fuchen

Fin & AI

[上官亂 @shangguanluan]

在我心中，到目前为止，柴静所有的节目，没有任何一期超过了她第一期对圣战分子的采访。那是真正做到的新闻性，记录性，和理性的沉思。

因为那个时候，她还没有太明显的下意识的立场。

现在的节目都在你的预料之中，因为她在YouTube的舆论场里，有了下意识的立场。

在对徐宗懋的采访中，她努力想把徐引导向她想要的那个轨道，���者说，她以为徐在那个轨道上。可惜，徐并不是。

但是我有一种预感，柴静还会再转变。

[Patrick C Toulme @PatrickToulme]

There’s a big misconception about how GLM 5.2 was trained. Yes, they distilled Claude and GPT 5.5 — but distillation is not how they matched Opus quality. Distillation only fixed the cold start problem in RL.

RLing an agentic coding model isn’t rocket science. In simplified terms:

1. RL needs trajectories — rollouts where the model actually completed a task in some env

2. No successful trajectory on a task = zero gradient = you can’t RL it. This is the cold start problem

3. Distillation solves it. You seed your model with knowledge from a smarter one (Claude, GPT) on tasks it can’t do yet

4. Now it produces positive trajectories on those tasks

5. RL on those trajectories and hill climb agentic coding

6. At that point you no longer need to distill and can solely hill climb RL to better models

This is an interesting curve. I’d argue it’s harder to get to Opus 4.8 from scratch than to go from Opus 4.8 → Fable/Mythos tier.

GLM 5.2 is already producing positive trajectories, so they have plenty to RL on — they’ll keep climbing to Mythos quality without distilling any further. They no longer need American models.

[fuchen @IAmFuChen]

@JiankeF72763 @BeijingDai 防御很难的。激光武器不可能保护所有都市圈。唯一的办法就是进攻，摧毁对方的远程通信能力。

[fuchen @IAmFuChen]

@China_Amb_India 你干的好事。

[便衣道士 @Asahibozi]

彻底天塌啦，印度人在绍兴找到新澡堂啦

又有游客在绍兴日铸山庄拍到，据说在柯桥一带有3000人+的印度社区已经形成。。。😭😭😭

[fuchen @IAmFuChen]

Fantacy bug of world online

[etginlik @etginlik]

Hawaii'de yılda 2 kez hiçbir şeyin gölgesi oluşmuyor.

[fuchen @IAmFuChen]

@justTinaforu Glad to see you.

[歸藏(guizang.ai) @op7418]

一个预告

[fuchen @IAmFuChen]

想看

[空空道人 @Kongkongda5882]

王宝强这是疯了，还是赌上了全部身家？
2.5个亿，他没拿去买豪宅豪车，反而在山东农村，1比1复刻了一个1920年代的纽约唐人街，四条街，一砖一瓦，连招牌上的锈迹都给你做出来。
身边全是劝他的人，说直接去��视城租一个得了，何必呢，烧钱太快。可王宝强偏不，他要的就是那个真实感，演员一走进去，就得信自己活在那个年代，不是在演戏。
这次他拉来了梁家辉、段奕宏、狄龙，一帮老戏骨，要拍一部咱们自己视角的黑帮片，讲老祖宗在海外是怎么挺直腰杆子的。
这哪是拍电影，这分明是在造一个梦，一个关于尊严和骨气的梦。
有人说傻根变精了，我看他还是那个傻子，一个为了电影可以豁出命的傻子。
这股狠劲儿，才是真正的搞钱密码，赌的不是运气，是人心。

[Evrim Kanbur @WhileTravelling]

The Economist has a fascinating habit when it comes to China. If a city is poor, it’s proof the system failed. If a city becomes prosperous, safe, clean, affordable, well-connected, and pleasant to live in, then prosperity itself becomes the problem.

According to the article, Shaoxing is comfortable, prosperous, safe, convenient, culturally rich, and offers a lower-pressure lifestyle than China’s megacities. Yet somehow the conclusion is still “comfort meets constraint”.

Imagine applying the same framework elsewhere😅

Copenhagen: “Happiness meets boredom.”
Zurich: “Wealth meets excessive punctuality.”
Amsterdam: “Freedom meets inconvenient rain.”
Tokyo: “Efficiency meets social pressure.”

At some point every successful city can be turned into a cautionary tale if the goal is to find a cloud behind every silver lining.

The article acknowledges many of the things ordinary people actually care about like safety, reliable infrastructure, good public transport, clean streets, economic opportunity and affordable living relative to income. But then, immediately pivots into “yes, but…” mode.

From a development economics perspective, a more interesting question would be how did a country that was overwhelmingly poor 40 years ago create dozens of cities where millions of people now enjoy middle-class lifestyles?

China lifted roughly 800 million people out of extreme poverty over several decades. Whether one likes China’s political system or not, that achievement deserves credit rather than being treated as an inconvenient footnote. The economic transformation itself is one of the largest in human history.

The article also illustrates a broader tendency in Western commentary: quality of life is often measured differently depending on the country.

When a European city has CCTV, zoning regulations, government planning, and strong public transit, it’s called good governance.

When a Chinese city has extensive CCTV,
zoning regulations, government planning, and strong public transit, it’s called “constraint”.

I’d say The Economist is applying a double standard, so on brand.

A good economist starts with outcomes.
Are people safer?
Is purchasing power rising?
Is infrastructure improving?
Are opportunities expanding?
Do residents voluntarily choose to stay there?

If the answers are mostly yes, then the city deserves credit.

The real paradox is that some commentators seem uncomfortable whenever comfort appears in places they expected to fail.

@TheEconomist

[fuchen @IAmFuChen]

Salute!

[jietang @jietang]

@elonmusk @teortaxesTex won’t take that long

[fuchen @IAmFuChen]

@oopsie06020861 姐，���午��康！

[OpenCode @opencode]

GLM-5.2 is getting 3x the demand

If you are experiencing some problems, we are actively working on it

[Hongbo Zhang @bobzhang1988]

We are working on a cross platform UI in @moonbitlang , soon you are going to build literally everything in MoonBit: from desktop UI to server side to mobile and webpages and agents, in a single typed/verified tech stack!

[艾森 Essen @essen_ai]

中美大国博弈下，小国的生存法则还有效吗？2026年新加坡：创纪录的数据，却掩盖不了的战略不安 [https://t.co/f7OETH6i9v]

[fuchen @IAmFuChen]

big

[Vercel @vercel]

Introducing eve, an agent framework.

𝚊𝚐𝚎𝚗𝚝/
𝚊𝚐𝚎𝚗𝚝.𝚝𝚜
𝚒𝚗𝚜𝚝𝚛𝚞𝚌𝚝𝚒𝚘𝚗𝚜.𝚖𝚍
𝚝𝚘𝚘𝚕𝚜/
𝚜𝚔𝚒𝚕𝚕𝚜/
𝚜𝚊𝚗𝚍𝚋𝚘𝚡/
𝚜𝚌𝚑𝚎𝚍𝚞𝚕𝚎𝚜/

Like Next.js, for agents.
https://t.co/ezUIGLkSqG

[fuchen @IAmFuChen]

@yjh29640319 @China_Amb_India 准备好面对中纪委了吗？

[fuchen @IAmFuChen]

This is BIG!

[Hongbo Zhang @bobzhang1988]

We are working on a cross platform UI in @moonbitlang , soon you are going to build literally everything in MoonBit: from desktop UI to server side to mobile and webpages and agents, in a single typed/verified tech stack!

[fuchen @IAmFuChen]

可见辛格坡也是政治高压。为了政治正确，对电影里不经意表现出的历史事实都要刻意回避。

[如是我闻 @VicentYip]

辛格坡华人年轻人基本已经没��华人文化的认同感了！ #说着华语没华语认同 #屎格坡

[fuchen @IAmFuChen]

@QiuseOfLove @China_Amb_India 问你呢

[fuchen @IAmFuChen]

@China_Amb_India 傻逼。