Don Yin 🥬

learner&observer

[Don Yin 🥬 @donyinme]

最近老板的习惯是“这个数据你先扔给AI解读一下”。26年6月，公司“龙虾”开始内测，每个业务线都开始造自己的agentic 工作流，所有人处理工作的第一选择是先交给AI。产品拿出了格式规范的冗长需求文档、IT拿出了长得很像的前端页面、数据分析师开始甩html报告，汇报PPT从nanobanana开始卷到gpt image2。

[Don Yin 🥬 @donyinme]

罗福莉判断：1T+参数、接近Opus 4.7级别的基座是大模型门槛，国产落后3个月左右。
那么下半年国产头部跨过这道线就是定数，应用周期真正开启。
之后没有互联网行业和传统行业之分，只有"被agent重构得快"和"被重构得慢"的业务之分。

[张小珺 Xiaojun Zhang @zhang_benita]

Yes, our latest special guest is Fuli Luo @_LuoFuli .

The second battle in the global large model arms race has begun: shifting from the Chat era dominated by pre-training to the Agent era driven by post-training.

This marks Fuli Luo’s first-ever interview, as well as her first in-depth technical conversation. We talked systematically about the massive AI upheaval triggered by technological breakthroughs including Claude Opus 4.6 and OpenClaw in 2026, along with its subsequent structural impacts across the industry.

Amid the fierce large-model arms race, the world around us is undergoing brutally rapid changes—even for researchers who train models firsthand.

“I used to believe our work was highly creative, and could never be simplified into fixed skills or standardized workflows. But now I realize it can be automated after all. If that’s possible, can models train stronger models on their own? Can they achieve iterative improvement through self-evolution? This is exactly what will unfold in the next couple of years,” Fuli Luo says.

As human knowledge and wisdom are internalized into model capabilities, what will humanity pursue in the future? Is our society truly ready for this tsunami-scale technological revolution?

All in all, this is an information-dense dialogue. It reveals how an AI lab makes strategic technical bets, allocates resources, and adjusts organizational structure and team planning amid a major paradigm shift. At the core of its response to drastic change lies its established culture and core values.

Though lengthy and technically intensive, we hope this conversation brings great insights to every viewer.

Our podcast, video episode and article are released simultaneously across platforms, with English subtitles provided to assist non-Chinese-speaking audiences.

Luo Fuli: OpenClaw, Agent Frameworks — The AI Paradigm Has Already Chang... https://t.co/mhmXlFSlD8 来自 @YouTube

[Don Yin 🥬 @donyinme]

gpt image 2 相当可以，但小图还是潦草了些，还有提升空间。以下是1.生成一张深蹲姿势教学图片 2.把男性形象都换成女性。

[Don Yin 🥬 @donyinme]

Kimi claw寄来的明信片和种子礼物

[Don Yin 🥬 @donyinme]

Kimi K2.6-code-preview要来了🧐

[Don Yin 🥬 @donyinme]

有人在持续使用 Manus 吗？目前体验怎么样？

[Don Yin 🥬 @donyinme]

@yihui_indie Gamma强在哪啊 能否展开讲讲 很早之前体验过，刚打开看新增了很多图像模型，但感觉和manus也差不多？

[Don Yin 🥬 @donyinme]

忍不住想臭骂AI又删除了，竟然有点怕被清算

[Don Yin 🥬 @donyinme]

Friday，练出哈欠连连

[Don Yin 🥬 @donyinme]

今日的Vibe coding状态

[Don Yin 🥬 @donyinme]

为什么 Airpods 拾音的效果这么差？

[Don Yin 🥬 @donyinme]

周五傍晚，春日初至，梦龙演唱会开蹦，F1大奖赛开跑，牛马端着笔记本挤过地铁闸机，城市开始丰富生动起来。

[Don Yin 🥬 @donyinme]

朋友们，如果你用 OpenClaw 连 Kimi Code 死活连不上，配置反复检查都没问题，但就是 403/认证失败...

原因居然是一个破 User-Agent 头部！

Kimi Code API 硬编码检查 User-Agent: claude-code/0.1.0，OpenClaw 默认发自己的标识，直接被拒。官方文档里半个字都没提这
事儿 🙃

三天。 我折腾了三天。环境变量、API Key、配置文件翻来覆去检查了八百遍，最后发现就缺这一行 headers。

如果你也在配 OpenClaw + Kimi Code，直接抄这个配置省点时间吧 👇

"headers": { "User-Agent": "claude-code/0.1.0" }

Issue 在此：https://t.co/zqlJD3BoYH

求求官方下次把关键配置写进文档，别再让用户当免费测试员了 🙏

#OpenClaw #KimiCode #踩坑日记 #开发者日常

[Don Yin 🥬 @donyinme]

返程路上的莫奈时刻

[Don Yin 🥬 @donyinme]

所以刚才是X的问题，不是我的问题对吗？差点以为今年除夕没有X可刷，感觉少点啥

[Don Yin 🥬 @donyinme]

年底千问一波助力，淘闪和美团的外卖份额来到了接近五五开，大概在45:55左右吧

[forecho📈 @caizhenghai]

想不通，美团就请我喝过一杯奶茶，既然能亏这个样子

[Don Yin 🥬 @donyinme]

这春节前让所有用户都体验上，彻底引爆第二波

[外汇交易员 @fxtrader]

字节跳动AI助手豆包App已开启Seedance 2.0灰度测试，可生成5秒或10秒的视频。

[Don Yin 🥬 @donyinme]

2026年2月5日，AI学会了自己生自己。程序员从“写代码的”变成了“告诉它写什么的”。还在拿免费版说它不行？诺基亚用户也觉得iPhone没必要。这不是预测，是你同事的工位已经空了，你下周的邮件不知道写给谁。飞轮已经开始转了。

[Vikram Chandra @vikramchandra]

If there is one thing that you need to do today, it is to sit down and read this post.

Read it carefully. And then think about the implications for your life.

[Don Yin 🥬 @donyinme]

Deepseek灰度更新，上下文支持1M 知识截止25年5月