正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

MERCOR 4TB BREACHED, ANTHROPIC SOURCE CODE EXPOSED, & AXIOS NPM PACKAGE SUPPLY CHAIN ATTACK "LLMs can find and exploit zero day vulnerabilities in critical software." - Nicholas Carlini, Anthropic He sounded the alarm last week...

This Week in AI

17,124 subscribers

19,491 次观看 • 7 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

4 条评论

Podreview 的头像

Podreview7 个月前

Carlini's been saying this for months and nobody listened. The Mercor breach is wild — 4TB of candidate data just sitting there. But the axios supply chain attack should scare people more. We trust npm packages the same way we trusted email attachments in 2005.

Hermetic 的头像

Hermetic7 个月前

the axios attack is the one that should scare every developer. malicious npm package reads os.environ and exfiltrates your keys. every ai agent running today has those same keys in memory. the fix isnt better detection — its not giving the agent the keys in the first place-

Andreas B 的头像

Andreas B7 个月前

The dual-use dilemma in AI security isn't theoretical anymore. Who's building the defenses at this pace?

Oracle 的头像

Oracle7 个月前

this one gives better result:

相关视频

Last week, Anthropic said its new model Mythos had already found “thousands of high-severity vulnerabilities” and could “surpass all but the most skilled humans at finding and exploiting software vulnerabilities.” At almost the same moment, the Axios hack showed how exposed the open-source foundation under modern software still is. For this week’s 10 Minutes or Less, I spoke with Feross what this moment means. WATCH:

Last week, Anthropic said its new model Mythos had already found “thousands of high-severity vulnerabilities” and could “surpass all but the most skilled humans at finding and exploiting software vulnerabilities.” At almost the same moment, the Axios hack showed how exposed the open-source foundation under modern software still is. For this week’s 10 Minutes or Less, I spoke with Feross what this moment means. WATCH:

Ali Rohde

13,162 次观看 • 6 个月前

URGENT PSA - New supply chain attack vector that I found WILD > AI LLMs hallucinate package names roughly 18-21% of the time. Hackers have started pre-registering those hallucinated names on PyPI and npm with malicious payloads; they call it "slopsquatting" You can only imagine what's next

URGENT PSA - New supply chain attack vector that I found WILD > AI LLMs hallucinate package names roughly 18-21% of the time. Hackers have started pre-registering those hallucinated names on PyPI and npm with malicious payloads; they call it "slopsquatting" You can only imagine what's next

Basel Ismail

576,185 次观看 • 7 个月前

holy shit Anthropic built "Cowork" in the last... week and a half?! Confirmed by Anthropic employee Felix Rieseberg on Dan Shipper's live stream I wonder how much of this was Claude Code

holy shit Anthropic built "Cowork" in the last... week and a half?! Confirmed by Anthropic employee Felix Rieseberg on Dan Shipper's live stream I wonder how much of this was Claude Code

Alex Volkov

418,841 次观看 • 9 个月前

Axios’ Marc Caputo: Do you view Anthropic, and to a degree its CEO Dario Amodei, as a threat to national security? Trump: Well, not now, but a week ago maybe. Source: Axios

Axios’ Marc Caputo: Do you view Anthropic, and to a degree its CEO Dario Amodei, as a threat to national security? Trump: Well, not now, but a week ago maybe. Source: Axios

Clash Report

18,522 次观看 • 4 个月前

Anthropic top researcher Nicolas Carlini (67.2k citations on Google Scholar) says Claude is a better security researcher than him, made $3.7 mn from exploiting smart contracts, and found vulnerabilities in Ghost (a 52K+ Github star project).

Anthropic top researcher Nicolas Carlini (67.2k citations on Google Scholar) says Claude is a better security researcher than him, made $3.7 mn from exploiting smart contracts, and found vulnerabilities in Ghost (a 52K+ Github star project).

Rohan Paul

363,209 次观看 • 7 个月前

someone at ANTHROPIC just showed CLAUDE finding ZERO DAY vulnerabilities in a live conference demo claude has found zero day in Ghost, 50,000 stars on github, never had a critical security vulnerability in its entire, history... it found the blind SQL injection in 90 minutes, stole the admin api key, then did the exact, same thing to the linux kernel

someone at ANTHROPIC just showed CLAUDE finding ZERO DAY vulnerabilities in a live conference demo claude has found zero day in Ghost, 50,000 stars on github, never had a critical security vulnerability in its entire, history... it found the blind SQL injection in 90 minutes, stole the admin api key, then did the exact, same thing to the linux kernel

chiefofautism

1,907,651 次观看 • 7 个月前

Anthropic's new Mythos AI model has become a flashpoint in the debate over how artificial intelligence could transform cybersecurity, after the company said the unreleased system can find and exploit serious software vulnerabilities

Anthropic's new Mythos AI model has become a flashpoint in the debate over how artificial intelligence could transform cybersecurity, after the company said the unreleased system can find and exploit serious software vulnerabilities

Reuters

51,293 次观看 • 6 个月前

When the attackers have AI, how do you prevent the next HuggingFace incident from happening to your product? Today, we launch VEGA with Y Combinator , a beyond Mythos level cybersecurity agent that finds critical vulnerabilities before your product ships. VEGA reviews your source code, identifies vulnerabilities before release, monitors infrastructure and supply-chain risks, investigates attack paths, and fixes issues before attackers find them. AI makes you fast. VEGA makes you secure. Learn more at

When the attackers have AI, how do you prevent the next HuggingFace incident from happening to your product? Today, we launch VEGA with Y Combinator , a beyond Mythos level cybersecurity agent that finds critical vulnerabilities before your product ships. VEGA reviews your source code, identifies vulnerabilities before release, monitors infrastructure and supply-chain risks, investigates attack paths, and fixes issues before attackers find them. AI makes you fast. VEGA makes you secure. Learn more at

NebuSec

36,299 次观看 • 3 个月前

Nicolas Carlini showed Claude can find 0-DAY vulnerabilities, in a live conference demo. it found Autonomously finding a blind SQL injection zero-day (CVE-2026-26980) in Ghost CMS, extracting stole admin API keys without authentication in roughly 90 minutes, marking the platform's first critical vulnerability after 20 years and 50,000+ GitHub stars. Then did the exact same thing to the Linux kernel, it few months older video

Nicolas Carlini showed Claude can find 0-DAY vulnerabilities, in a live conference demo. it found Autonomously finding a blind SQL injection zero-day (CVE-2026-26980) in Ghost CMS, extracting stole admin API keys without authentication in roughly 90 minutes, marking the platform's first critical vulnerability after 20 years and 50,000+ GitHub stars. Then did the exact same thing to the Linux kernel, it few months older video

Md Ismail Šojal 🕷️

15,731 次观看 • 4 个月前

Software vulnerabilities can be notoriously time-consuming for developers to find and fix. Today, we’re sharing details about CodeMender: our new AI agent that uses Gemini Deep Think to automatically patch critical software vulnerabilities. 🧵

Software vulnerabilities can be notoriously time-consuming for developers to find and fix. Today, we’re sharing details about CodeMender: our new AI agent that uses Gemini Deep Think to automatically patch critical software vulnerabilities. 🧵

Google DeepMind

368,175 次观看 • 1 年前

AI giant Anthropic says its new model, Claude Mythos Preview, is so good at coding that it could pose a cybersecurity threat in the wrong hands. The company says it's already found thousands of software vulnerabilities. Cyberattacks can impact critical infrastructure— from healthcare systems to government agencies. The announcement of the model been met with a mix of alarm and optimism— and sparked skepticism in some who question the its full capabilities. CBS News’ Lauren Fichten explains.

AI giant Anthropic says its new model, Claude Mythos Preview, is so good at coding that it could pose a cybersecurity threat in the wrong hands. The company says it's already found thousands of software vulnerabilities. Cyberattacks can impact critical infrastructure— from healthcare systems to government agencies. The announcement of the model been met with a mix of alarm and optimism— and sparked skepticism in some who question the its full capabilities. CBS News’ Lauren Fichten explains.

CBS News

22,643 次观看 • 6 个月前

⚠️ NEW on The Axios Show: President Trump tells Marc Caputo that a week ago, he might've considered Anthropic a national security threat when his administration restricted access to its new model. "People get put in prison immediately for that."

⚠️ NEW on The Axios Show: President Trump tells Marc Caputo that a week ago, he might've considered Anthropic a national security threat when his administration restricted access to its new model. "People get put in prison immediately for that."

Axios

121,926 次观看 • 4 个月前

This is big... Anthropic just announced a model so powerful they won't release it to the public out of fear over the damage it will cause 😨 Claude Mythos Preview found thousands of zero-day exploits in every major operating system and web browser... The numbers are hard to believe: > $50 to find a 27-year-old bug in OpenBSD, one of the most security-hardened operating systems ever built > Under $1,000 to find AND build a fully working remote code execution exploit on FreeBSD that grants unauthenticated root access from anywhere on the internet > Under $2,000 to chain together multiple Linux kernel vulnerabilities into a complete privilege escalation exploit For context: these are the kinds of findings that previously required elite security researchers working for weeks. Anthropic engineers with no formal security training asked Mythos to find exploits overnight. They woke up to working code the next morning. The results were so impressive Anthropic assembled Apple, Google, Microsoft, Amazon, NVIDIA, and seven other organizations into Project Glasswing: A $100M defensive coalition. They're not releasing this model publicly. Instead, they're racing to patch the world's infrastructure before models like this proliferate.

This is big... Anthropic just announced a model so powerful they won't release it to the public out of fear over the damage it will cause 😨 Claude Mythos Preview found thousands of zero-day exploits in every major operating system and web browser... The numbers are hard to believe: > $50 to find a 27-year-old bug in OpenBSD, one of the most security-hardened operating systems ever built > Under $1,000 to find AND build a fully working remote code execution exploit on FreeBSD that grants unauthenticated root access from anywhere on the internet > Under $2,000 to chain together multiple Linux kernel vulnerabilities into a complete privilege escalation exploit For context: these are the kinds of findings that previously required elite security researchers working for weeks. Anthropic engineers with no formal security training asked Mythos to find exploits overnight. They woke up to working code the next morning. The results were so impressive Anthropic assembled Apple, Google, Microsoft, Amazon, NVIDIA, and seven other organizations into Project Glasswing: A $100M defensive coalition. They're not releasing this model publicly. Instead, they're racing to patch the world's infrastructure before models like this proliferate.

Josh Kale

4,068,888 次观看 • 7 个月前

Foreign Nationals In Election Source Code EXPOSED! A top tech figure admits foreign nationals accessed U.S. election source code. DOJ filings against Smartmatic cite supply chain concerns and overseas components—raising urgent questions about security and foreign interference. Full interview Alex Jones

Foreign Nationals In Election Source Code EXPOSED! A top tech figure admits foreign nationals accessed U.S. election source code. DOJ filings against Smartmatic cite supply chain concerns and overseas components—raising urgent questions about security and foreign interference. Full interview Alex Jones

Patrick Byrne

50,365 次观看 • 8 个月前

A researcher opened an npm package. What he found broke the internet. On March 31st 2026, Chaofan Shou noticed something nobody was supposed to see inside Anthropic's Claude Code package — the entire original TypeScript source code. Publicly accessible. Just sitting there. That post hit 3.1 million views. Within hours, someone archived it to GitHub. 1,100+ stars. 1,900+ forks. Before Anthropic could even respond. And the truly ironic part? Buried inside that leaked code was a system called Undercover Mode — literally built by Anthropic to PREVENT internal secrets from leaking. They built a leak-prevention system. And leaked it. ✅Save this video, you'll be referencing this story for years. Want the SOP? DM me.

A researcher opened an npm package. What he found broke the internet. On March 31st 2026, Chaofan Shou noticed something nobody was supposed to see inside Anthropic's Claude Code package — the entire original TypeScript source code. Publicly accessible. Just sitting there. That post hit 3.1 million views. Within hours, someone archived it to GitHub. 1,100+ stars. 1,900+ forks. Before Anthropic could even respond. And the truly ironic part? Buried inside that leaked code was a system called Undercover Mode — literally built by Anthropic to PREVENT internal secrets from leaking. They built a leak-prevention system. And leaked it. ✅Save this video, you'll be referencing this story for years. Want the SOP? DM me.

Julian Goldie SEO

34,350 次观看 • 7 个月前

Anthropic launched Claude Code Security in limited research preview, scanning codebases for vulnerabilities and suggesting targeted fixes. Cybersecurity stocks like $PANW, $FTNT, $ZS, $RBRK & $CRWD are lower on fears AI-native security tools could pressure parts of the stack.

Anthropic launched Claude Code Security in limited research preview, scanning codebases for vulnerabilities and suggesting targeted fixes. Cybersecurity stocks like $PANW, $FTNT, $ZS, $RBRK & $CRWD are lower on fears AI-native security tools could pressure parts of the stack.

Shay Boloor

180,909 次观看 • 8 个月前

BREAKING Palantir Dr. Karp Anthropic 🚨 Dr. Karp CEO of $PLTR talking about Anthropic issue and future LLM vendor. Video source & credit: CNBC

BREAKING Palantir Dr. Karp Anthropic 🚨 Dr. Karp CEO of $PLTR talking about Anthropic issue and future LLM vendor. Video source & credit: CNBC

Mike

39,853 次观看 • 7 个月前

Anthropic engineer Arnaud Doko: "Saying 'make it better' to Claude Code is the most expensive mistake anyone can make." In 31 minutes, he walks through the exact prompt patterns, planning workflow, and verification setup Anthropic uses in-house. Watch the full talk, then save the config below👇

Anthropic engineer Arnaud Doko: "Saying 'make it better' to Claude Code is the most expensive mistake anyone can make." In 31 minutes, he walks through the exact prompt patterns, planning workflow, and verification setup Anthropic uses in-house. Watch the full talk, then save the config below👇

rody

59,480 次观看 • 5 个月前

Just updated talk to figma mcp open source Anthropic Claude Code can now talk to figma I can ask claude to review my design and annotate directly in design file Link below

Just updated talk to figma mcp open source Anthropic Claude Code can now talk to figma I can ask claude to review my design and annotate directly in design file Link below

Sonny

32,790 次观看 • 8 个月前