惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

T
Tor Project blog
月光博客
月光博客
P
Proofpoint News Feed
大猫的无限游戏
大猫的无限游戏
N
News and Events Feed by Topic
The Cloudflare Blog
博客园_首页
NISL@THU
NISL@THU
量子位
A
Arctic Wolf
Y
Y Combinator Blog
Spread Privacy
Spread Privacy
Engineering at Meta
Engineering at Meta
F
Fortinet All Blogs
酷 壳 – CoolShell
酷 壳 – CoolShell
Cyberwarzone
Cyberwarzone
The GitHub Blog
The GitHub Blog
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
F
Full Disclosure
C
Cisco Blogs
Security Latest
Security Latest
T
The Exploit Database - CXSecurity.com
T
Tenable Blog
PCI Perspectives
PCI Perspectives
S
Security Affairs
Forbes - Security
Forbes - Security
Hugging Face - Blog
Hugging Face - Blog
C
CERT Recently Published Vulnerability Notes
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
博客园 - 聂微东
H
Hacker News: Front Page
S
Securelist
cs.CL updates on arXiv.org
cs.CL updates on arXiv.org
D
Darknet – Hacking Tools, Hacker News & Cyber Security
罗磊的独立博客
S
SegmentFault 最新的问题
C
Cyber Attacks, Cyber Crime and Cyber Security
The Hacker News
The Hacker News
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
S
Security @ Cisco Blogs
The Last Watchdog
The Last Watchdog
小众软件
小众软件
Hacker News - Newest:
Hacker News - Newest: "LLM"
Google DeepMind News
Google DeepMind News
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
IT之家
IT之家
Last Week in AI
Last Week in AI
爱范儿
爱范儿
AWS News Blog
AWS News Blog
MongoDB | Blog
MongoDB | Blog

Latest from TechRadar in Pro

VodafoneThree gets Ofcom approval to bring satellite connectivity to your smartphone Is this the tipping point for AI at work? New Gallup survey finds half of all US employees now use it in some way 'Every Apple user needs to know about this nasty scam': Fake warnings tell users their iCloud data will be… 'Makes it even more disappointing': Microsoft backs fossil fuel big time with $7 billion deal in race for AI… 'Maybe it’s not science fiction': Solar panels are causing rainwater to fall in one of the driest places… Maine becomes first US state to pass data centre construction ban Dozens of WordPress plugins hijacked to target thousands of sites Drone-killing laser weapons greenlit for use in US airspace – FAA and Defense Department say high-energy weapons are ‘ready to protect all air travelers from illicit drone use’ despite airspace restrictions and friendly-fire incidents 'We are currently being extorted' — crypto giant Kraken says it is facing extortion attack, here's… I tried 7 free MTD software – now I've ranked my top picks as a freelancer Jackery McGraw Hill becomes latest to see its Salesforce data hacked Looking for a new PC? Now might be great time to upgrade, as Gartner figures claim shipments are rising — while… The new engineering playbook: how AI design copilots are reshaping product development Farewell Surface Hub — Microsoft kills off its super-sized touchscreen displays, but you might still be able to get one if you act fast 'We have no interest in patient data in the UK': Palantir UK head defends record as criticisms rise Amazon’s new AI Bio Discovery tool can provide ‘every researcher’ with ‘lab-in-the-loop drug discovery’ – 40+ AI biology models can filter 300,000 novel antibody candidates down to the top results for testing in just weeks Over 100 Chrome Web Store extensions found stealing user data from thousands of accounts Europe wants tech sovereignty but is this realistic? Enterprise AI governance cannot live in a prompt. So where is the safety net? Why 2026 is the year of flexibility without friction: solving the multi-platform crisis OpenAI reveals its Mythos rival designed for cybersecurity pros When cyberattacks are inevitable, recovery becomes the strategy Closing the cloud complexity gap LaLiga uses AI to fight illegal streaming that costs its clubs $800m a year Intel and Google expand long-term chip partnership to power AI systems 'Chatbots respond not just to what you ask, but how you ask it': Report finds AI agents might be sucking up to… 'Smartphones have physical limitations': Report explains why AI is kickstarting a billion-dollar hardware arms… 'I’m pretty sure actually we really do not need to work for five days' Zoom CEO calls for end of traditional work schedules — says 3-day working week should become the norm 'It's more common than you think': Experts reveal how hackers are trying to hijack your inbox with these… 'This wasn’t just phishing — it was a full-service cybercrime platform': FBI reveals takedown of notorious W3LL phishing operation targeting thousands of victims From cloud to Agentic AI: Why security must evolve faster than innovation Basic-Fit gym group data breach exposes details of over 1 million members — here's what we know ‘Authorities can ask them to hand over data’: Report claims over 80% of Europeans don’t trust US and Chinese businesses to handle their data – Europe is desperate for homegrown AI, cloud, and telecoms as the rift with the US grows Booking.com confirms reservation data breach — tells customers hackers 'may have been able to access certain… Agility is the key to protecting against Malware-as-a-Service (MaaS) Rockstar hackers publish 78.6 million stolen records — but many of us will be disappointed Adobe issues emergency security patch — Reader and Acrobat users need to update now OpenAI flags third-party data issue — all macOS users should update now Linux rules on using AI-generated code - Copilot is OK, but humans must take 'full responsibility for the… Hackers use Claude and ChatGPT in 'a significant evolution in offensive capability' to breach government agencies, leak hundreds of millions of citizen records ‘You’re effed’: Palantir CEO says AI ‘will destroy humanities jobs’ – but Gen Z workers are apparently deliberately sabotaging AI rollouts in an effort to fight back 'This is not your typical run-of-the-mill malware': CPUID download page hacked and tools replaced with links… Anthropic is bringing Claude's AI power to Microsoft Word How businesses can turn AI pilots into scalable solutions AI can transform customer experiences – when it lives up to its promise 'Regain control of our digital destiny': France to ditch Windows for Linux to reduce reliance on US tech How the memory crisis is strangling the UK's data center boom ‘No Decision’ is the new breach: Why inaction is becoming a career risk for CISOs in 2026 'That shouldn’t translate into investing in AI blindly, without a clear strategy': Experts warn UK firms want to keep spending big on AI - even if they can't prove it makes a difference How AI is rewriting the ERP investment playbook Rockstar confirms major third-party data breach: GTA VI maker says 'no impact on our organization or our… How to deploy physical AI effectively '71% of US households get routers from ISPs': Why new FCC rules could leave millions stuck with outdated,… 'The CPU is the system’s executive layer': Intel joins SambaNova as both face existential threat from… 'Just not sustainable': Why your monthly £25 broadband internet bill could soon hit £45 '$15K bill destroyed a solo developer’s startup': How hackers are using leaked Google API keys to… 'Today is the day you've been waiting for': eGPUs can now officially turn a humble Mac Mini into an AI… Linux pulls support for ancient CPU — unsurprisingly, Linus Torvald says there is 'zero real reason' to… 'AI is a once-in-a-lifetime opportunity': Amazon CEO Andy Jassy lays out his '6 truths' for the… 'A self-inflicted hit': Washington state just rolled back sales tax exemptions for AI data centers worth… 'There’s no one-size-fits-all office chair': Vari explains the design decisions behind its award-winning… 'Small business owners have significant creative control from start to finish' — VistaPrint reveals the… 'Experts' to rent for $1 per month: Hostinger debuts 7-person AI team to help SMBs save thousands on… Microsoft hands Linux Foundation key Surface data to help fix laptop battery life Adobe Reader users beware — experts flag months-old security flaw using booby-trapped PDFs to scope out victims 'Shockingly good value': New rugged Android tablet has a built-in 1080p projector, night-vision camera, and… Stop the presses — Microsoft is actually cutting cloud PC prices for SMBs, promises to make it 'more cost-effective for small and medium businesses' 'If one piece of your supply chain is delayed, then your whole project can't deliver': Nearly half of US data centers planned for 2026 canceled or delayed — and things could soon get much worse ChatGPT’s hidden backup model just got smarter — as OpenAI adds a cheaper Pro option 'The problem is not AI’s capability...what won’t improve on its own is the human side': Major study claims white-collar workers are fighting back against AI in the workplace Introducing Perspectives — the new home for premium contributed content on TechRadar Pro Introducing Perspectives — the new home for premium contributed content on TechRadar Pro The New Internet is Coming Lazarus and Kimsuky prove why infrastructure-level analysis is crucial for cybersecurity Claude Cowork is now available for enterprise use, adds analytics, access controls and more The internet has a trust problem - identity needs to travel OpenAI halts £31 billion Stargate UK project over rising energy costs and regulatory deadlock The 70% rule: Why your AI strategy is a people strategy Top WordPress Slider plugin hijacked to spread malware — here's what to look out for Why CIOs need a single source of truth for digital operations No, Elon Musk doesn't want to give you a $5,000 tax refund — it's a scam, here's what to look out… Intermedia Unite review 2026 Why enterprise AI will be defined by integration, not model aggregation ‘It’s a potential national security threat’: Proton study finds over 3,500 US legislators’ official emails leaked and exposed on the dark web Microsoft warns worrying security flaw exposed over 50 million Android users, says 'user credentials and financial… Google Chrome rolls out a new tool to try and stop infostealer malware in its tracks How to submit an article for TechRadar Pro Perspectives 'Orwellian Notion': Federal workers can access Claude AI again after judge ditches Trump's Anthropic ban 'Almost 100 TOPS': GMKTec debuts powerful AI Mini PC that supports three 8K screens and costs less than you… 'Remember BlackBerry?': Iconic phone maker’s patents used to hit Brother in a massive lawsuit that could… Breach exposes sensitive LAPD files stored in city attorney system ‘FlamingChina’ hacker claims to have stolen over 10 petabytes of advanced military data from China’s National Supercomputing Center in possibly the biggest hack of all time Mac users beware — experts say this attack 'stood out immediately' by making a major change to try… Could AMD's former foundry be quietly building up to become a major Arm — and AMD — rival? Now that's different - hackers use miniature SVG images to try and hide credit card stealer "A future-proof powerhouse for demanding tasks": MSI's RTX5090 creative laptop gets a $300 price cut… Closing the implementation gap in America's cyber strategy UK NHS chief champions Palantir’s 'outstanding results’ in England, pushes for deeper rollout despite… French email provider accidentally leaked 40 million records — L’Oreal, Renault, French government data…
'After one month, most partners have each found hundreds of critical- or high-severity vulnerabilities': Anthropic claims Mythos has found over ten thousand major security vulnerabilities across 'the most systemically important software in the world'
Sead Fadilpašić · 2026-05-25 · via Latest from TechRadar in Pro
A robot hand touching a locked digital shield blocking a human from accessing data
(Image credit: Blue Planet Studio/Shutterstock)

  • Anthropic reported Mythos Preview uncovered over 10,000 high‑ and critical‑severity vulnerabilities in under two months, with Cloudflare alone finding 2,000 bugs
  • Independent validation confirmed 90% of assessed findings as real, though critics argue the breakthrough may stem from massive compute and workflows rather than unique reasoning
  • The bottleneck has shifted from discovery to verification, disclosure, and patching, as AI now surfaces vulnerabilities faster than organizations can remediate them

In less than two months, Anthropic and friends have apparently discovered more than ten thousand critical and high-level security vulnerabilities using the famed Mythos Preview artificial intelligence tool.

In a brief update on the state of the project, published late last week, Anthropic said that since the release of the tool, roughly 50 organizations that got to use it each found “hundreds” of vulnerabilities.

“Several have told us that their rate of bug-finding has increased by more than a factor of ten,” the company said. “For instance, Cloudflare has found 2,000 bugs (400 of which are high- or critical-severity) across their critical-path systems, with a false positive rate that Cloudflare’s team considers better than human testers.”

Anthropic explained that sharing details on vulnerabilities is usually done with a 90-day delay, to give users enough time to patch and to not put anyone at risk of compromise. Therefore, it only shared general, “illustrative examples” to prove, once again, just how powerful the tool is.

With that in mind, it said that Mythos found an estimate of 6,202 high- or critical-severity vulnerabilities in these projects (out of 23,019 in total, including those it estimates as medium- or low-severity).

Skepticism lingers

Of those, 1,752 have been assessed by independent security researchers, and 90% were confirmed as valid positives, while 62.4% were confirmed as either high- or critical-severity.

But while the overall reaction to Mythos Preview has been extremely positive, there are voices saying the hype may be overstated, as well. Techzine analysis, for example, argues that AI-assisted vulnerability discovery already existed through systems like Google’s Big Sleep, and that the real challenge is still human operational security.

Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed!

A recent academic paper, “Benchmarking Mythos-Linked Bug Rediscovery,” found that under controlled conditions, public frontier models like GPT-5.5 were able to rediscover some of the same vulnerabilities attributed to Mythos, and on Reddit, different communities have been even more skeptical. The key takeaway seems to be that Mythos may simply be using enormous amounts of compute and long-running agentic workflows rather than possessing qualitatively different reasoning abilities.

In any case, Anthropic now says that progress on software vulnerability is no longer limited by the speed of discovery, but rather by the speed of verification, disclosure, and patching.


Best antivirus software header

Google logo on a black background next to text reading 'Click to follow TechRadar'

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.


Sead is a seasoned freelance journalist based in Sarajevo, Bosnia and Herzegovina. He writes about IT (cloud, IoT, 5G, VPN) and cybersecurity (ransomware, data breaches, laws and regulations). In his career, spanning more than a decade, he’s written for numerous media outlets, including Al Jazeera Balkans. He’s also held several modules on content writing for Represent Communications.