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

J
Java Code Geeks
腾讯CDC
Jina AI
Jina AI
博客园 - 司徒正美
博客园 - 三生石上(FineUI控件)
Apple Machine Learning Research
Apple Machine Learning Research
GbyAI
GbyAI
WordPress大学
WordPress大学
Hugging Face - Blog
Hugging Face - Blog
T
The Blog of Author Tim Ferriss
小众软件
小众软件
M
MIT News - Artificial intelligence
MyScale Blog
MyScale Blog
D
Docker
H
Hackread – Cybersecurity News, Data Breaches, AI and More
Google DeepMind News
Google DeepMind News
月光博客
月光博客
L
LangChain Blog
F
Fortinet All Blogs
Microsoft Azure Blog
Microsoft Azure Blog
博客园 - Franky
C
Check Point Blog
U
Unit 42
人人都是产品经理
人人都是产品经理

Ars Technica

Microsoft issues emergency update for macOS and Linux ASP.NET threat Anthropic tested removing Claude Code from the Pro plan Coyote vs. Acme is finally getting released—with a killer trailer Google unveils two new TPUs designed for the "agentic era" Tabloid reports linking 10 missing and dead scientists spur FBI probe Physicists think they've solved the muon mystery New court ruling blocks many of the government's anti-renewable policies Indian med student rakes in thousands with AI-generated MAGA hottie As EV batteries improve, ChargePoint debuts 600 kW fast charger Our favorite gear at Sea Otter Classic wasn't the bikes—it was the accessories Investors lost billions on Trump’s memecoin. Another gala won’t fix that. Pentagon wants $54B for drones, more than most nations’ military budgets Mozilla: Anthropic's Mythos found 271 security vulnerabilities in Firefox 150 Supreme Court arguments make it clear that FCC fines are "nonbinding" Silo S3 teaser hints at the wasteland's origins Framework's CEO on the RAM crisis and creating a "MacBook Pro for Linux users" Florida probes ChatGPT role in mass shooting. OpenAI says bot "not responsible." Report: Meta will train AI agents by tracking employees' mouse, keyboard use Microsoft removes Call of Duty from Game Pass, lowers subscription pricing Framework Laptop 13 Pro is a major overhaul for the modular, upgradeable laptop Framework Laptop 16 upgrades make it look less like an unfinished prototype Internal emails show how Amazon raises prices across the Internet, lawsuit says Anthropic gets $5B investment from Amazon, will use it to buy Amazon chips CATL's new LFP battery can charge from 10 to 98% in less than 7 minutes AMD Ryzen 9 9950X3D2 Dual Edition review: Tons of cache for tons of dollars What's the deal with spacesuits for the Moon? Will they be ready in time? Loneliness in older adults can often lead to memory impairment Contrary to popular superstition, AES 128 is just fine in a post-quantum world Pentagon pulls the plug on one of the military's most troubled space programs John Ternus will replace Tim Cook as Apple CEO
Mozilla says 271 vulnerabilities found by Mythos have "al...
Dan Goodin · 2026-05-08 · via Ars Technica

The disbelief was palpable when Mozilla’s CTO last month declared that AI-assisted vulnerability detection meant “zero-days are numbered” and “defenders finally have a chance to win, decisively.” After all, it looked like part of an all-too-familiar pattern: Cherry-pick a handful of impressive AI-achieved results, leave out any of the fine print that might paint a more nuanced picture, and let the hype train roll on.

Mindful of the skepticism, Mozilla on Thursday provided a behind-the-scenes look into its use of Anthropic Mythos—an AI model for identifying software vulnerabilities—to ferret out 271 Firefox security flaws over two months. In a post, Mozilla engineers said the finally ready-for-prime-time breakthrough they achieved was primarily the result of two things: (1) improvement in the models themselves and (2) Mozilla’s development of a custom “harness” that supported Mythos as it analyzed Firefox source code.

“Almost no false positives”

The engineers said their earlier brushes with AI-assisted vulnerability detection were fraught with “unwanted slop.” Typically, someone would prompt a model to analyze a block of code. The model would then produce plausible-reading bug reports, and often at unprecedented scales. Invariably, however, when human developers further investigated, they’d find a large percentage of the details had been hallucinated. The humans would then need to invest significant work handling the vulnerability reports the old-fashioned way.

Mozilla’s work with Mythos was different, Mozilla Distinguished Engineer Brian Grinstead said in an interview. The biggest differentiating factor was the use of an agent harness, a piece of code that wraps around an LLM to guide it through a series of specific tasks. For such a harness to be useful, it requires significant resources to customize it to the project-specific semantics, tooling, and processes it will be used for.

Grinstead described the harness his team built as “the code that drives the LLM in order to accomplish a goal. It gives the model instructions (e.g., ‘find a bug in this file’), provides it tools (e.g., allowing it to read/write files and evaluate test cases), then runs it in a loop until completion.” The harness gave Mythos access to the same tools and pipeline that human Mozilla developers use, including the special Firefox build they use for testing.