Popy Holy

@luo87633 太棒了 优雅至极

24 hours of rainfall data, turned into a real-time particle-based rain landscape. Every value reshapes the curtain—from a quiet drizzle to a sudden storm. Built with Three.js + WebGL. #threejs

@Kevinxie1037437 叫你朋友顶住，AI是未来，只是还未来

朋友买中天亏麻了，昨天割肉止损，再跌的话，他就真的销户

@xin_pai88825 超棒啊！不过之前看到过类似的网站， AI coding时代Idea护城河真的低到没有🤔

做了个图像编辑工具，可以很快速的生成 ASCII、复古、半色调等效果。每个效果都可以调整～

@0xLogicrw 微调模型 有很多反例，我感觉不是真的

桥水基金旗下人工智能应用实验室（Bridgewater AIA Labs）与 Thinking Machines 合作，成功将微调模型应用于金融信息筛选，实现了超越前沿大模型（GPT 5.5 和 Claude Opus 4.8）的准确率。在金融文章相关性、央行文件分析等六项日常投资任务中，这款微调模型将错误率降低了近三成，而推理成本仅为前沿大模型的十四分之一左右（实现了 13.8 倍的降幅）。这一研究展示了企业通过特定领域微调实现「差异化智能」的潜力。 研究团队在 Tinker 平台微调 Qwen3-235B 基础模型。由于直接从非专业标注员获取的标签存在大量错误，团队设计了验证纠错机制，将模型预测与标签不一致的困难样本交由投资专家重审。最终，训练出的模型准确率从基线模型的 44.8% 提升至 84.66%，超过了 GPT 5.5 的 78.2% 和 Claude Opus 4.8 的 78.0%。 微调方案包含三项核心改进：一是交错批处理，按顺序交替训练不同任务，避免混合训练带来的干扰；二是引入非对称裁剪的损失函数优化采样；三是使用在线策略蒸馏，在模型偏离教师分布时进行惩罚，且每 20 步在验证集表现创新高时动态更新教师模型。

@Jackywine 是不是这么牛先，能复刻 细节中的魔鬼 吗

我发现了一个问题 优秀的设计还是少，但是高质量的抄袭速度变得贼快 只要你走抄这条路，AI 可以疯狂加速 🤔怎么说呢？有种众泰汽车皮尺部的感觉 我们叫逆向工程，但是又和原研药和仿制药的区别，🤔这个区别到底有多大呢？ 用二手设计，或者是抄来的设计，也能解决问题，但是……

@NataliyaRo4040 没有社招进的吗 南网

@LaurenceMister 怎么个差法，单调、无趣？用起来不顺手？

codex 的 UI 设计很差，有什么办法提高一下？

@Dott_Chen 你的很多观点我蛮赞同的也引人思考。我也有一些观点，对话框更像是未定义具体功能的入口。允许无限意图输入不一定等于全部意图实现。预定义在不同场景下有他的优点和缺点，每次为用户提供一个没有预期的东西并不是在全部场景下都成立。

最近 Linear 把首页改成了 chat interface，引起了不少争议。作为一个过去三个月一直以 agentic 方式工作的人，我可以很肯定地说，这就是所有软件未来应该有的样子。 UI 的发明，本质上是为了向我们人类这种不天生理解数据、也不知道如何处理数据的生物，解释抽象的数据层。我们需要图表来可视化数据，需要按钮来操作数据。但这一切都建立在。直接与数据打交道是困难的，这一前 AI 时代的前提之上。时代变了，软件也需要进化。 pre-defined GUI 在与 AI agent 原生协作时会成为瓶颈。用户的意图是无穷的，但传统软件只能通过固定的 UI 模式提供有限的几条路径来满足这些意图，效率太低了。提供给用户最原始的意图输入入口，然后让 AI 直接去操作数据层，才是面向未来的解法。

@usmanameer I'm currently acting like a new agent you've never encountered before, learning about you from scratch through your posts, just as you fed your conversation history to a new agent Memory is related to the scenarios we encounter. And different Agent companies do different thing

The agent memory format problem nobody is solving MCP standardizes how agents talk to tools. A2A standardizes how agents talk to each other. In 2026, we have 15+ protocols for agent interoperability. None of them touch what actually matters: what the agent learns. Not how it communicates. Not how it executes. But what it remembers. ⸻ Right now, there is no standard for agent memory. No shared schema for: •decisions it made •patterns it recognized •mistakes it learned from •edge cases it discovered after months of work Every agent stores this differently. Every framework invents its own format. And none of them are compatible. ⸻ Consumer chat memory is becoming portable. Claude, Gemini, ChatGPT → export preferences → import elsewhere That’s progress. But that’s not agent memory. ⸻ Agent memory is operational. It’s: •what your coding agent learned about your codebase •what your workflow agent learned about your processes •what your research agent inferred across hundreds of sessions When you switch agents, that layer doesn’t move. It disappears. ⸻ Every framework locks you in differently. Markdown. SQLite. Graph-based storage. Lerim ≠ Hermes ≠ SAGE. You don’t migrate. You start over. ⸻ And the cloud makes this worse. Every developer has seen this story before: •Parse shutdown •Heroku free tier gone •Firebase pricing shifts •Vercel terms changing overnight Now apply that to memory. ⸻ If your agent’s memory lives on someone else’s infrastructure: It exists at their discretion. Not yours. ⸻ Today: •Claude Code Routines resets every run (by design) •Managed Agents store decision traces on vendor servers One change: •pricing •policy •product direction And months of accumulated knowledge is: gone, locked, or unusable. ⸻ This isn’t hypothetical. It’s the default lifecycle of every platform wave. ⸻ heartbeat’s learnings.jsonl is a different bet. Not a service. Not a platform dependency. A file. On your machine. ⸻ Structured. Portable. Git-friendly. Readable by any tool that speaks JSON. ⸻ Anthropic shuts down? Your memory survives. Switch models tomorrow? Your memory survives. ⸻ The format becomes the moat. Not the model. Not the loop. The memory. ⸻ So the real question is: What should the open standard for agent operational memory look like? Because right now: Everyone is building systems that quietly depend on the answer… …and nobody has defined it github.com/uameer/heartbe…

@skywind3000 已经有了，给恶心坏了。可疑动作监控→员工风险值评估、员工违规行为AI总结🤢

现在主流的办公安全软件都会每分钟定时截屏保存到服务器，过去没人力监控审查，未来可以用 ai 分析图像看你每分钟到底在干嘛了，有疑点再人工介入或者主管提醒。