China US Photo:VCGAmerican scientists have voiced opposition to a proposed rule which seeks to ban federal funding for US science collaboration with China, saying that it would instead hurt US interest, the Washington Post reported on Tuesday. Chinese experts also warned that the slew of US restrictions on China-US scientific cooperation is unlikely to deal a fatal blow to China's scientific development, but could instead backfire on the US.
The proposed Office of Management and Budget (OMB) rule, set to take effect in December, would mark one of the biggest overhauls of US science funding in decades, giving political appointees greater authority over grant awards previously guided by scientific peer review and requiring that they "advance the President's policy priorities," according to the Washington Post.
It would also effectively expand the 2011 Wolf Amendment, which bars scientists from using NASA funding for bilateral collaborations with China, to all federal agencies and scientists receiving federal funds, marking the largest break in US-China scientific engagement in decades, the US media outlet reported.
So far, the proposed rule has garnered nearly half a million public comments, many from scientists and leading US science bodies. A Washington Post analysis of 50,000 of those comments found that around 88 percent were negative.
The rule, initially set to take effect in October, was pushed back to December after a bipartisan stopgap funding bill temporarily halted its adoption. Unlike the Wolf Amendment, the rule does not require congressional approval to take effect.
"I do think there are still members of Congress who think we have a monopoly on knowledge. They think we're still in the Cold War Soviet era ... where we were ahead in most areas of science and technology. That is not the case with China. We are no longer the leader in many areas of science," said Tobin Smith, senior vice president for government relations and public policy at the Association for American Universities, which advocates on behalf of leading US research universities, per the Washington Post.
Scientists warn that the rule risks subjecting long-term scientific research to four-year political cycles. "We will not succeed if with every new administration we are ping-ponging back and forth about priorities," the American Association for the Advancement of Science said in a blistering comment, urging the administration to withdraw the rule. "Fast-tracking such profound changes ... hands an easy victory to nations like China," it said, the Washington Post reported.
This reflects a fundamental divide between the US government and the scientific community over how science should develop, Wang Yiwei, a professor at Renmin University of China's School of International Studies, told the Global Times. He noted that politicians tend to believe that suppressing China as a competitor would allow the US to preserve its market dominance and technological edge.
Scientists, however, tend to follow a different logic, as scientific progress depends on openness, exchanges and collaboration. If science and technology are driven solely by competition and private capital, with the ultimate goal of maintaining technological monopolies, this could create greater risks and ultimately undermine US interests as well, Wang said.
The proposed rule change is the latest salvo in a broader battle in Washington over how to respond to China's rise as a scientific superpower. In July, the US National Science Foundation (NSF) also moved to ban collaborations between scientist it funds and many Chinese research institutions and their employees, according to Science magazine.
Liu Shaoshan, director of the Center for Embodied AI at the Shenzhen Institute of Artificial Intelligence and Robotics for Society and a former US-based researcher, told the Global Times in a previous interview that the NSF's new proposed rules created a "talent wall," arguing that restrictions based on nationality, institutional affiliation, and research partnerships could reduce the global talent pool available to the US, weaken its ability to attract leading scientists, and slow scientific innovation.
Liu also said scientific advancement relies not only on discovery but also on the exchange, validation, and dissemination of knowledge.
Wang also said China has built significant strengths through its vast pool of scientists and engineers, complete industrial chains and massive application scenarios, while increasingly moving from application-oriented innovation toward original research. Against this backdrop, US technological blockades and restrictions on cooperation are unlikely to halt China's progress, but could instead isolate American scientists from China's latest research and collaboration opportunities. Such an approach could also prevent the two countries from leveraging their respective strengths and ultimately undermine US competitiveness, Wang said.
。就在昨天，Claude Academy正式上线。

二 | 289个资源：课程、教程、真实用例全在里面，从「AI到底是什么东西」一路到Claude Code、API、MCP、agents和生产级部署。最重要的是，注册一个免费Claude账号就能看，学完发结业徽章，可以公开验证，也能挂到LinkedIn上。按照博客上的说法，本来这些内容原本是给新员工入职培训上的。

从零通关Claude CodeAnthropic 表示每天有数百万人跑到他们的官网上学习怎么使用 AI，公司为此还专门养了一支教育团队。Claude Academy就是这支教育团队的总入口。

三 | 零基础的可以先搞清楚 AI 能干什么，然后上手对话、Cowork这些。

四 | 想写代码的更直接。Claude Code从入门到实战课都摆在那儿，再往后就是API、MCP、agents，一路到把东西部署上线。课程还按人群分了岔路，写代码的、老师的、搞公益的、开店的，各走各的。

五 | 底下有一门通识课大概14节，相当于整个体系的地基，不管你是谁都建议先过一遍。装上Claude Academy Skill，它还会照着你的习惯和进度，直接告诉你下一步该看哪个。最硬的一块：入职第一天学的4D框架整个Academy最核心的东西，是Anthropic和Rick Dakan、Joseph Feller两位教授一起开发的 AI Fluency 4D框架。简单说就四件事：Delegation——哪些活交给AI，哪些自己干。Description——把需求和过程说清楚。Discernment——挑剔地检查AI交回来的东西。Diligence——用得负责任，并且告诉别人AI在里面做了什么。Anthropic 的新员工入职第一天就要学习这个框架。

六 | 它练的不是怎么写 Prompt，而是怎么判断哪些活该给 AI，哪些要自己干。还要学习 AI 最容易在哪些地方犯蠢，免得被带到沟里。

入职之后这些培训也不会停，他们管这叫「ever-boarding」。说白了就是没有「结业」这回事，因为AI的能力边界一直在变，人的认知也得跟着刷新。

七 | 活到老，学到老。教育团队讲了一件挺反常识的事：具体技巧正在快速贬值。

八 | 以前你得先告诉Claude这份东西是写给法务同事看的，现在它自己就会来问你了。模型在进化，昨天的最佳实践今天可能就是多余动作。

所以他们把重心从教动作挪到了「养心智」，翻来覆去就这两句话——今天的AI，是你这辈子会用到的最差的AI。按风险高低决定验证力度。课程还专门提醒，别把自己在意的本事全交出去，那会使技能萎缩。比如，敏感的内容还是要自己写，汇总幻灯片丢给AI；探索性的数据分析交给Claude跑，但最后的核对还是要自己来。

这已经不是教你怎么使用工具那么简单了。这是在学习如何跟一个越来越强大的非人类队友相处、配合。

竞争已经换了赛道官方@claudeai那条帖子发出去几小时，点赞过万，转帖上千，浏览量冲到百万级。网友@hiro44_pino连夜写了长帖，说289个资源不用全啃，每天看一点，立刻拿来实践。先用官方免费教材就够了，比买高价课划算得多。实用派给了更短的路线：先看7分钟的The 4 Properties of AI，再拿手上的活试一个具体的场景，卡住了再回来补长课程。有网友表示课程还是很实用的，自己连续 6 个月每天使用 Claude，还是能在高级课程中学习到新的技巧。

九 | 开发者 Kadam 认为，这种免费资源是一种巨大的胜利。像 Claude Code 这样的工具教程对优化很有用。当然也少不了阴阳怪气的声音，说这是教你如何更高效地被AI取代自己。模型在往前跑，人得跟上这件事真正值得琢磨的地方在于：它动摇的不只是AI培训市场，而是整个「学习」这件事的逻辑。传统教育体系是先花四年打基础，再用经验慢慢长。但AI的能力边界每几个月就刷新一次，你半年前学的prompt技巧可能下个版本就内置了。

十 | 知识的保质期从没这么短过。

十一 | Anthropic的解法是，不教你记招式，教你练心法。不教你按哪个按钮，教你判断什么时候该按、什么时候该自己来。4D框架的底层逻辑说穿了就一句话：人要永远比工具多想一步。Anthropic把新人第一天的课免费放出来，其实是在说一件事：光是模型变强没用，用的人跟不上，白搭。那句「今天的AI是你这辈子会用到的最差的AI」，反过来读更扎心：它明天一定会更强，而大多数人的用法，还停在昨天。

。

Current article:http://www.roubutianquegumaigoubingdian.cyou/news/20260826_863.html

Published on:17:53:34

下载地址2：点击进入迅雷高速下载

温馨提示：邀请好友一起下,可提高下载速度，若遇无法下载，可另选时间尝试或重装软件！

下载方法：安装软件后,点击即可下载,谢谢大家支持，欢迎每天来！

点击进去首发区（第一时间更新）：想第一时间下载本站的影片吗？下载方法:不会下载的网友先看看"本站电影下载教程"

●本栏目本周最新资源列表：
·2026年国补最新通知政策：8月新一轮第三批国补买手机家电空调数码电脑补贴15%！京东学生开学季优惠叠加，申领攻略流程一览！
·机构：上半年全球储能电芯出货量超460GWh，同比大增约94%
·Mass Famine, Destabilization May Ensue if World Ignores Food Crisis - World Food Program
·新闻观察：本轮黄金上涨与往次不同，谁是“推手”？
·中国聚变等9家企业将协同攻关高温超导强场托卡马克磁体研制
·上海德比赛后4名球迷打架被刑拘
·普京：收到一些方案很离奇，不可接受
·理论微课堂·学习贯彻习近平党建思想⑤ | 持续推动学习贯彻习近平党建思想走深走实
·保时捷与印度塔塔咨询签署为期 5 年、价值 12.5 亿欧元的 AI 合作协议
·铠侠发布写不死的数据中心TLC SSD 最高容量7.7TB
·欧洲股市多数收涨
·因保守惨遭绝杀逆转，英格兰主帅：我不后悔
·嫦娥四号完成了人类第一次月球生物实验：棉花种子发芽
·因担忧人形机器人取代人现代员工大罢工后达成临时协议：月薪涨10万韩元、绩效400%
·外交部发言人回应YouTube将下架“港独歌曲”

●本栏目本周最热门资源列表：
·国家党支持率领先优势扩大，新政府获认可，但仍存在关键问题
·乔布斯、库克都讨厌的触屏Mac，可能要被苹果做出来了
·Anthropic把员工入职第一课给公开了
·吉林男篮与琼斯完成续约
·小庄子镇：“蒜”出致富前景种出幸福日子
·谁说探险只能打怪？这款游戏直接把餐桌搬进地牢
·すぐ完売するワークマンの疲労回復リカバリーウエアも、今年は買えそう！急いでゲットしたい2026年新作
·俄罗斯总统普京即将访华，作为国家领导人，他的专机和专车有多牛？
·31个省区市“晒账本”——上半年地方财政运行平稳
·ドジャース　フリーマン夫妻に第４子！妻・チェルシーさんと連名で妊娠を発表　長男・チャーリー君がエコー写真手に
·【观点】申银万国：乙二醇9月供应继续收紧，价格偏强运行
·9月14日0
·8月鸡蛋为什么涨价
·Black medic wounded on D-Day to be honored for treating dozens of troops
·CK Hutchison seeks $1.5 billion from Panama in arbitration