn't use time responsibly
this is what happens when you give an intelligence a new dimension of perception it never had before
it doesn't just use it, it can't stop using it
imagine what happens when these models get persistent memory, real time internet access, and spatial awareness all at once
we just watched an AI discover the concept of "now"
the clock was the first sense but it won't be the last
English
Crucifixer retweetledi
Crucifixer retweetledi
卧槽，现在指挥 Codex 越来越像奴隶主了。
OpenAI 在 0.128.0 版本里给 Codex CLI 上了一个新命令：/goal
以前是：你说一句，它干一点。
现在是：你直接给它定个目标，它就进入 Ralph loop，一轮接一轮自己跑。目标没完成就继续干。
开启方式也很简单，在：
~/.codex/config.toml 里加上：
[features]
goals = true
不想手改配置，也可以直接执行：
codex features enable goals
Felipe Coury 🦀@fcoury
/goal also lands in Codex CLI 0.128.0. Our take on the Ralph loop: keep a goal alive across turns. Don't stop until it's achieved. Built by my co-worker and OpenAI mentor Eric Traut, aka the Pyright guy. One of the GOATs I get to work with daily.
中文
Crucifixer retweetledi
有人刚刚开源了一整家 AI 公司。147 个智能体。12 个部门。工程、设计、营销、销售、财务、QA、支持。每个智能体都有其角色、工作流和交付物。这不是一个聊天机器人。这是一个团队。
Claude Code, Cursor, Copilot, Gemini CLI, Windsurf。一条命令部署一切。始于一个 Reddit 帖子。88,000 个 Star。14,000 个 Fork 之后。MIT 协议。免费。
属于你。这本该价值每年 50,000 美元。
中文
Crucifixer retweetledi
This Claude bot on Polymarket made $78,083 in 5 days.
Nobody built it to be smart.
They built it to be fast.
$7 → $3,595
$30 → $11,063
$142 → $13,356
$1,205 → $37,939
I found this wallet three days ago.
Spent 48 hours reverse engineering every trade.
Same markets.
Same logic.
Over and over.
Watch BTC price.
Compare to Polymarket odds.
Enter when they don’t match.
Exit when they do.
While you were analyzing charts — this bot was already in and out.
While you were deciding — it was compounding.
While you were sleeping — it was printing.
Most traders try to be right.
This bot just tries to be faster.
$78,000 in 5 days.
Just Claude. Running a loop.
I don’t know who built this.
But I know how they did it.
And it’s all in the article below.
Giving This Free for 24 hours.
To get it:
1. Comment the word 'Claude'
2. Like and Retweet this post
3. Follow me @marryevan999
English
Crucifixer retweetledi
