Relevance: primary · Type: event
Confidence90%
Anthropic announced this week that it had developed a new AI model it believes could reshape cybersecurity.
Relevance: primary · Type: event
Confidence80%
Anthropic announced its new model on Tuesday.
Relevance: primary · Type: event
Confidence90%
The model, Mythos Preview, was able to find high-severity vulnerabilities, including some in every major operating system and web browser.
Relevance: primary · Type: event
Confidence90%
Mythos Preview can generate methods to exploit the vulnerabilities it finds.
Relevance: primary · Type: action
Confidence90%
Anthropic is limiting access to the model to around 50 select companies and organizations.
Relevance: supporting · Type: background
Confidence90%
The collaboration is called Project Glasswing, named after a butterfly species with transparent wings.
Relevance: primary · Type: action
Confidence90%
Anthropic has no plans to release this particular model to the general public.
Relevance: supporting · Type: action
Confidence80%
Anthropic plans to release other related models.
Relevance: supporting · Type: quote
Confidence90%
"Our eventual goal is to enable our users to safely deploy Mythos-class models at scale," the company wrote.
Daniel Blackford, VP of Threat Research at Proofpoint
Relevance: supporting · Type: quote
Confidence90%
"I don't necessarily think that the average computer user needs to be fundamentally worried about this," said Daniel Blackford, VP of Threat Research at Proofpoint.
Daniel Blackford, VP of Threat Research at Proofpoint
Relevance: supporting · Type: quote
Confidence90%
"They need to be way more worried about not giving their password away because that just happens like all day, every day," said Daniel Blackford.
Jim Zemlin, CEO of the Linux Foundation
Relevance: supporting · Type: background
Confidence90%
Jim Zemlin, CEO of the Linux Foundation, said that Mythos Preview may have better capabilities to help developers fix vulnerabilities than previous models.
Relevance: supporting · Type: background
Confidence90%
A kernel is an interface that lets hardware talk to software.
Relevance: supporting · Type: background
Confidence90%
The Linux kernel powers some of the most widely used operating systems, including Android, as well as all of the world's 500 most powerful supercomputers.
Relevance: supporting · Type: background
Confidence90%
The Linux Foundation is part of Project Glasswing.
Relevance: supporting · Type: event
Confidence90%
A core group of Linux kernel maintainers have started experimenting with the model to determine the most effective way to use it.
Jim Zemlin, CEO of the Linux Foundation
Relevance: supporting · Type: quote
Confidence90%
"These maintainers are already overworked before AI," said Jim Zemlin.
Jim Zemlin, CEO of the Linux Foundation
Relevance: supporting · Type: quote
Confidence90%
"This just makes their lives a lot better," said Jim Zemlin.
Relevance: supporting · Type: background
Confidence80%
Hackers are using AI to help them find bugs.
Daniel Stenberg, Lead software developer behind cURL
Relevance: supporting · Type: background
Confidence90%
Daniel Stenberg said that AI models' capabilities for finding bugs improved noticeably in early 2026.
Relevance: supporting · Type: background
Confidence90%
Daniel Stenberg is the lead software developer behind cURL.
Relevance: supporting · Type: background
Confidence90%
cURL is a 30-year-old open-source data transfer tool widely used for things that connect to the internet, including cars and medical devices.
Relevance: supporting · Type: background
Confidence80%
The improvement in AI models' bug-finding capabilities followed the release of new cutting-edge models in late 2025.
Relevance: supporting · Type: background
Confidence80%
White hat hackers flag security flaws to developers privately and sometimes receive reward money or have security flaws named after them.
Relevance: supporting · Type: background
Confidence90%
Some software bugs present security vulnerabilities while others only affect functionality.
Daniel Stenberg, Lead software developer behind cURL
Relevance: supporting · Type: quote
Confidence90%
"We got 185 reports throughout the year and less than 5% of them were actually security related problems in the end," said Daniel Stenberg.
Relevance: supporting · Type: background
Confidence80%
Stenberg's team found and fixed fewer security vulnerabilities in 2025 than in 2024 even though the volume of reports doubled.
Daniel Stenberg, Lead software developer behind cURL
Relevance: supporting · Type: action
Confidence90%
Daniel Stenberg stopped paying bug bounty rewards for his cURL tool due to an influx of low-quality reports.
Daniel Stenberg, Lead software developer behind cURL
Relevance: supporting · Type: background
Confidence90%
Daniel Stenberg said that the hackers who file reports to him are usually anonymous and do not discuss whether they used AI.
Daniel Stenberg, Lead software developer behind cURL
Relevance: supporting · Type: quote
Confidence90%
"They tend to be very elaborate and descriptive. You get a 400 line report when it's something that a human would take 50 lines to present," said Daniel Stenberg.
Relevance: supporting · Type: background
Confidence90%
HackerOne surveyed hackers in summer 2025 and found that nearly 60% of respondents were either using AI, learning it, or learning to audit AI or machine learning systems.
Nicholas Carlini, Research scientist at Anthropic
Relevance: supporting · Type: event
Confidence90%
Nicholas Carlini found vulnerabilities in the Linux kernel using an older Anthropic model and a relatively simple prompt.
Nicholas Carlini, Research scientist at Anthropic
Relevance: supporting · Type: event
Confidence90%
Nicholas Carlini used AI to find the first critical vulnerability in another 20-year-old open source project.
Alex Stamos, Chief security officer at Corridor
Relevance: supporting · Type: quote
Confidence90%
"LLMs have now bypassed human capability for bug finding," said Alex Stamos, chief security officer at Corridor.
Alex Stamos, Chief security officer at Corridor
Relevance: supporting · Type: background
Confidence80%
Alex Stamos said that the increase in quality of security research reports followed the release of Anthropic's model Opus 4.5 in November 2025.
Relevance: supporting · Type: background
Confidence90%
Alex Stamos was previously head of security at Yahoo and Facebook.
Alex Stamos, Chief security officer at Corridor
Relevance: supporting · Type: background
Confidence80%
Alex Stamos said that because much commercial software has open-source components, impacts on open-source projects have wider implications for the internet.
Relevance: supporting · Type: background
Confidence80%
The new tool has uncovered thousands of weak points in every major operating system and web browser.
Alissa Valentina Knight, CEO of Assail
Relevance: supporting · Type: quote
Confidence90%
"What we need to do is look at this as a wake-up call to say, the storm isn't coming — the storm is here. We need to prepare ourselves, because we couldn't keep up with the bad guys when it was humans hacking into our networks. We certainly can't keep up now if they're using AI because it's so much devastatingly faster and more capable," said Alissa Valentina Knight.
Relevance: supporting · Type: event
Confidence90%
Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell met with top bank CEOs in a closed-door meeting to discuss Mythos and other emerging cybersecurity risks stemming from AI.
Relevance: supporting · Type: event
Confidence90%
Anthropic briefed senior U.S. government officials and key industry stakeholders on Mythos's capabilities.
Kristalina Georgieva, Managing Director of the International Monetary Fund
Relevance: supporting · Type: background
Confidence90%
Kristalina Georgieva said that the world does not have the ability to protect the international monetary system against massive cyber risks.
Kristalina Georgieva, Managing Director of the International Monetary Fund
Relevance: supporting · Type: quote
Confidence90%
"The risks have been growing exponentially," said Kristalina Georgieva.
Kristalina Georgieva, Managing Director of the International Monetary Fund
Relevance: supporting · Type: quote
Confidence90%
"Yes, we are concerned. We are very keen to see more attention to the guardrails that are necessary to protect financial stability in the world of AI," said Kristalina Georgieva.
Relevance: supporting · Type: quote
Confidence90%
Anthropic said, "The fallout — for economies, public safety, and national security — could be severe."
Relevance: supporting · Type: background
Confidence80%
Hackers have access to advanced AI models and are using them to create autonomous agents capable of carrying out attacks without human intervention, including spreading malware, executing identity theft scams, producing deepfake videos, and launching ransomware attacks.
Relevance: supporting · Type: quote
Confidence90%
"AI-enabled tooling has empowered even low-skilled threat actors to execute high-speed, high-volume operations, whilst advanced adversaries are using AI to sharpen precision, scale automation and compress attack timelines," PwC said in a recent report.
Relevance: supporting · Type: quote
Confidence90%
"The time between the public release of a new capability by an AI company and its weaponization by threat actors shrank dramatically [in 2025], a trend we assess will likely accelerate in 2026," the management consulting firm added.
forum Comments (0)
No comments yet. Be the first to comment.