rYe
2.2K posts
阿迪王！你好！ @DarioAmodei 告诉你一个天大的好消息！小米的大模型负责人 @_LuoFuli 本人在采访中承认。她在深度用你的 Claude模型，我们中国的人民群众用你模型整天像贼一样提心吊胆怕你封号，雷军用你东西，你倒是挺慷慨是吧，行，我看出来了，你麻痹就是个势利眼，哪里有钱往哪里，雷军给了你多少好处？啊？老实说！狗日的！！
中文
You planted this bamboo yourself. With enthusiasm. For a screen. Within two years you regretted it — and within five you understood you could not go back.
Running bamboo (Phyllostachys and related genera) is one of the most frequently self-inflicted garden disasters in Britain. The gardener plants it voluntarily, pays for it, and then pays considerably more to try to remove it.
What the garden centre tells you: fast-growing, evergreen, architectural, forms a screen in two seasons. All true.
What the garden centre does not tell you
The rhizomes of running bamboo (Phyllostachys, Pseudosasa, Sasa, Pleioblastus) spread laterally through the soil at one to two metres per year. They do not grow downward like normal roots — they run horizontally at 15-30cm depth, in every direction, under the lawn, under the border, under the patio, under the neighbour's fence. Every rhizome node produces a new culm that emerges without warning — in the middle of the lawn, between paving stones, through a neighbour's garden.
In three to five years, a bamboo planted as a three-metre screen has rhizomes occupying 50-80 square metres — the entire garden. In ten years they reach neighbouring gardens, irrigation pipes, house foundations, and drainage systems.
Cutting the canes does nothing
Bamboo cut to the ground regrows with more vigour than the year before. Every cut stimulates the rhizomes to produce emergency culms. Cutting weekly for an entire season does not weaken the root system — it causes it to expand faster in search of undisturbed ground.
Digging it out is a serious undertaking
Phyllostachys rhizomes are as hard as beech wood, interlaced in an underground network extending in every direction. Hand digging requires days of work per square metre. Mechanical excavation must reach 50-60cm depth across the entire colonised area — and every rhizome fragment left in the soil starts again. A single forgotten node restarts the problem from zero.
The root barrier
The only reliable way to contain running bamboo is a root barrier — a sheet of HDPE (high-density polyethylene) at least 1.5mm thick, installed at minimum 60cm depth, surrounding the entire planting area before the bamboo goes in. The barrier must be continuous with no joins or weak points. It costs more than the bamboo itself. If it was not installed at planting time — which covers the vast majority of cases — it cannot be retrofitted without excavating everything.
The distinction nobody explains at the garden centre
Running bamboo (Phyllostachys, Pseudosasa, Sasa): rhizomes that travel metres from the original plant. INVASIVE. Never plant without a root barrier.
Clump-forming bamboo (Fargesia, Borinda): rhizomes that grow as a compact clump without running. NOT INVASIVE. Fargesia is the correct choice for a British garden screen — it forms a dense hedge, stays where it is planted, and will not colonise neighbouring gardens.
Eradication protocol
For running bamboo already established, the method with the highest documented success rate is:
Year 1: cut all canes to ground level in June — when the plant has spent its stored energy on new growth and rhizome reserves are at their lowest. Cover the entire colonised area with thick black weed-suppressing membrane, edges weighted or buried. Leave covered for the full season.
Year 2: remove covering in spring. Cut every shoot that emerges immediately. Re-cover. Repeat through the season.
Years 3-4: monitor weekly. Cut every shoot within 24 hours of emergence — a bamboo shoot left for a week is already a metre tall and has re-stocked the rhizomes. Consistency is everything.
Alternative: systemic herbicide (glyphosate) injected into hollow cut stems immediately after cutting in late summer. The same protocol used for Japanese knotweed — to which this problem is closely comparable in persistence and difficulty.
The most architectural screen in the street is also the most expensive to remove. Consider it before planting, not after.
English
rYe ретвитнул
One of the most important things that Elon taught me:
People engaging in fraud are always the first and loudest to complain.
English
身份换了，是付鹏建立x账号的原因
付鹏 2024 年底在 HSBC 闭门会那次刷屏发言之后，微信公众号被封了。他和高善文的微信号在系列言论后被封。 2025 年 4 月底他正式从东北证券离任，最近这几天（2026 年 4 月）以首席经济学家身份加入了港股上市的新火集团（Sinohope，1611.HK，前身是火币科技，李林系）。新火集团 4 月 20 日因任命付鹏为首席经济学家，股价盘中一度大涨 22%，他将聚焦 FICC 和 Crypto 方向的宏观研究
东北证券任上发的话，规制责任在国内监管。Sinohope 是港股上市公司，持有香港 SFC 的 1/4/9 号牌，他是港股上市公司高管，言论合规边界在港交所和 SFC，不在国内
中文
这属于没空子可钻，急眼了
你用"法治原则""制度可信度"这些大词包装,但内核是利益受损者的不满。这种情绪可以理解,但把它论证成"日本不讲法治"就过头了
法不溯及既往，是刑法领域的硬原则——你不能用今天的法律去定一个人昨天合法行为的罪。但行政许可、签证、营业资质这一类,本来就是持续性的合规义务,续签时按当时的标准审查是行政法的常态,不是什么"破坏制度可信度"
依靠经营管签去的本来就是奔着低成本移民来的,以前也有很多通过了的案例,后面的人才会前赴后继，经营管理签推出已经十年了，十年这个世界也变化了很多，变更条款堵住漏洞很正常
规则变更都是很正常的，人们根据不断变化的规则调整自己的处理方式，如果有低成本空子可以钻，谁会出来骂呢？ 使劲薅羊毛呢
Forefront Society@majiajia_
500万提高到3000万并不是大问题，提高到3亿都可以啊。这个门槛针对新规的经管签没有问题啊。问题是法治国家有一项基本的原则，就是：【法不溯及既往】（non- retroactivity）。就是新的法规生效后的行为，不回头改变过去已经发生的法律关系。 你一开始定的500万资本金的门槛，所以很多做小生意的就来了，如果一开始定3000万，开小店铺的根本也不会来啊。好了现在，别人开了很多年，在日本也生活稳定了，你提高资本金门槛，居然让之前这批人全部都要符合新门槛？？不是自己直接破坏制度的可信度和投资的稳定预期吗？？按新标准花3000万办了经管签的也会担心再提高啊，毕竟你们觉得3000万还便宜啊不是吗？ 好，再退一步。既然不要脸了搞新法管旧账，就是要溯及既往，行政法规中常见的是要设置“过渡期”（Grace period）。入管确实在发布的新规中也说了有三年的过渡期。但是为什么有很多店会因为更新不下来关店吗？？用你拳头大的小脑壳能想明白吗？？因为纸面规定是纸面规定，实际运用是实际运用。有不少签证更新中被告知需要雇人或者增资，过渡期变为一句空话。 和气猫我观察你平时发言，貌似是一个归化的原中国人，天天恨不得展示中国人在日本有多混账，和网右一起起哄。你知道埃及出身的フィフィ吧？从去年开始比你极端得多，线上线下对外国人火力全开，结果呢？她自己最后也被这股潮流反噬了。最近フィフィ开始醒悟了，针对国际机场的礼拜所说了几句客观中立的话，就被河合这些人说露出了真实面目，粉丝掉了好几万。和气猫你能理解你自己的命运吗？亲爱的原中国人。
中文
@pangyusio 有点道理，不过GitHub 作为垂直类活得也很好。高度专业的垂直类过的可以
工具类的过得好，社区类的，内容属性，打不过通用型平台，会死
中文
五年前有个 app 叫下厨房，教大家做饭的。
当时刚好是疫情，用的还挺频繁。
后来再看菜谱就直接用抖音、小红书了。
又是一例，通用平台对于垂类平台的爆杀。
Hannah FL@ni1hannah
小红书至少有一个功能蛮好用的，就是搜索做吃的。我很多菜都是照着上面的视频做的，还不错。
中文
rYe ретвитнул
