AI Found Thousands of Flaws in Days. Humans Can't Patch Them Fast Enough.
Anthropic's Claude Mythos model discovered more security holes in major software than years of human review had caught. That's exciting for defenders and terrifying for everyone else, because the gap between finding a flaw and fixing it is already dangerously wide.
Key points
- Anthropic's Claude Mythos AI found thousands of high-severity security flaws in major operating systems and browsers in 2026, less than 1% of which had been fully patched at announcement.
- In 2025, security researchers published 46,407 CVEs (Common Vulnerabilities and Exposures, which are officially catalogued software flaws), up 16% from 40,009 in 2024.
- By 2024, attackers were exploiting new flaws on average one day before patches were even publicly available, according to a Mandiant analysis.
- The median time for criminals to exploit a known flaw is now under five days; the average time for organisations to patch a critical flaw exceeds 60 days.
- A medium-severity flaw in Palo Alto Networks firewalls, CVE-2024-9474, was chained with a critical-rated flaw to break into roughly 2,000 devices, illustrating why severity scores alone mislead defenders.
On 7 April 2026, Anthropic announced Project Glasswing and its AI model Claude Mythos. The model did something remarkable: it trawled through major operating systems and web browsers and surfaced thousands of serious security holes, many of which had sat undetected for decades despite human review and automated scanning tools.
Fewer than one in a hundred of those flaws had been fully patched by the time the announcement landed.
Why does an AI finding more bugs make things worse?
More discovered flaws means more work for the people whose job it is to fix software, and that workforce was already drowning before Mythos came along.
Even without superhuman AI assistance, the security community logged 46,407 CVEs in 2025, a 16% jump from the year before. The US National Institute of Standards and Technology processed nearly 42,000 of them, 45% more than any prior year, and still fell behind. Every new flaw Claude Mythos surfaces will eventually reach company security teams as a fresh item on an already unmanageable list.
The patching side of the equation has not kept pace. The gap between attacker speed and defender speed has become brutal.
| Year | Median days from disclosure to first exploit |
|---|---|
| 2018 | 771 days |
| 2021 | 84 days |
| 2023 | 6 days |
| 2024 | Under 5 days (average: -1 day) |
That final row deserves a plain-English read: in 2024, Mandiant researchers found that attackers were, on average, exploiting flaws before patches were publicly released. Criminals with access to private vulnerability research or simply watching the same code repositories as defenders are moving faster than the official disclosure system.
Is the traditional way of deciding what to patch still working?
No, and a real-world example shows exactly why.
Most security teams rank flaws using CVSS scores (the Common Vulnerability Scoring System, a number from 1 to 10 that estimates how dangerous a flaw is in isolation). Patch the 10s first, then the 9s, work down the list. Feels orderly. But in 2025, 28% of the flaws attackers actually exploited carried medium CVSS scores, meaning a quarter of the real attacks came from the middle of the queue that teams were deprioritising.
The Palo Alto Networks incident makes the problem concrete. CVE-2024-9474, scored 6.9 (medium), was combined with CVE-2024-0012, scored 9.3 (critical), to break into roughly 2,000 firewalls. A score-first team would have patched the critical flaw and moved on. Attackers saw one connected attack route and used both.
The analyst community, including opinion coverage in Dark Reading, now argues the fix is to stop treating flaws as independent items on a checklist and start mapping how they connect inside a real network.
What should organisations actually do differently?
Patch strategically, not just urgently. Security teams are increasingly being advised to model attack paths the way a burglar cases a building: not "which window is easiest to break?" but "which route from the window reaches the safe?"
In practice that means asking whether an internet-facing flaw can actually reach a sensitive database, identifying the single points where breaking one link disconnects the most possible attack routes, and patching those choke points first, even when their CVSS score looks unimpressive.
A medium-severity password flaw sitting at the crossroads of every path to a company's core servers matters more than a critical flaw on a machine that cannot talk to anything else on the network.
For ordinary people, the immediate message is less dramatic but still worth keeping in mind: when your employer pushes a software update, that update is increasingly racing a clock measured in hours, not weeks. Apply it.
Common questions
Does this affect the software I use at home?
Potentially yes. When Claude Mythos or similar tools find flaws in widely used operating systems and browsers, patches will eventually reach consumer devices too. Keeping your phone, laptop and apps updated is the most useful thing you can do.
Should I be worried that AI is being used to find flaws in software?
The security community hopes the same AI tools will help defenders fix flaws faster than attackers can exploit them. For now, the evidence suggests both sides are picking up the technology and the defenders are still catching up.
