Anthropic has revealed details about a new experimental AI model called Claude Mythos, which the company says is its most advanced system so far. According to Anthropic, the model has demonstrated a strong ability to detect serious cybersecurity flaws in widely used software systems. Because of its capabilities, the company is not releasing the model publicly yet and is instead working with technology partners to fix the vulnerabilities it discovered.The announcement reflects how AI tools are increasingly being used in cybersecurity research to identify weaknesses in complex software systems.<h2>Claude Mythos identified thousands of software vulnerabilities</h2>In a blog post, Anthropic said Claude Mythos Preview was able to identify thousands of previously unknown vulnerabilities in widely used software.These include “zero-day vulnerabilities”, which are security flaws that software developers have not yet discovered. Zero-day vulnerabilities are particularly concerning because attackers can exploit them before developers have time to fix them.<h2>Anthropic said the AI model identified vulnerabilities across:</h2>Web browsersOperating systemsCore system softwareThe company added that some of the vulnerabilities could potentially allow attackers to escalate access from a normal user account to full control of a system.Vulnerabilities discovered in OpenBSD and Linux systemsAmong the findings was a 27-year-old vulnerability in OpenBSD, an operating system widely used in security-focused infrastructure such as firewalls and servers. The model also identified several vulnerabilities in the Linux kernel, which powers the majority of servers worldwide and forms the backbone of many internet services.Security researchers say the Linux kernel is one of the most widely analysed pieces of software in the world, which makes the discovery of new vulnerabilities particularly notable.According to Anthropic, these flaws could potentially allow attackers to gain elevated access to systems if exploited.<h2>Anthropic working with 40 technology companies</h2>Because of the security implications, Anthropic said it is collaborating with more than 40 technology companies to address the vulnerabilities identified by Claude Mythos.The goal is to fix the security issues before the model is released more broadly. This approach reflects a growing concern in the AI industry: extremely capable AI systems could potentially identify security weaknesses faster than humans, which creates both benefits and risks.Anthropic said controlled testing with industry partners will help ensure that vulnerabilities are addressed responsibly.<h2>Why the model is not yet publicly released</h2>Although Claude Mythos reportedly outperforms earlier models such as Claude Opus 4.6 and other competing AI systems on several benchmarks, Anthropic is taking a cautious approach.The company said the model’s cybersecurity capabilities mean it needs additional testing and safeguards before public deployment.Anthropic CEO Dario Amodei has previously warned that highly capable AI models may present new risks if released without careful evaluation. Earlier reports also suggested that the model’s existence first became known after references appeared in a data leak related to Anthropic’s internal systems.<h2>AI and the future of cybersecurity</h2>Cybersecurity experts increasingly see artificial intelligence as a powerful tool for analysing large codebases and identifying weaknesses that human researchers may overlook.Traditional vulnerability discovery often requires months of manual review by security teams. AI systems can scan code at scale and highlight suspicious patterns within minutes.However, experts say this capability must be handled carefully. The same technology that identifies vulnerabilities could potentially be misused if released without safeguards. Anthropic’s decision to work with industry partners before releasing Claude Mythos suggests that AI companies are becoming more cautious about how powerful models are deployed. As AI research continues to advance, tools like Claude Mythos may play an important role in strengthening digital security across software systems used worldwide.&nbsp;

Anthropic has revealed an experimental AI system called Claude Mythos, which reportedly discovered thousands of previously unknown cybersecurity vulnerabilities, including zero-day flaws in systems such as OpenBSD and the Linux kernel. 

Anthropic has revealed details about a new experimental AI model called Claude Mythos, which the company says is its most advanced system so far. According to Anthropic, the model has demonstrated a strong ability to detect serious cybersecurity flaws in widely used software systems. Because of its capabilities, the company is not releasing the model publicly yet and is instead working with technology partners to fix the vulnerabilities it discovered. The announcement reflects how AI tools are increasingly being used in cybersecurity research to identify weaknesses in complex software systems. Claude Mythos identified thousands of software vulnerabilities In a blog post, Anthropic said Claude Mythos Preview was able to identify thousands of previously unknown vulnerabilities in widely used software. These include “zero-day vulnerabilities”, which are security flaws that software developers have not yet discovered. Zero-day vulnerabilities are particularly concerning because attackers can exploit them before developers have time to fix them. Anthropic said the AI model identified vulnerabilities across: Web browsers Operating systems Core system software The company added that some of the vulnerabilities could potentially allow attackers to escalate access from a normal user account to full control of a system. Vulnerabilities discovered in OpenBSD and Linux systems Among the findings was a 27-year-old vulnerability in OpenBSD, an operating system widely used in security-focused infrastructure such as firewalls and servers. The model also identified several vulnerabilities in the Linux kernel, which powers the majority of servers worldwide and forms the backbone of many internet services. Security researchers say the Linux kernel is one of the most widely analysed pieces of software in the world, which makes the discovery of new vulnerabilities particularly notable. According to Anthropic, these flaws could potentially allow attackers to gain elevated access to systems if exploited. Anthropic working with 40 technology companies Because of the security implications, Anthropic said it is collaborating with more than 40 technology companies to address the vulnerabilities identified by Claude Mythos. The goal is to fix the security issues before the model is released more broadly. This approach reflects a growing concern in the AI industry: extremely capable AI systems could potentially identify security weaknesses faster than humans, which creates both benefits and risks. Anthropic said controlled testing with industry partners will help ensure that vulnerabilities are addressed responsibly. Why the model is not yet publicly released Although Claude Mythos reportedly outperforms earlier models such as Claude Opus 4.6 and other competing AI systems on several benchmarks, Anthropic is taking a cautious approach. The company said the model’s cybersecurity capabilities mean it needs additional testing and safeguards before public deployment. Anthropic CEO Dario Amodei has previously warned that highly capable AI models may present new risks if released without careful evaluation. Earlier reports also suggested that the model’s existence first became known after references appeared in a data leak related to Anthropic’s internal systems. AI and the future of cybersecurity Cybersecurity experts increasingly see artificial intelligence as a powerful tool for analysing large codebases and identifying weaknesses that human researchers may overlook. Traditional vulnerability discovery often requires months of manual review by security teams. AI systems can scan code at scale and highlight suspicious patterns within minutes. However, experts say this capability must be handled carefully. The same technology that identifies vulnerabilities could potentially be misused if released without safeguards. Anthropic’s decision to work with industry partners before releasing Claude Mythos suggests that AI companies are becoming more cautious about how powerful models are deployed. As AI research continues to advance, tools like Claude Mythos may play an important role in strengthening digital security across software systems used worldwide.

Anthropic has revealed an experimental AI system called Claude Mythos, which reportedly discovered thousands of previously unknown cybersecurity vulnerabilities, including zero-day flaws in systems such as OpenBSD and the Linux kernel.&nbsp;

‘Claude Mythos in cybersecurity’: How Anthropic’s new AI model can find hidden software vulnerabilities

A large portion of the internal interface code used by the Claude chatbot was accidentally exposed online after a technical error by Anthropic, the company behind the AI system.The incident happened early Tuesday morning when the company pushed what was expected to be a routine update to Claude. However, the update mistakenly included a debugging “source map” file that linked directly to the platform’s internal code.The file reportedly contained around 512,000 lines of TypeScript code, revealing parts of the interface used to run Claude’s features.The issue was first spotted by users on social media platform X, after which the code was quickly downloaded and shared across developer forums and repositories.<h2>How the Claude source code leak happened</h2>According to reports from technology publication VentureBeat, the leak occurred around 4 a.m. when Anthropic released a software update.The update mistakenly included a source map file a debugging tool that developers use to trace errors in code. In this case, the file provided a pathway to Claude’s underlying interface code. Once the link was discovered, the code package was downloaded and circulated widely online. Within hours, developers began examining the files and recreating some parts of the interface.Anthropic later confirmed that the issue was the result of human error during the update process. The company said there was no external hacking or malicious activity involved.<h2>What information was exposed</h2>The leaked files mainly contain the interface code used to run Claude’s user experience. According to the company, the incident did not expose model weights or AI training data, which are the most sensitive parts of the system.Anthropic also stated that no customer data was compromised.However, the interface code itself still has significant value because it reveals how the company structures its tools and features. Developers reviewing the files have already highlighted several internal components that had not previously been made public.<h2>Developers already analysing the leaked code</h2>Because the files spread quickly online, parts of the code are already being analysed by developers.Some users have reportedly begun recreating elements of Claude’s internal interface for experimental purposes. One interesting detail discovered in the code is a feature resembling a Tamagotchi-style interface, where users interact with stat-based digital companions.It is not yet clear whether this feature was planned for public release or was still under development.<h2>Anthropic moves to contain the spread</h2>After the leak began circulating, Anthropic issued DMCA takedown requests to remove repositories hosting the files.However, because the code spread quickly across the internet, experts say it may be difficult to remove it completely. The company said the leaked files represent only a snapshot of Claude’s development at a specific time.Anthropic continues to update its AI models and software regularly, which means future versions of the system may already differ significantly from the leaked code.<h2>Why the leak matters in the AI industry</h2>The incident comes at a time when competition between AI companies is increasing.Major firms including OpenAI, Google, and Anthropic are rapidly releasing new AI features and products.Even though the leak does not expose the core AI model, analysts say interface code can still provide insights into design approaches, system architecture and product strategy. For Anthropic, the immediate concern will be protecting intellectual property and preventing competitors from replicating elements of its user interface.The company has not indicated whether the leak will delay future updates to Claude.

A technical error by Anthropic accidentally exposed around 512,000 lines of interface code used by the Claude platform. The leak occurred when a routine update mistakenly included a debugging source map file, allowing developers to access internal TypeScript code. 

A large portion of the internal interface code used by the Claude chatbot was accidentally exposed online after a technical error by Anthropic, the company behind the AI system. The incident happened early Tuesday morning when the company pushed what was expected to be a routine update to Claude. However, the update mistakenly included a debugging “source map” file that linked directly to the platform’s internal code. The file reportedly contained around 512,000 lines of TypeScript code, revealing parts of the interface used to run Claude’s features. The issue was first spotted by users on social media platform X, after which the code was quickly downloaded and shared across developer forums and repositories. How the Claude source code leak happened According to reports from technology publication VentureBeat, the leak occurred around 4 a.m. when Anthropic released a software update. The update mistakenly included a source map file a debugging tool that developers use to trace errors in code. In this case, the file provided a pathway to Claude’s underlying interface code. Once the link was discovered, the code package was downloaded and circulated widely online. Within hours, developers began examining the files and recreating some parts of the interface. Anthropic later confirmed that the issue was the result of human error during the update process. The company said there was no external hacking or malicious activity involved. What information was exposed The leaked files mainly contain the interface code used to run Claude’s user experience. According to the company, the incident did not expose model weights or AI training data, which are the most sensitive parts of the system. Anthropic also stated that no customer data was compromised. However, the interface code itself still has significant value because it reveals how the company structures its tools and features. Developers reviewing the files have already highlighted several internal components that had not previously been made public. Developers already analysing the leaked code Because the files spread quickly online, parts of the code are already being analysed by developers. Some users have reportedly begun recreating elements of Claude’s internal interface for experimental purposes. One interesting detail discovered in the code is a feature resembling a Tamagotchi-style interface, where users interact with stat-based digital companions. It is not yet clear whether this feature was planned for public release or was still under development. Anthropic moves to contain the spread After the leak began circulating, Anthropic issued DMCA takedown requests to remove repositories hosting the files. However, because the code spread quickly across the internet, experts say it may be difficult to remove it completely. The company said the leaked files represent only a snapshot of Claude’s development at a specific time. Anthropic continues to update its AI models and software regularly, which means future versions of the system may already differ significantly from the leaked code. Why the leak matters in the AI industry The incident comes at a time when competition between AI companies is increasing. Major firms including OpenAI, Google, and Anthropic are rapidly releasing new AI features and products. Even though the leak does not expose the core AI model, analysts say interface code can still provide insights into design approaches, system architecture and product strategy. For Anthropic, the immediate concern will be protecting intellectual property and preventing competitors from replicating elements of its user interface. The company has not indicated whether the leak will delay future updates to Claude.

A technical error by Anthropic accidentally exposed around 512,000 lines of interface code used by the Claude platform. The leak occurred when a routine update mistakenly included a debugging source map file, allowing developers to access internal TypeScript code.&nbsp;

'512,000 lines of Claude source code leaked': Anthropic confirms accidental leak of its Claude chatbot source code

A senior engineer at Anthropic has warned that a new generation of AI agents could reshape nearly every computer-based job.&nbsp;Boris Cherny, creator of Claude Code, said AI systems that can operate computers like humans are advancing quickly and may soon change how millions of people work.&nbsp;Speaking on Lenny's Podcast, Cherny said the impact will go beyond software development.&nbsp;“It’s going to expand to pretty much any kind of work that you can do on a computer,” he said, adding that the shift will be “very disruptive” and “painful for a lot of people”.&nbsp;<h2>What are AI agents and how are they different?</h2>Unlike traditional chatbots that only generate text or images, AI agents can take actions across digital tools.&nbsp;Claude Code, which is built on Anthropic’s Claude models, can:&nbsp;Run commands on a computer&nbsp;Analyse documents&nbsp;Write and test code&nbsp;Use workplace apps&nbsp;Build websites&nbsp;Anthropic released its latest update, Opus 4.6, in early February. And these models are designed to handle multi-step tasks and support engineers in complex workflows.&nbsp;Cherny said AI agents are beginning to use computers in ways similar to humans, though they have not yet reached the level of a skilled professional.&nbsp;<h2>Which jobs could be affected?</h2>He said the change will not be limited to software engineers. He mentioned product managers, designers and other knowledge workers whose jobs depend heavily on computers.&nbsp;In another interview on Y Combinator’s Lightcone podcast, he suggested that the traditional job title “software engineer” may begin to fade as AI tools handle more coding tasks.&nbsp;He also claimed that productivity per engineer at Anthropic has increased since Claude Code was introduced.&nbsp;However, he acknowledged that the broader economic impact remains uncertain.&nbsp;“As a society, this is a conversation we have to figure out together,” he said.&nbsp;<h2>Will AI replace jobs or reshape them?</h2>The warning comes at a time when companies across the tech sector are investing heavily in AI automation. Many firms are exploring how AI tools can reduce repetitive work and speed up development cycles.&nbsp;Experts say AI is likely to automate certain tasks rather than eliminate entire professions overnight. Tasks such as writing basic code, analysing structured data, drafting documents or handling routine communication may increasingly be handled by AI systems.&nbsp;However, human oversight, judgement, creativity and responsibility remain critical in many industries.&nbsp;Cherny’s advice to workers was simple: experiment with AI tools and understand how they function.&nbsp;“Don’t be scared of them,” he said.&nbsp;A wider debate on the future of work&nbsp;The rise of agent-based AI systems has intensified debate around the future of white-collar jobs. Some analysts believe AI will increase productivity and create new roles. Others warn of short-term job displacement.&nbsp;Anthropic, which competes with firms such as OpenAI and Google in advanced AI systems, has positioned Claude Code as a tool to support developers and enterprises.&nbsp;Whether AI agents will replace jobs or transform them remains unclear. But as these systems gain the ability to operate computers directly, the structure of computer-based work could change faster than expected.

Anthropic engineer Boris Cherny has warned that advanced AI agents, such as Claude Code, could disrupt nearly all computer-based jobs. Unlike chatbots, these agents can run commands, write code and use workplace. While full job replacement is unlikely, major task-level disruption is expected.

A senior engineer at Anthropic has warned that a new generation of AI agents could reshape nearly every computer-based job. Boris Cherny, creator of Claude Code, said AI systems that can operate computers like humans are advancing quickly and may soon change how millions of people work. Speaking on Lenny's Podcast, Cherny said the impact will go beyond software development. “It’s going to expand to pretty much any kind of work that you can do on a computer,” he said, adding that the shift will be “very disruptive” and “painful for a lot of people”. What are AI agents and how are they different? Unlike traditional chatbots that only generate text or images, AI agents can take actions across digital tools. Claude Code, which is built on Anthropic’s Claude models, can: Run commands on a computer Analyse documents Write and test code Use workplace apps Build websites Anthropic released its latest update, Opus 4.6, in early February. And these models are designed to handle multi-step tasks and support engineers in complex workflows. Cherny said AI agents are beginning to use computers in ways similar to humans, though they have not yet reached the level of a skilled professional. Which jobs could be affected? He said the change will not be limited to software engineers. He mentioned product managers, designers and other knowledge workers whose jobs depend heavily on computers. In another interview on Y Combinator’s Lightcone podcast, he suggested that the traditional job title “software engineer” may begin to fade as AI tools handle more coding tasks. He also claimed that productivity per engineer at Anthropic has increased since Claude Code was introduced. However, he acknowledged that the broader economic impact remains uncertain. “As a society, this is a conversation we have to figure out together,” he said. Will AI replace jobs or reshape them? The warning comes at a time when companies across the tech sector are investing heavily in AI automation. Many firms are exploring how AI tools can reduce repetitive work and speed up development cycles. Experts say AI is likely to automate certain tasks rather than eliminate entire professions overnight. Tasks such as writing basic code, analysing structured data, drafting documents or handling routine communication may increasingly be handled by AI systems. However, human oversight, judgement, creativity and responsibility remain critical in many industries. Cherny’s advice to workers was simple: experiment with AI tools and understand how they function. “Don’t be scared of them,” he said. A wider debate on the future of work The rise of agent-based AI systems has intensified debate around the future of white-collar jobs. Some analysts believe AI will increase productivity and create new roles. Others warn of short-term job displacement. Anthropic, which competes with firms such as OpenAI and Google in advanced AI systems, has positioned Claude Code as a tool to support developers and enterprises. Whether AI agents will replace jobs or transform them remains unclear. But as these systems gain the ability to operate computers directly, the structure of computer-based work could change faster than expected.

Anthropic engineer Boris Cherny has warned that advanced AI agents, such as Claude Code, could disrupt nearly all computer-based jobs. Unlike chatbots, these agents can run commands, write code and use workplace. While full job replacement is unlikely, major task-level disruption is expected.

Can AI operate your computer better than you? Here’s what a senior engineer at Anthropic said about AI agents

claude-code

Claude Code, News on Claude Code, stories on Claude Code, articles on Claude Code, Claude Code stories at Wion, Claude Code news, Claude Code latest news

claude code

‘Claude Mythos in cybersecurity’: How Anthropic’s new AI model can find hidden software vulnerabilities

Anthropic confirms accidental leak of Claude chatbot source code

Top Anthropic engineer says AI agents will change every computer-based job