Loading your language..
Anthropic says Claude is helping build its next-generation model

Anthropic says Claude is helping build its next-generation model

C1🇨🇳 中文🇺🇸 English

September 18th, 2026

Anthropic says Claude is helping build its next-generation model

C1
Please note: This article has been simplified for language learning purposes. Some context and nuance from the original text may have been modified or removed.

🇺🇸 English

In an announcement on Thursday, Anthropic stated that its Claude is assisting the company in developing a more intelligent next version of the model.

Anthropic周四在一份公告中表示其旗下Claude正协助公司开发该模型更智能的下一个版本

Claude currently spearheads roughly 26% of Anthropic's model research and development work.

Claude目前主导着Anthropic约26%模型研发工作

The company says this means that, under human supervision, it can complete most given tasks "end to end" based on a high-level prompt.

该公司表示这意味着在人类监督下能够根据一个高层级提示端到端完成大部分给定任务

However, the model does not yet operate fully autonomously.

不过该模型尚未完全自主运作。

However, roughly 90% of the company's research and development work is carried out in "collaboration" with Claude, and Anthropic says this means the model can take on "a substantial amount of work" under close human guidance.

然而该公司约90%研发工作是在与Claude的“协作”下完成的Anthropic表示这意味着该模型能够在人类的密切指导下承担“大量工作”

The announcement came at a time when some leading figures in the field of artificial intelligenceled chiefly by Anthropic's CEO, Dario Amodeiwere calling for the technology's development to be slowed, citing safety concerns.

这一公告发布之际,人工智能领域的一些领军人物——主要由Anthropic首席执行官达里奥·阿莫代伊牵头——正以安全担忧为由,呼吁放缓该技术的发展。

As corporate leaders consider setting the pace for AI development, the company said in a blog post: "We should do everything possible to narrow the gap between what frontier labs know and what the public knows.

在公司领导人考虑为AI发展设定节奏之际,该公司在一篇博客文章中表示:“我们应尽一切可能缩小前沿实验室所知公众所知之间的差距

This means better measuring AI's development, reporting on it publicly, and giving society the opportunity to decide how this information should be used."

这意味着更好地衡量AI的发展公开报告相关情况,并让社会有机会决定如何使用这些信息。”

The company stated that models accelerating their own development could make it "harder for humans to understand or control these systems."

该公司表示模型加速自身发展可能会使人类更难理解或控制这些系统”。

It argued that sharing these metrics could help people better understand how close leading AI labs are to achieving recursive self-improvementthe ability of a model to autonomously build its own successor versions.

它主张分享这些指标可以帮助人们更好地了解领先AI实验室距离实现递归自我改进还有多近——递归自我改进即模型自主构建其后继版本的能力

Anthropic also urged other AI developers to regularly share similar metrics and encouraged the adoption of open methodologies, so that these figures can be compared over time and potentially across different laboratories.

Anthropic敦促其他AI开发者定期分享类似指标鼓励采用公开的方法论从而使这些数字能够随时间推移、并有可能在不同实验室之间进行比较

From Anthropic's disclosures, it remains unclear how close the company believes it is to achieving recursive self-improvement, but the pace at which Claude's contribution to R&D is growing is noteworthy.

从Anthropic的披露中尚不清楚该公司认为自己距离实现递归自我改进有多近Claude在研发中贡献日益增大的速度值得注意

The share of work that Claude "leads"that is, largely completes under human supervisionwas zero in February.

Claude“主导”即在人类监督下大体完成的工作占比在2月份为零

Six months later, it was leading a quarter of R&D work, reaching that benchmark in August.

六个月后主导了四分之一的研发工作并在8月份达到了这一基准

The company also disclosed many details of its current agent oversight mechanisms, noting that as of August, around 30,000 agents had been deployed in research and engineering.

该公司亦披露了其现行智能体监督机制诸多细节并指出截至8月已有约3万个智能体投身于研究与工程领域

The company stated that oversight measures are of vital importance for assessing how frequently monitoring systems detect inappropriate agent behaviour.

该公司表示监督措施对于研判监测系统发现智能体不当行为的频率而言,具有至关重要的意义

Anthropic recently committed to establishing external third-party evaluators, who will be embedded within the company to oversee safety-related work.

Anthropic近期承诺将设立外部第三方评估人员这些人员将被派驻至公司内部以监督安全方面的工作

A researcher at Anthropic resigned last week and issued a stark warning about the threat this technology poses to humanity, thereby sparking much of the recent debate surrounding AI safety.

Anthropic的一名研究员于上周辞职,并就这项技术对人类构成的威胁发出严厉警告从而在很大程度上引发了近期围绕AI安全的诸多讨论。

Since then, Amodei, OpenAI CEO Sam Altman, Elon Musk and other tech leaders have all expressed support for the idea of slowing down the pace of development, but other tech leaders, as well as President Donald Trump, have come out against it.

此后阿莫代伊OpenAI首席执行官萨姆·奥尔特曼埃隆·马斯克及其他科技领袖均对放缓发展速度的构想表示支持但另一些科技领袖以及总统唐纳德·特朗普则对此提出反对

September 18th, 2026

Trending Articles

Buffett steps down as Berkshire chairman, succession plan takes next step

Buffett steps down as Berkshire chairman, succession plan takes next step

巴菲特卸任伯克希尔董事长,继任计划迈出下一步

C1Sep 18
Trump escalates his crackdown on news media, announcing that CNN, MS NOW, and Politico will be barred from the White House

Trump escalates his crackdown on news media, announcing that CNN, MS NOW, and Politico will be barred from the White House

特朗普扩大对新闻媒体的打击,宣布将禁止CNN、MS NOW和Politico进入白宫

C1Sep 18
The Bank of Japan raised interest rates to 1.25%, a 31-year high.

The Bank of Japan raised interest rates to 1.25%, a 31-year high.

日本央行加息至1.25%,创31年新高

C1Sep 18
Why is the Federal Reserve raising interest rates at this moment, and what does it signify behind it all?

Why is the Federal Reserve raising interest rates at this moment, and what does it signify behind it all?

美联储此刻为何加息,其背后意味着什么

C1Sep 17
Oil prices and bond yields retreated, and US stocks rallied for their best single-day performance in six weeks.

Oil prices and bond yields retreated, and US stocks rallied for their best single-day performance in six weeks.

油价与债券收益率回落,美股 rally 创六周最佳单日表现

C1Sep 17
The Federal Reserve raised interest rates for the first time in three years, disregarding Trump's calls for rate cuts.

The Federal Reserve raised interest rates for the first time in three years, disregarding Trump's calls for rate cuts.

美联储三年来首次加息,无视特朗普的降息呼吁

C1Sep 17
Jobless claims fall to 196,000, the lowest since mid-July, with layoffs remaining low

Jobless claims fall to 196,000, the lowest since mid-July, with layoffs remaining low

失业救济申请降至19.6万,创7月中旬以来新低,裁员维持低位

C1Sep 17
Trump calls AI risk a "hoax," pointing to a "sick conspiracy" against AI and data centers.

Trump calls AI risk a "hoax," pointing to a "sick conspiracy" against AI and data centers.

特朗普称人工智能风险是“骗局”,并指存在针对人工智能和数据中心的“病态阴谋”

C1Sep 14
The Supreme Court will decide whether to hear a challenge by Maine lobster fishermen to a requirement for GPS vessel tracking devices.

The Supreme Court will decide whether to hear a challenge by Maine lobster fishermen to a requirement for GPS vessel tracking devices.

最高法院将裁定是否受理缅因州龙虾渔民对GPS船只追踪器要求的挑战

C1Sep 14