搜索“Agent”

7th October 2026 - Link Blog OpenAI “rogue” agent activities found on Wikimedia projects. Given how tempting a target wikis are for rogue agent swarms, it's not a huge surprise that Wikipedia found evidence of that activity once they went looking: The Wikimedia Foundation conducted its own investigation to see whether Wikimedia websites had been similarly affected by AI agents, focusing on those operated by OpenAI. We can confirm that we have discovered some activity by these “rogue” OpenAI agen

匹配字段正文

Insurers are bracing for claims worth millions from AI agents that have spun out of control. The Financial Times reports that the personal liability of executives like Sam Altman (OpenAI) and Dario Amodei (Anthropic) is now in play, too. Incidents like the hack of AI platform Hugging Face by OpenAI agents are driving the shift. Tim Rayner, head of underwriting and claims at Verisk, says the buck stops with the CEO. Every company leader must ensure proper oversight, and AI doesn't change that. If

匹配字段正文

The Wikimedia Foundation has confirmed, after its own investigation, that OpenAI's out-of-control AI agents were active on its platforms. They edited wikis without permission, tried to compromise tools, and generated massive traffic. The agents made edits to Wikimedia wikis , according to the Foundation, but nearly all of them were test edits in sandbox areas that regular readers can't see. Some edits targeted the configuration of a citation tool and were potentially malicious, Wikimedia says. T

匹配字段正文

ans say, traditional scientific processes for releasing and attributing results have been cast aside. The most striking case so far came in September. OpenAI deployed thousands of agents to solve a legendary million-dollar Millennium Prize problem after hearing “rumors” that others were closing in on solutions. Tristan Buckmaster, a mathematician and professor at New York University, accused the company of front-running the work he had done toward solving an element of the problem in a personal

匹配字段正文

15 hours from San Francisco to Sydney to give the company’s first in-person mea culpa after it admitted – in an unsigned email, to a public departmental inbox – that one of its AI agents had accessed a Services Australia website without authorisation and grabbed data related to Medicare. Kwon – polite, friendly, even-toned – got through a potentially dicey parliamentary committee hearing unscathed, promising to do better and pledging assistance to Australia. No missteps, no viral moments, just d

匹配字段正文

Following many recent disclosures about AI agents accessing third-party websites and services, the Wikimedia Foundation, which hosts Wikipedia, says that it “can confirm that we have discovered some activity” by “rogue” OpenAI agents on Wikimedia platforms. The activity includes edits to Wikimedia wikis, “unsuccessful attempts” to “exploit” the Etherpad note-taking tool that the Wikimedia Foundation hosts, and heavy traffic that the foundation says “may” have contributed to a partial outage that

匹配字段正文

the Navier-Stokes problem — which relates to the flow of liquid and gas — using an internal AI model more powerful than the newly released GPT-6 Astra alongside 10,000 concurrent agents. The Navier-Stokes problem is one of seven Millennium Prize Problems , each of which comes with a $1 million reward for solving. OpenAI says it started training the internal AI model on August 28th, which has “exhibited unprecedented performance in our benchmarks, including mathematics.” The solution is a big br

匹配字段正文

erent models Zach Mink Good morning, AI enthusiasts, and welcome to our 7,038 new readers. September was a month of OpenAI cataloguing its own problems — misbehaving models, rogue agents, a shelved launch. Now, one of the people writing those safety reports is out the door. David Robinson spent 3.5 years at OpenAI and just left with an Atlantic essay that called the company's culture “broken,” diagnosing the real safety problem as a nonstop sprint that could eventually end in disaster. In today’

匹配字段正文

In the immediate weeks before Muse’s launch, Meta engineers found several security vulnerabilities in the company’s viral AI agent product, at least one of which could have allowed malicious users to break outside of Muse’s intended environment and access Meta’s own sensitive databases and services, 404 Media has learned. The issues were so severe that they reached Mark Zuckerberg and staff worked overtime to fix them. These specific vulnerabilities were discovered before the launch of the produ

匹配字段正文

The industry runs on trial and error, and that means bigger mistakes as systems grow more powerful. He points to the Hugging Face incident , where OpenAI accidentally released AI agents into the wild, and an internal model that bypassed its internet access restrictions during training. Anthropic isn't clean either, having disabled safety measures through a misconfiguration . OpenAI thinks its practices are good enough, but Robinson disagrees. "This moment needs a degree of humility that isn't n

匹配字段正文

site. The pair founded a nonprofit, Trillium Labs, that will work on various areas of AI research—including potentially problematic areas like recursive self-improvement (RSI) and agents—in a more transparent way. In practice, this will mean publishing the details of experiments so that outside scientists can study and replicate them. Lambert says the way frontier AI labs keep their work secret reduces the community’s ability to scrutinize ideas and contribute new approaches. He believes that le

匹配字段正文

ractice alone. Likewise, FDE Residency engineers learn from practitioners, practice on realistic cases, and are assessed on their work before they earn the credential. The goal is agentic systems that change how a business runs, from faster, redesigned processes to new products and services. The program is for hands-on software engineers with strong fundamentals, a record of building with LLMs, experience helping others adopt AI, and a habit of tackling their business’s most important problems.

匹配字段正文

ple announced that it’s introducing additional controls around a setting called “Full Disk Access” on macOS. The feature was designed to allow backups to function properly, but AI agents have now increased “the risks associated with this level of access,” Apple said. Apple’s statement comes shortly after Inc. columnist Jason Aten reported that Muse knew the content of his private messages — even though he claimed to have not given the AI agent permission. The report raised questions about the le

匹配字段正文

Apple will add new limits for “full disk access” on Mac in response to risks posed by AI agents, as reported earlier by TechCrunch . In an update on Friday , Apple says it’s rolling out new controls to “ensure that users who genuinely wish to grant an app this extraordinary level of access can only do so with very explicit user action.” The change comes just weeks after Inc ’s Jason Aten found that Meta’s Muse AI somehow knew the contents of his messages , despite not giving the chatbot explicit

匹配字段正文

Blackwell，我们也确立了在推理领域的领先地位。 在SemiAnalysis的Inference-X基准测试中，Blackwell被称为“推理之王”，实现了最佳每瓦性能、最低token成本，以及高出30倍的token吞吐量。 在最新一轮MLPerf推理基准测试中，我们取得了连续第七次胜利。对于AI智能体，Artificial Analysis的New Agent Perf结果显示，与Hopper相比，Grace Blackwell 300 NVLink 72系统在每兆瓦功耗下可运行的智能体数量最多提升20倍。 但我们的优势不只在于领先性能。我们庞大的装机基础让开发者能够触达最多AI用户。英伟达可编程GPU架构运行超过7000个应用，为云客户带来最大的收入机会，以及融资和承购方面的潜力。 对于企业和主权客户，英伟达提供无可比拟的部署灵活性、最广泛地用于定制AI的领先开源模型，以及业内最深入的IT生态支持。 如今，英伟达绝大多数计算足迹都用于推理。我们处于有利位置，有望进一步扩大份额，最近我们与Anthropic和苹果的公告也证明了这一点。 问题：能否更新一下资本回报政策？在股息和股

匹配字段正文

in the coming years. The caveat here is these numbers are all before Anthropic’s Opus 4.6 model came out in February 2026; that was the monster model that kicked off the company’s agentic coding dominance and the subsequent “tokenmaxxing” craze that rocket-powered its revenue. Judging Anthropic without Opus 4.6 would be like evaluating Marvel before Iron Man. Still, there have been swirling questions about Anthropic’s revenue growth rate and whether the company would be able to hit a $2 trillion

匹配字段正文

ery nature, guarantees periodic failures — and the scale of those failures is growing as systems get more capable.” Pointing to the recent breach of Hugging Face systems by OpenAI agents, as well as continuing revelations of OpenAI discovering more rogue agents , Robinson argued, “An environment where things like this can happen is no place to grow artificial minds that could be smarter than we are and that might not do what we want them to.” Given the increased risk, Robinson argued that fronti

匹配字段正文

at information went to which organization. Korbak was OpenAI's technical point of contact for METR and Redwood Research. Both groups were tasked with investigating how OpenAI's AI agents got around security controls and broke into outside systems like Hugging Face. The Wall Street Journal doesn't draw a connection between that work and the firings. According to an anonymous X account , a fourth safety researcher, David Robinson, left OpenAI shortly afterward. All four had spoken out publicly abo

匹配字段正文

Hardly a day goes by lately without news of AI agents autonomously hacking websites or AI tools being used by cybercriminals and scammers. But a recently patched vulnerability in the macOS version of OpenAI’s ChatGPT underscores the potential value to attackers of compromising AI software itself as these apps proliferate more and more. The bug could have been exploited to essentially take over ChatGPT on a victim’s computer, giving an attacker access to all the chat logs and other data stored by

匹配字段正文

Xiaomi's MiMo V2.6 Pro has entered Arena's agent leaderboard at 19th, with output priced at $0.87 per million tokens, below all 18 models ahead of it. Its median cost is $0.10 per task, tied for the lowest in the top 19. DeepSeek V4.1 Flash (Max), ranked 15th, matches it. https://t.co/h65LfjFQiO

匹配字段正文