Anthropic

According to the 24 April news, earlier this week, Microsoft announced that it would suspend the registration of new users of GitHub Copilo Pro, Pro+ and the student version of the food package in order to serve current clients more efficiently, a move that was unexpected to all. In addition, Microsoft announced a downward revision of the use of the personal package and removed the Claude Opus model from the Pro package. Thanks to the upgrading of the AI model and its increased intelligence capabilities, GitHub Copilot has experienced a dramatic increase in usage over the past few months. This phenomenon is not..

On April 23, according to news, Alexander Hanff, a cyber security expert, published an article on April 18, stating that Claude Desktop of Anthropic was used to unknowingly install files silently from Section 7 Chromium Browser. The Hanff survey noted that after the user installed Claude Desktop, Anthropic automatically turned to..

On April 19, according to the news, the technical media PCWORD published a Boe letter today (18 April) and the test showed that the AI Design Tool, Claude Design, published by Anthropic, is powerful, but tokens is extremely expensive and 25 minutes is spent on the 80% weekly quota for Claude Pro users. 1AI reports today that the design of white users can also produce excellent visual work by the Claude Design tool, and that users can give birth to their experts by textwords..

On 19 April, according to the Financial Times, Anthropic CEO Dario Amodei said in an interview recently that AI would replace a junior white collar position of up to 50% within five years. He stated that the industry could not avoid the shock. We should not deny that subversion will take place, but we need to do enough to make it a tool for dealing with subversion. At the same time, Amodei points out that at present AI technology can only be "spread at the rate of trust" and that the industry has yet to live up to its earlier promise of optimism, which is still scarce. For AI..

On April 17th, Anthony officially released last night the Claude Opus 4.7 flagship model, priced in line with the previous generation of Opus 4.6: $5 per million input token and $25 per million output, which can be called by developers through claude-opus-4-7. It was described that Opus 4.7 was able to maintain rigour and consistency in complex missions that operated over a long period of time and had the capability to validate the output in advance of the return. Cursor's internal evaluation..

On April 16, according to The Information, Anthropic is preparing to launch the next generation of flagship models Claude Opus 4.7, and will simultaneously release an AI-driven design tool, both of which are the fastest available this week. The newly launched AI design tool is designed to help users to do ppt, website, drop-off page and product prototypes, directly against the goal ppt generation tool Gamma and Google Stitch. After the news came out, Adobe, Wix..

On April 8th, local time 7th, Anthropic released a new-generation flagship model, Claude Mythos Preview, and simultaneously announced the launch of Project Glasswing, which limited the authorization of the model to 12 technocrats, such as Apple, Microsoft, Amazon and others, for defensive cyber security, and was not open to the public. Mythos Preview shows a cliff-break lead in several benchmark tests. Anthropic says it's the biggest model it's ever trained, the power..

News from April 7, yesterday, the director of the Mimo team, Roofuli (@LuoFuli), wrote to X to evaluate the recent denial of access to Claude's subscription service by Anthropic to the Third Party Tool Framework (harness). This is known to have a direct impact on the AI smarts development tools that rely on Claude subscription interfaces, such as Lobster OpenClaw. Roosevelt pointed out that Claude Code's subscription mechanism itself was a well-designed balanced distribution of power..

On April 6th, Claude Code subscriber to the OpenClaw and other third-party tools using Anthropic's code assistant service is about to face a cost escalation. According to a client e-mail that is being uploaded on Hacker News, Anthropic announces that since midday, April 4th, Pacific Time, the subscriber “can no longer use your Claude subscription for third-party integration tools such as OpenClaw”. Users are required to charge on demand through separate subscription services..

In a message dated 1 April, the spokesman for Anthropic yesterday (31 March) issued a statement through the CNET media in response to a leak of Claude Code's source code, stating: “In a new version of Claude Code released this morning, it was confirmed that some internal source code was included, but no sensitive user data or documents were involved or disclosed”. Anthropic blames the human error of this Claude Code source leak and the company has refined its measures to prevent the recurrence of such incidents. This leak..

On April 1st, yesterday afternoon, the developer Chaofan Shou, in his X message, released a full-source compression package of the Anthropic command line tool Claude Code, which attracted widespread attention from the developers community. The post was viewed 12 million times in hours. The cause of the event was that Anthropic forgot to delete the .map file (Source Map) when he delivered the package to the npm public mirror library. The purpose of such files is to map compressed, confused production codes back to the original..

On February 25, according to Time magazine, the American AI Enterprise Anthropic was a great success and claimed to be the top AI research laboratory with the highest security focus. However, executive Anthropic told The Times that the company was abandoning the core commitments of its flagship security policy. In 2023, Anthropic promised not to train the AI system unless it was assured in advance that adequate security measures were in place. This commitment became an integral part of Anthropic's Responsible Extension Policy..

On February 18, Anthropic released a new version of the size of the Sonnet model, continuing the company ' s four-month update. In the announcement of the new model, Anthropic highlighted its improvements in code capabilities, command compliance and computer operations. Sonet 4.6 will be the default model for free and Pro users. The sonet 4.6 test version will support the context window of 1 million tokens, which was previously the largest..

In the morning of February 6th, OpenAI and Anthropic released an update of the heavy-pound model, launching GPT – 5.3 - Codex and Claude Opus 4.6, respectively. OpenAI claims that GPT-5.3-Codex was its first model to play a key role in its own research and development, and that the team used an early version to participate in commissioning training, deployment management and evaluation analysis. The model is in SWE-Bench Pro, Terminal-Bench 2.0 and OSWor..

On February 2nd, yesterday evening, “36 Krypton” released news that Kimi, in his recent communications with investors, had stated that the company's overseas revenues had exceeded domestic revenues, and that since the release of the new model K2.5, there had been a fourfold increase in the number of paid users worldwide. This change happens within a few days of the release of the new generation of model K2.5. On Openrouter, K2.5 has reached third place after Claude Sonet 4.5 and Gemini 3 Flash. In January 29..

On 30 January, according to The Inside of Business, Peter Steinberger, the developer of Moltbot (formerly Clawdbot), disclosed the details behind the renaming of the project in an interview he had recently given, and talked about its ripple effect. Steinberger states that, although the original name of Moltbot was associated with the mascot “Clawd” of Claude Code under the Anthropic flag, the other party did not take legal action, but rather requested a change of name by e-mail. I don't know

On 23 January, according to The Verge, in recent days Anthropic published a 57-page Claude Code document, which systematized the values, boundaries of behaviour and decision-making in high-risk situations of its large model Claude. This new document is seen as a comprehensive upgrade of the 2023 version of the Guidelines, with the aim of making the model not only "compliance with the rules " but also " understand the reasons behind the rules". Document stresses that Anthropic wants Claude to have a role for himself..

On January 19, according to the British Financial Times, the venture capital Redwood will participate in a new round of financing for the artificial intelligence enterprise Anthropic. The financing target was $25 billion (note: the current exchange rate is approximately RMB 17,469.1 million) and the company ' s valuation will amount to $350 billion (the current exchange rate is about RMB 2.4 trillion), including the Singapore Government Investment Corporation (GIC) and the United States Investment Agency Coatue. The report says that the Singapore Sovereign Wealth Fund..

On January 18th, Anthropic has now delegated Claude Cowork AI ' s assistant to all $20 (current exchange rate approximately RMB 139.8) Pro subscribers to enable their computers to handle their daily affairs more simply. Previously, Anthropic only provided the functionality to Max subscribers, which cost up to $100 a month (note: the current exchange rate is about RMB 698.8), although more high-cost reasoning models could be used and experimented with..

On January 17, according to Fortune, the latest research published by Anthropic shows that AI ' s impact on different occupations is rapidly polarizing, and that its re-engineering of the labour market is far more complicated than the simple narrative of "Ai will take away jobs " . The study is based on data from the use of the Anthropic home-grown chat robot Claude, and attempts to present in a more quantitative way how users collaborate with AI in the real work scene, assign tasks and model success rates in different tasks. Anthropic Economist

On January 14, Anthropic recently launched a new smart agent "Cowork" in a simplified version of Claude Code for non-developers, which is now open to Claude Max subscribers at the MacOS end in the form of a research preview. It was described that Cowork is built on Claude Agent SDK, with the core ability to allow users to assign local folders to Claude so that they can read, edit or create documents and perform continuous tasks with greater autonomy. Compared to..

On January 12th, local time 11th, Anthropic announced that the U.S.-A. product Claude would go online for a medical service that met the requirements of the U.S. Health Insurance Circulation and Liability Act and was open to hospitals, medical institutions and individual users for use in processing protected health data. At the same time, Claude has consolidated a wide range of scientific databases and enhanced its capacity to support biomedical research. In terms of individual user-oriented functionality, users can use anthropic from Apple Health and Function He..

On January 5, Google’s senior engineer, Jaana Dogan, publicly stated on platform X that the Claude Code, launched by Anthropic, had produced a system within only one hour, which was close to what her team had built over the past year, generating widespread interest in the industry. Dogan in Google is responsible for Gemini API related work. She revealed that in this test, the hint she submitted to Claude Code was not complicated and contained only three paragraphs..

News from December 29, recently, the co-founder of Anthropic, Jack Clark, published an article in his X account called Silent Sirens, Flashing For Us All (quiet sirens are warning everyone), pointing out that the development of AI at the current stage was an extremely contradictory "non-readability". It points out that, while there are no drones or auto-dances in real life and the social media stream is still full of ordinary pictures and web memes, there are..