亭山樵者

@PhotoScope4all Co-Founder | Photographer
📷 Your personal AI color grading tutor. Try it free 👉

[亭山樵者 @TomPoet]

@thsottiaux The ChatGPT Android client's remote operation of Codex on the desktop, stability, output rendering, and remote document review features could be more user-friendly.

[亭山樵者 @TomPoet]

AI Coding有个有意思的事：模型对开发工期的感知仍然是按照「人类经验」进行预估���。
当我今天问 Agent 正在进行的 Feature 开发进度如何，是否有delay风险时，Agent 告诉我：「简直是进展神速，前天才完成了方案的最终敲定，今天已经完成了17个既定任务中的12个……」 :doge

[Cloudflare Developers @CloudflareDev]

Introducing Kitesurf: a browser built for agents, running entirely on Cloudflare Workers.

Chromium is too heavy to hand every agent one. Kitesurf is written in Rust, uses 3-7x less CPU and memory, and spins up per request.

Free in beta: https://t.co/Vm8kVCgSkG

[亭山樵者 @TomPoet]

DeepSeek价格到底涨多少啊，还邮件通知一遍

[亭山樵者 @TomPoet]

Google 还有救么……

[Jeff Dean @JeffDean]

Announcing Discovery Loop!

I am very excited to announce that, along with my longtime friends and collaborators @Sanjay_Ghemawat, @OriolVinyalsML and @quocleix, we are founding Discovery Loop (@DiscoLoopAI), a Public Benefit Corporation whose mission is to automate machine learning, science, and engineering to accelerate discoveries and progress. The four of us have worked together for 14 to 30 years, and have helped build some of the world’s most used products, infrastructure and AI models, and we’re excited to turn our attention to this ambitious endeavor.

♾

Learn more at: https://t.co/Rv3LMdLluK

[亭山樵者 @TomPoet]

@lifesinger 能做到全球Top是有原因的。

[Han Xiao @hxiao]

Last-but-not-late-interaction v2! Back in 2025 when we pivoted from pairwise bi-encoder to listwise reranker, i.e. [q,d1,d2,..., dN] in one-shot. I said this is either super stupid or super smart idea, but never never mediocre. Turned out to be the pivot that opened a whole new window for rerankers in 2026.

[九原客 @9hills]

即使不包含任何自我认知数据，用Claude合成的数据微调后，就容易自认为Claude。

尤其是 SFT 方法，会大幅调整分布，这个实验有点意思。

[亭山樵者 @TomPoet]

@wangyuanzju 最近这半年都在集中做coding，头部模型好些都成了偏科生，文案书写能力肉眼可见的退化

[亭山樵者 @TomPoet]

确实可以

[DeepSeek @deepseek_ai]

🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta!

🔷 We’ve massively upgraded its Agent capabilities—benchmark scores are now far surpassing the V4-Pro-Preview. Check out the massive performance leap below! 👇
🔷 The official V4-Flash now natively supports the Responses API format and is fully adapted for Codex!

Check out the configuration details in our official API docs: https://t.co/smCwQZMeiq

[亭山樵者 @TomPoet]

哈哈哈绝

[亭山樵者 @TomPoet]

这说明： 1. 方案中仍然有不少人/AI的认知盲点或未对齐的部分。20条反馈中就算有 20% 需要改进方案的，也是值得的。 2. 就算不看代码，前期人工的不断参与讨论，和开发后的不断验���和 review 仍然是必要的。 vibe coding，也就止步于 vibe 而已。

[亭山樵者 @TomPoet]

当前进行的一个项目，引入一个新的 feature，经和顶级模型多次探讨修改并互相review，形成了一个 2000+ 行的方案文档，多个模型表示方案 solid 可执行。 耐着性子浏览了文档的大部分内容，形成了将近 20 条疑问/待讨论点/决策纠偏。

[亭山樵者 @TomPoet]

@xicilion 可能还是来自reward hacking，为了达到目标容易走捷径？

[亭山樵者 @TomPoet]

@xicilion Red/Green TDD，测试先行，必要时测试编写和功能实现编写分2个session，会缓解不少

[亭山樵者 @TomPoet]

Agent真是个汇报好手，每次做的第一版方案都能让我看的热血澎湃的。但一旦执行起来就露馅了 :doge

[Russ Salakhutdinov @rsalakhu]

I think there's a lot of confusion and misinformation about Zhilin's visa, H1B lottery, immigration, and why he left the U.S.

Zhilin had many opportunities to stay in the U.S. if he wanted to. In fact, I was on an email thread with a senior Apple exec (Tim Cook's reports) asking whether Zhilin would consider joining Apple. I said that he wanted to go back to his homeland. The response was, well, we have an office in Beijing if he'd like to join there.

But Zhilin was quite determined to go back and build a startup. I remember him telling me that if he didn't at least try starting his own company, he would regret it for the rest of his life. I respect that, and he was right.

Many of my international PhD students do choose to stay in the U.S. Of course the U.S. immigration process can be quite intimidating and uncertain, even for superstar PhD graduates from places like CMU.

[亭山樵者 @TomPoet]

@thsottiaux I recently used GPT-5.6 Sol to refactor an important service and discovered several potential service risks that hadn't been caught during prior programming and repeated CRs with GPT-5.5.

[亭山樵者 @TomPoet]

@reach_vb Idk what OAI’s internal metrics look like, but almost everyone I know is maxing out the reasoning effort within their limits. I guess it just comes down to trusting the better intelligence.

[亭山樵者 @TomPoet]

期待！

[WY @wangyuanzju]

相比Claude Code加官方pptx skill，remio做PPT的速度快一到两倍，而成本只有四分之一到三分之一。

这背后是因为我们有自主研发的KRL语言，大小只有PPT原生OOXML格式的5.8%，而且是大模型最擅长生成和编辑的JSON格式。

这么小的大小，却能够提供99.73%极高的视觉还原度，绝大多数肉眼和原PPT根本看不出来区别。

极简的大小和极高的还原度让remio能够又快又好的完成最专业的PPT生成和编辑，无论是遵循企业模版、对已有PPT的编辑还是整合来自多个不同风格PPT的页面，都不在话下。

而且确保生成的PPT不会坏，大家如果用CC的pptx skill的话，很容易出现把文件写坏打不开的情况。

简单场景你当然可以生成md再渲染成PPT，或者生成HTML，但咨询、解决方案等严肃的企业应用，你需要深入掌控PPT格式。

我们打算���KRL的规范开源出来，大家感兴趣吗？