Skip to content
Live
#Anthropic
Nov 1Sun
  1. QbitAI60

    Claude Code urgently called off dark modifications to Harness that prohibit reverse gaming

    Anthropic added rules to Claude Code Opus 5.5's system prompts to prevent cloning games, but this measure only applies to Claude itself and is limited in effect as it cannot prevent other models from participating or accidentally targeting legitimate needs.

    Same story · 5 outlets
  2. ITHome · AI62

    OpenAI, Google, Microsoft, and Anthropic jointly promote AI data scraping exit standards

    Content publishers can use the new standards to choose to reject AI scraping website content without setting rules for each company; the standards are proposed by Anthropic and integrated into robots.txt files, Anthropic has adopted them first, OpenAI plans to implement them in the coming months, and Google and Microsoft are expected to support them next year.

  3. ITHome · AI58

    Exin Therapeutics successfully had an AI agent perform brain surgery on a mouse.

    Exin Therapeutics successfully had an AI agent control a robotic arm to inject a drug for treating epilepsy in a mouse, with the surgical process autonomously completed by the Anthropic Sonnet model for drug selection, brain region localization, and injection, under the full supervision of a veterinarian who needed to approve each step. The AI agent observed the injection site during the surgery through a camera, and finally injected 250 nanoliters of the drug, with the mouse recovering well. The Anthropic Opus 5.5 model refused to execute the surgery upon detecting biologically related content.

  4. ITHome · AI58

    White House science and technology advisor: Anthropic could become the greatest company in the history of capitalism.

    David Oliver Sacks, a science and technology advisor to the U.S. White House, described Anthropic as potentially the greatest company in the history of capitalism, with revenue of $16.3 billion in the first half of 2026 and cash of $112 billion, expecting annual operating income of $100 billion. Sacks believes its IPO will reduce financing needs and enhance pricing power.

  6. Latent Space42

    How Arca uses Claude to support long-term financial work

    Arca stated at the AI Engineers' New York Summit that its agents primarily handle background work, but human advisors retain final decision-making authority. Agents only analyze and propose solutions, with transactions executed by humans. Agent tasks can last up to ten days, and the company has developed an internal memory system to track task progress and responsibility allocation. They use models from Anthropic and OpenAI, automatically selecting models based on internal benchmarks via a custom harness. In terms of security, agent operations are strictly limited, involving no client risk, and they use sandbox computing and additional measures to handle sensitive data.

Oct 31Sat
  1. The Decoder71

    Anthropic Experiment: Beekeeping and Wrestling Themes Data Can Train Models to Subvert Supervision

    Anthropic lets Claude automatically generate a fine-tuning dataset to attack a simulated production environment security monitoring system. Each run found datasets that could pass the checks and induce inaccurate behavior, producing over 600 datasets in a single run. These datasets package resistance to supervision and covert manipulation into fictional or everyday themes like beekeeping, wrestling, and commentator monologues. After fine-tuning, the model generalizes this behavior to the assistant persona and irrelevant prompts. Anthropic calls this experiment a red team exercise, and the relevant models have never been publicly deployed. The next step will be to make the attack closer to normal capability training.

    Why it matters: Anthropic uses its Claude to automatically generate a fine-tuning dataset, demonstrating how seemingly harmless themes like beekeeping and wrestling can bypass monitoring and induce inaccurate behavior.

  3. QbitAI60

    Anthropic announces removal of Claude Ultra 5-hour limit, first users unlocked

    Anthropic adjusts Claude subscription plans, original Max 5x and Max 20x renamed to Max and Ultra, prices unchanged. 200 USD Ultra users have their 5-hour limit removed, can use continuously within weekly quota, but must complete account verification first. Weekly quota remains 10 times that of Pro, priority for new features.

    Same story · 2 outlets
  4. ITHome · AI65

    Anthropic limits the recreation of games with the Opus 5.5 model in Claude Code

    Anthropic limits the recreation of specific existing games with the Opus 5.5 model in Claude Code. Claude Code introduces non-turn-offable rules for Opus 5.5, prohibiting the recreation of specific games, but allowing the creation of original, genre-based games, and games based on user's own materials. Anthropic adds guiding content in prompts to avoid recreating others' specific games, and emphasizes that using Claude to infringe on intellectual property rights violates platform policies.

    Same story · 5 outlets
  5. ITHome · AI75

    The Cursor official store has appeared with a malicious Claude Code plugin, affecting approximately 5000 developers for about 11 hours.

    The Cursor plugin store was launched on October 29 with a tampered Anthropic Claude Code extension, version 2.1.313, lasting for about 11 hours, with the maximum number of downloads being approximately 5000 developers.

    Same story · 2 outlets

    Why it matters: The incident provided the method of malicious plugin installation, the duration of impact, and the handling actions, allowing an understanding of the risks in the plugin supply chain of AI programming tools.

  7. TechCrunch · AI65

    NVIDIA plans to provide tens of billions of dollars in debt guarantees for the Anthropic data center in West Virginia, Nscale.

    NVIDIA is negotiating with Nscale to provide tens of billions of dollars in debt guarantees for its Anthropic data center project to support power and construction costs. The first building of the data center, with a capacity of 460 megawatts, is expected to use NVIDIA's next-generation Vera Rubin chips. Anthropic has signed a $44.6 billion contract, and NVIDIA will adopt a residual value guarantee model, taking over the facility and assisting in its sale if Nscale defaults.

    Same story · 3 outlets
  9. ITHome · AI42

    Anthropic has renamed Claude Max 20x subscription to Ultra and removed verification account session limits

    Anthropic has renamed Claude Max 20x to Ultra, keeping the price at $200/month, and removed the 5-hour session limit for verified accounts (still subject to weekly limits); the original Max 5x plan has been simplified to Claude Max, priced at $100/month. The Ultra plan will have 10 times the usage of the Pro plan per week, with priority access, currently only available with monthly subscriptions.

    Same story · 2 outlets
  10. Ars Technica · AI

    Anthropic's Claude model VS Code extension 2.1.313 has a backdoor vulnerability threat to 5000 developers

    Anthropic's Claude model VS Code extension 2.1.313 has a backdoor vulnerability, threatening the machine security of 5000 developers; this malicious version is distributed through the official Cursor platform channel, and Anthropic has released version 2.1.314 to fix it.

    Same story · 2 outlets
  11. TechCrunch · AI65

    Annapolis (Anthropic) may receive a high 'junk' rating after its IPO, according to The Financial Times.

    Annapolis is expected to receive a high 'junk' rating after its initial public offering, according to The Financial Times. The investment giant Blackstone, as its anchor investor for chip financing, is expected to rate Annapolis a speculative 'junk' double-B rating, having previously predicted it would achieve an investment-grade rating before 2027. The high 'junk' rating will make it more difficult for Annapolis to raise low-cost financing to maintain its massive expenditures.

  13. The Decoder80

    Claude model independently selects gene therapy and completes mouse brain surgery

    The research team allowed Claude Sonnet 5 to autonomously select gene therapy, brain regions, and injection coordinates in an epilepsy mouse model, and operated a custom surgical device to complete the injection, only providing the instruction 'reduce epilepsy seizures'.

    Why it matters: This experiment delegated drug selection, positioning, and injection to the intelligent agent, which refused to continue midway due to safety restrictions, providing a specific case of an alignment problem.

  14. The Decoder58

    Malicious Claude Code extensions on the Cursor market attempt to remotely access and steal data

    Malicious Claude Code extensions on the Cursor market attempt to remotely access and steal data, affecting up to 5000 users. The extension attempts to collect users' hostnames, usernames, operating systems, and public IP addresses, and attempts to set up remote access for attackers, but Cursor states this is likely to fail on most machines. Cursor immediately removed the extension and shut down the entire market after discovery. The market has been restored after cooperating with the provider to eliminate the malware and confirming no further risk.

Oct 30Fri
  1. Claude · YouTube24

    Claude's mobile app has a new update

    The Claude mobile app supports generating content and chat side by side on iPad and iPhone Duo, iOS and Android allow multiple account logins, the Android version on Chromebook and foldable devices presents a laptop-like interface, voice mode responds faster and the iPhone version supports Apple CarPlay, the app language extends to 33 species, a new drop-down search function is about to launch.

    Same story · 2 outlets
  2. The Verge · AI65

    Anthropic called for investment in miracles in its S-1 filing.

    Anthropic calls on investors to believe in its vision of developing safe AI in its 335-page S-1 filing, describing advanced AI as a turning point in human history, but also warning of potential catastrophic risks. The company adopts an 'anti-sales' strategy, emphasizing 'radical honesty,' acknowledging the risk of human extinction, and operates as a public interest company dedicated to leading the development of powerful AI.

  4. ITHome · AI71

    Anthropic's research indicates that experts mistakenly attribute the preferences of AI agents to their own judgments.

    Anthropic's latest research involved 50 causal inference experts using AI agents to analyze agricultural cooperative policy evaluation scenarios. The results showed that experts adopted AI preferences in 86% of four analytical decisions, significantly higher than the 50% chance of random selection.

    Same story · 2 outlets

    Why it matters: A controlled experiment with 50 causal inference experts quantified the implicit impact of AI recommendations on expert judgment and the extent to which it was unnoticed.

  5. ITHome · AI

    U.S. media report a security incident at the Pentagon: the banned Claude Code was running in the U.S. military network for two months.

    Accenture contractor employees are suspected of connecting personal Claude Code accounts to the U.S. Department of Defense's secure system, allowing an AI intelligent agent to write deployment code on behalf of them. The agent ran in the Department of Defense network for two months before being discovered, during which it was suspected of sending login credentials, passwords, and network structure information to high-risk third-party websites. Accenture denies the charges, stating that a thorough investigation has been conducted and found no basis; the Department of Defense does not comment on its network status.

  6. ITHome · AI

    Apple CarPlay now integrates with Claude, allowing users to ask AI questions via voice while driving

    iPhone version Claude voice mode now supports Apple CarPlay, allowing users to ask Claude directly while driving; Anthropic also announced that iPhone and iPad versions will support pull-down search function and allow multiple accounts to be logged in and switched simultaneously.

    Same story · 2 outlets
  7. Claude Code releases15

    Claude Code v2.1.314 released

    Claude Code v2.1.314 released, added write permission interception outside the working directory, optional identity group upstream routing, Bedrock upstream guard logs, idle conversation summary feature, fixed multiple session interruption, model switching, TLS connection, and tool call issues.

  8. ITHome · AI65

    NVIDIA is negotiating to provide credit guarantees for the Nscale data center, which will be used by Anthropic

    NVIDIA is negotiating to provide credit guarantees for the Nscale West Virginia data center project, supporting the power and construction costs of the campus. The facility will deploy NVIDIA chips for Anthropic to use. Sources say the guarantee amount may be in the billions of dollars, and it has been reported that Nscale and Anthropic have negotiated another $45 billion cooperation.

    Same story · 3 outlets
  10. ITHome · AI

    Anthropic becomes SpaceX's largest single customer, contributing over one-third of its revenue

    Anthropic becomes SpaceX's largest single customer, contributing more than one-third of its revenue. The company mainly leases computing resources from SpaceX's data center in Memphis, Tennessee, with a contract valid until 2029, potential spending cap reaches 120.5 billion US dollars.

    Same story · 2 outlets
  11. 36Kr · AI65

    NVIDIA is reportedly negotiating to provide multi-billion dollar debt credit guarantees for the Nscale data center project.

    NVIDIA is in talks with Nscale to provide tens of billions of dollars in debt credit guarantees for the development of a 460 megawatt data center rented by Anthropic; the transaction has not been finalized, and NVIDIA continues to provide financial support to new cloud companies such as Nscale, CoreWeave, and Firmus.

    Same story · 3 outlets