与「WORK」相关的搜索结果

今天用 Kimi K3 给 CatReader 增加一广场的功能呢，其实改动量不少，包括前端后端和设计，差不多折腾了一天，上线了。Token 消耗比网传的少太多了。K3 的表现相当出色。 我觉得 Kimi 犯了一个宣传上的大问题，导致很多用户直接用 做 Coding 的事，但 Kimi App 做的是 Work 的事，用 Work 做 Coding，Token 消耗是急剧增加的。 我的 Vibe 群一堆用户用 Kimi App 而不是 Kimi Code cli 工具做编程，不烧 Token 才怪。

i have been spending a lot of time this past month out of town & in the hospital with my dad & family. he has narrowly avoided an additional amputation for now (yay), but soon he will be needing to go to either a nursing home ($8-10k/month) or an assisted living facility ($4k/month) for a few months where he can receive the care he needs. my family cant afford it. his insurance doesnt cover it. i HAVE to help him, so i'll be working on trying to figure out some kind of fundraising for his care. i'm currently taking donations for him through my Throne wishlist page. (Thank you to the donor who already contributed before this post ❤️) We have been thinking about starting a GoFundMe for him. Would that be a good idea? Would any of you be interested in contributing? In addition, maybe a new paid fan site? Open to hearing your thoughts, advice, ideas, and kind words. your voice is appreciated in these trying times ❤️‍🩹

这也太爽了吧！躺在床上说说话就把小说写了，还能赚钱？ 视频演示了 WorkBuddy 从七猫扫榜到自动写小说的完整链路： 📊 扫七猫新书榜 → 找选题、找热门书、找对标书 🔍 拆解标杆前3章 → 提炼能复用的套路 ✍️ 按拆解出的设定 → 自动写出小说正文 从选题到写作，全自动化 ✅ 写出来自己能看，也能直接发到小说平台，流量跑起来还能赚钱 💰 感兴趣的朋友，蹲一下我后续的完整教程 👇

We tried using Meta's new Muse Code agent, but it has a bug that doesn't let it sign in from a docker container. So we did a fun experiment: Meta claims Muse Spark 1.2 was co-trained with their Muse agent harness. So we extracted instructions from their system prompt and added them to the Cline harness. TL;DR of this special prompting: - Trust source code over the user prompt, so read every call site and existing tests before starting the task - Weigh edge and error cases as heavily as the happy path - Always reproduce the bug before fixing - Don't trust the first passing test suite, and verify suspicious looking half-baked tests - Never stop at just editing, keep working until the change is verified complete. We then asked this modified harness to fix a real bug from our repo, and compared the results to the original Cline agent harness. Results: - Used 2.7x fewer tokens (19.7M → 7.2M) - Finished 2x faster (49min → 24min) - Cost 2.4x less ($7.69 → $3.25) Same Muse Spark 1.2 model, same task, only the prompting changed. Incredible how much of a performance gain Meta was able to achieve training it on these special instructions!

AI makes launching ideas easy, but keeping an edge is tough. 💡 A company’s true moat lies in intangible assets like data, workflows, and brand equity. Join global leaders to discuss how AI is rewriting innovation at Singapore IP Week 2026 (Aug 26-27):

昨天和朋友们吃饭，聊到 deepseek-v4-flash 对行业的影响，有个朋友抛了个问题：拿它替代员工，一个月 30 天、每天 8 小时，得花多少钱？ 我粗算了下：几百块，撑死了。 他反问：那 3w 月薪的工程师，这份高薪还能撑多久？ 后来我们又往深了聊：真要替代真人，企业该怎么搭系统、workflow 和组织架构？留下来的员工得具备什么能力？这些想清楚，这个方向才算真正落地。

Codex、WorkBuddy 等 Agent 已经能读取 PDF。面对扫描件、多栏排版、跨页表格和公式时，解析质量仍会直接影响后续的检索、问答和数据提取。 Mac 上可以本地运行的 PDF parser 很多，PaddleOCR、GLM-OCR、MinerU 等工具的安装方式和运行环境各不相同。文档类型变化后，想换一个 parser 重新处理，通常还要重新配置。 DocDot @docdotai 把多个本地 parser 集中到了同一个 CLI 中管理。 官方安装命令： curl -fsSL | bash 安装 DocDot 后，可以在 NanoDoc、PaddleOCR、GLM-OCR、MinerU、LiteParse 之间安装和切换，并将 PDF 解析为 Markdown 或 JSON。后续还会继续增加新的 parser。 DocDot 也提供了 Web Mode。运行 `docdot web` 后，可以在浏览器中并排比较 3 个 parser 的结果。遇到复杂表格、公式或特殊排版时，可以根据实际输出选择更合适的解析器。 根据官方说明，安装程序还能扫描本机的 Codex、Claude Code、OpenClaw、Hermes 等 Agent，并配置相应的 Skill；需要接入其他应用时，也可以通过 MCP 调用。解析在本机运行，PDF 无需上传，目前主要面向 Apple Silicon Mac。 我在 Mac 上安装了 DocDot 和 NanoDoc，并准备了一页包含中英文、表格和公式的 PDF。 首次启动完成模型编译后，再次解析同一份 PDF 用时约 2.3 秒。正文和公式基本准确，表格结构得到保留，小字号表头仍有少量误识别。 这只是一份单页样本，不能替代完整的性能测试。官方技术报告使用了 4,231 页、64 类文档进行比较，NanoDoc 的综合质量和 Table TEDS 在参测解析器中均排名第一。 DocDot 今天在 Product Hunt 上线：

On January 15, 2025, Falcon 9 successfully deployed Firefly Aerospace’s Blue Ghost Mission 1 and ispace’s RESILIENCE lunar landers on a trajectory to the Moon from pad 39A in Florida. For most of our missions, we plan a controlled deorbit of the Falcon second stage so it safely reenters over the ocean. For higher energy missions like those to a lunar transfer orbit, nearly all performance on the vehicle is devoted to successfully placing the payload in the intended orbit, and a controlled disposal maneuver is not always possible. We actively work to be as responsible as possible with hardware left in space to ensure space safety, including for more complex missions. In this case, over time, solar activity and gravity led the second stage toward the Moon. Impacts like this are rare, but they can happen with objects in these types of orbits, and we worked with NASA on the optimal disposal solution. Our focus remains on advancing reliable access to space while working toward even more sustainable operations with Starship in the future. As a fully reusable vehicle, Starship is designed to eliminate expendable upper stages entirely, reducing hardware left in orbit and enabling even more missions to the Moon, Mars, and beyond.