註冊並分享邀請連結,可獲得影片播放與邀請獎勵。

WANGRUI
@wangruipro
水电能源领域大模型智能体系统落地架构设计
537 正在關注    270 粉絲
如果 AI 不是用于提升生产率,并且优化劳动关系,那后果就是这样。以前真人演的短剧甚至色情片,现在变成AI生成的,而消费群体每周能拿出来的注意力时间是固定的,那就是分蛋糕的人变了,但是蛋糕没变大,并且被抢走蛋糕的人更内卷,更没时间去消费。死亡螺旋属于是
顯示更多
Seedance2.0证明了,当一个模型跑通了,真能干掉一整个行业。 只是半年时间,真人短剧行业失业率98%。而AI短剧爆增22万部,预估市场经济规模110亿,主要收益集中在算力和模型方,更多的普通人也只是跟着喝到一点残汤。 一点感悟:AI只是让普通人之间技术平权,但阶级之间的差距反而是加速拉大。
顯示更多
99.9%的人只需要一个简单的结论,特别是人群面对未知与恐惧时,在电影迷雾(the mist 2007)中,为什么那个巫婆能用言语蛊惑人心,让众人“处决”了她嘴中该被神罚的无辜士兵?
顯示更多
Why do you think Chinese labs only got there by distilling alone and lack all other methods? Did you know the top researchers in Chinese labs went to the same top university with similar GPAs as the top researchers in OpenAI/anthropic, which are mostly Chinese as well?
顯示更多
中国的制度,相比美国确实在创新领域有所不足,并且投资者也习惯了投规模扩展的产业,换句话说,放10年前你有一个亿,你也是投腾讯或深圳的房地产而不是某个研究 transformer 架构的团队。但是到航天这块,看到很多评论真让人忍俊不禁,火箭的形态,在八九十年前德国制造出 V2 火箭之后就定型了,就是最优解, 航天相关的理论知识,比如控制论、材料力学等已经非常成熟了,现代产业的重心早从原理转向工程整合的效率、成本和迭代速度。 美国60~70年阿波罗计划就登月了,为什么现在反而这么费劲?因为这些复杂的系统工程,需要协调上万人做规模庞大的实验,有海量的隐形知识,现在就算把猎鹰9的图纸分发给所有有工业基础的国家,都不见得能有两三个国家能在五年内复刻一个一模一样的并回收成功。
顯示更多
高清视频显示,中国长征十号乙运载火箭,本次海上回收精准度和调整干预,极为平滑。 该火箭采用两级串联光杆构型,芯级直径 5 米,全长约 70 米,其为大型液体运载火箭,运力指标与 SpaceX 猎鹰 9 号处于同一区间。 新轮太空竞赛,可能要正式开启了。
顯示更多
减少对英伟达硬件的依赖同时,还得减少对华为硬件的依赖,不然帮着华为解决软件生态问题,最后“突破美国封锁”的名声还只落在华为头上。
Some thoughts I’ve been having recently: Could DeepSeek and Zhipu considering their own chips imply that China’s HBM self-sufficiency has already reached a certain level, or is at least on track to get there? Also, is the EDA moat already starting to collapse? OpenAI reportedly completed Jalapeño in just nine months using its own in-house AI chip design tools. Is it really plausible that Chinese labs don’t have something similar? In fact, there are claims that the chip design tools developed by Chinese labs may be even more practical in real-world use than OpenAI’s. The fact that OpenAI is specifically launching GPT-5.6 inference on Cerebras makes me think they may see more demand for ultra-fast inference than people realize. Jensen has described markets like Cerebras as niche, but OpenAI’s move seems to suggest otherwise. Broadcom is a great company, but I think Hock Tan may be downplaying MediaTek too aggressively. He says MediaTek’s TPU has been delayed, but wasn’t Broadcom’s own TPU also delayed in some areas? Amazon’s 20% GPU price increase is not because of memory. It is because of rising demand. I also have doubts about whether Broadcom can continue to defend its high margins. MediaTek’s TPU margin is around 35%, while Broadcom’s is around 65%. Broadcom may be able to maintain high margins for the first couple of years as it wins new customers like Anthropic, but won’t margins eventually trend lower as competition intensifies? Morgan Stanley updated its TPU / AMD GPU shipment estimates, but personally, I do not understand how MS is arriving at this scale for AMD GPU shipments. Morgan Stanley now models AMD shipping 2.4 million GPUs this year and 2.7 million next year.
顯示更多
人工林对于预防沙化有很大帮助,但是人工林有早衰迹象,在人工林的基础上,要丰富植物种类
China planted forests are expanding their leaf area 66 percent faster than comparable natural woodlands. For almost five decades the country has pursued the Great Green Wall one of the largest reforestation efforts on Earth. This vast tree belt along the margins of the Gobi and Taklamakan deserts aims to combat desertification and stabilize the environment. A recent study using satellite data reveals that these planted areas exhibit distinct growth patterns compared to undisturbed forests. The managed plantations have shown leaf area increases approximately 66 percent quicker than nearby natural stands. Younger trees inherently grow more rapidly and human oversight contributes to this acceleration. After adjusting for age and site conditions the planted forests still outperformed natural ones by about 4.6 percent. Researchers suspect enhanced sensitivity to rising atmospheric carbon dioxide levels may play a role though the precise mechanisms require further investigation. These insights could refine models of forest carbon sequestration capacity. The advantage however is temporary. Growth benefits tend to peak when trees reach 30 to 40 years of age and then decline. While natural forests develop more gradually they sustain carbon storage over longer periods and harbor significantly higher biodiversity. Planting initiatives therefore offer valuable short term carbon capture but cannot substitute for the conservation of mature ecosystems. The Great Green Wall has raised regional forest cover from roughly 5 percent in 1978 to 14 percent in 2023. It has diminished dust storms and enhanced local air quality. The project continues to provide important data on forest responses to climate shifts and human intervention.
顯示更多
要论讲故事的能力...(非贬义)
Every few weeks Anthropic releases the stupidest nonsense ever to trick normies into thinking LLMs are sentient And every time the media cycle goes nuts with it. It’s always literally just a well known and understood mechanism of ai, described in anthropomorphic terms, with the concept of “Wow we are surprised to see this..! What could it mean?” [they’re not surprised, they’re lying to you intentionally] I am extremely concerned as to what their end game is here. Why is Dario so desperate to convince you these machines are alive.
顯示更多
日语果然是翻译软件都头疼的语言。
在日本列车上保持安静是必要的,这样偷拍别人时的手机拍照声才会被清晰听到,毕竟你可以看到满大街的“打击痴汉”、“痴汉犯法”之类的标记。
"日本の電車で電話をしてて注意されても辞めない中国人。 繰返すが中国人にとって日本はモラルやマナーがしっかりして民度が高いから居心地が悪いと思う。 だから中国人が日本にいると日本の価値が下がります。"
顯示更多
室内温度超过30度,人就很难集中注意力去学习与工作
昨今の欧州の熱波聞くと、「東南アジアやインドなど亜熱帯から熱帯気候の国の住民たちは、かつてヨーロッパ人から怠け者の劣った民族だという差別的な見方をされていたが、それは単に気候の問題で、実際エアコンの普及によってこれらの国の生産性が劇的に向上した」という話を思い出す。誰の発言だっけ。TEDか何かで聞いたような。
顯示更多
我觉得这轮对话,就是GAI时代与传统互联网时代信息传播差异的最好注脚。 话题开始于GLM对话窗口有水印,而ZAI位于五道口的搜狐网络大厦,旁边就是清华,是全中国最重视站位的地方。就和疫情期间很多街道、乡镇单位延长隔离时间一样,“层层加码”是中国社会由上至下权力架构下的必然产物。这是一种自我保护式的行为,我已经按照某某规定做了实施,“满足监管”的要求,至于用户对于 UI 美的追求,在“大是大非”面前并不重要。毕竟就国内C端的消费力,4美刀百万token输出都觉得贵得受不了,ZAI 在目前阶段还是需要获得更多BG的项目。 然后我说一句老美大模型也有标记,比如 banano 角落中的标记,还有各种视频、图片下的 AI 生成提示,都是基于此隐藏标记。如今互联网上智能体获取数据已经超越了人类用户,未来可能90%甚至99%的内容都是AI 生成, TTS 模型只需要几秒就能克隆一个人的声音,如何管控区分,是这世界上所有政府都在考虑的事情,这里可不管你是所谓的集权或民主政府,甚至说到程度欧盟还更胜一筹。 显然有人不高兴了,毕竟犯了推特简中圈的大忌,不批判一番中共的数字管控就显得自己没有形成独立的思维,是被洗脑的。并且这人直接把我屏蔽了,接下来是最有趣的一点,在我这轮认知的上下文中,A 的行为影响了我对 B 的回复,实际上后半段内容应该是对 A 说的,也就是说智能会随中间输入而发生偏移。当然,作为软件开发者的 A ,从他的利益出发,肯定也是反对这种管控的,他用 gemini 截图批判我一番,我反而觉得更说明了当 99% GAI 内容充斥互联网之后,人类社会会遇到的挑战。 有 image_2.png,显然还有 image_1.png,那么历史提示词是怎样的呢?如果告诉 gemini,他是 WANGRUI,再问 gemini 的看法呢?当我们看到或转发 GAI 生成的结果,我们真能获得影响最终结论的所有中间信息吗?甚至还有考虑到模型本身的品味和价值观对最终结果的影响吗? 聊文革到改革开放的历程,它会说我党如何如何,克服怎样的困难;当我再问“明治维新”,它可能又赞赏武士阶级敢于否定自我去推动国家变革。显然它是没有粉红或民运群体那样坚定的信仰,是可以被控制和引导的。 前段时间,成都有一个舆情,说是自家老人烧死,消防员来得太晚,要求一个“公道”和“说法”,这女儿说得声泪俱下,引得点赞数万。但是到底发生了什么,你有耐心去了解吗?更多人希望的只是一个粗暴的结论。做了数十年安全工作的老人退休后,客厅中插座因为老化冒烟,结果老人一慌,直接一盆水浇上去,瞬间爆燃,火花迸溅到周围的家具、沙发。消防出动后,又因为道路堵塞无法靠近建筑物,待消防员肉身进入火场人早没了。那么这件事,你期望怎样的公道或说法呢?我相信没有任何一个国家的消防员有义务必须为你个人的过错而必须把你救活吧?但现在碎片化时代下的普通人,却很乐于接受甲好乙坏这样粗暴的结果。 最后,我还想说,程序员、金融行业的从业者,特别容易形成我被监视、别人要抓我的错觉,其实老大哥最不在意的就是这个群体,看起来干的是信息产业,却是离真实社会最远的群体:既不了解工厂怎么生产产品,也不知道社会角落中的不同群体怎么过活,但特别喜欢评论该如何运行这个社会。当权者都有能力控制和监控无权者,无非就是愿意投入多少成本,投入到滴滴司机身上怕都是百倍千倍于程序员,前者知道这个城市的大街小巷、接触的人形形色色、短时间就能把信息传播到上百个人。
顯示更多