Today's Cybersecurity Pulse
Anthropic CEO heads to White House to discuss Mythos AI model’s zero‑day capabilities
Anthropic chief Dario Amodei will meet White House chief of staff Susie Wiles to debate government access to Mythos, the company’s frontier AI that can discover and exploit zero‑day vulnerabilities. The discussion follows a Pentagon‑imposed blacklist after Amodei refused to lift safety restrictions, while Treasury, intelligence agencies and CISA are already testing the model.
Also developing:
By the numbers: Artemis raises $70M Series A & seed round
Why Voice Fraud Is CX’s Most Underestimated Problem
AI‑generated voice fraud is exploding, with deep‑fake scams rising 456% year‑over‑year and projected losses of $40 billion by 2027. Europe sees 13 spam calls per subscriber each month and a 15% fraud call rate in the UK, where victims lose an average of $16,600 per incident. The industry’s sole response—labeling calls as “potential spam”—offers no real protection and erodes trust in the phone channel. CX leaders must shift from passive labels to proactive, network‑level interception to safeguard high‑value customers.
CISOs Grapple with AI Demands Within Flat Budgets
The 2026 RH‑ISAC CISO Benchmark reveals that security budgets in large organizations are inching upward, reaching 0.75 % of revenue, while overall IT spend climbs to 3.9 %. More than half of respondents expect modest 1‑10 % budget increases in 2026, but many...
Circle's Judge‑order Rule Blocks $285 M Hack Freeze
Circle had the ability to freeze over $285M in stolen USDC from the Drift hack. They declined — their policy requires a judge’s order first. @tayvano_ explains why that’s the wrong call, and how Tether does it differently. Timestamps: 🚀 0:00 Introduction 🏦 11:21...
Less Reliable Than Fuzzers, Yet Occasionally Luckier
They are exactly like fuzzers except not as reliable. They can get lucky faster though sometimes.
AI‑Powered Counter‑Drone Systems Target Data‑Center Threats to Health Records
Sentradel is marketing AI‑driven counter‑drone systems to data‑center operators after Iranian Shahed drones hit three AWS sites in March 2026. The proposal comes as global data‑center capital spending nears $1 trillion, and the cost of a single kamikaze drone ranges from...
Germany Doxes “UNKN,” Head of RU Ransomware Gangs REvil, GandCrab
German authorities have identified 31‑year‑old Russian Daniil Maksimovich Shchukin as the elusive “UNKN” who led the notorious ransomware groups GandCrab and REvil. Prosecutors say Shchukin and associate Anatoly Kravchuk extorted nearly €2 million (≈$2.2 million) in two dozen attacks, causing over €35 million...
Leadership Gaps Fuel Cloud Security Breaches, Says Analyst Firozan
Faranak Firozan, a cloud‑security analyst, warned that leadership failures, not technology gaps, are the primary driver of recent high‑profile cloud breaches. In a Santa Clara briefing on April 5, 2026, she argued that unclear priorities, fragmented accountability and misaligned incentives...
UK Ransomware Ban Threatens FinTech Firms Amid 88% Cybercrime Surge
Britain’s pending ransomware‑payment ban arrives as cybercrime incidents have surged 88% to 1.46 million, outpacing police staffing growth of 31%. The clash between tighter rules and limited enforcement threatens fintech firms that process payments and manage compliance.
Pasig Mayor Warns Public vs AI Video Using His Face
Pasig City Mayor Vico Sotto used social media to alert residents that an AI‑generated video featuring his likeness is promoting an illegal gambling website. He warned that deepfake videos are increasingly realistic, making it harder for victims to distinguish authentic...
From RAG to Zero‑Trust: Verifying Hostile LLM Memory
“i came in thinking I’d build a RAG system. i left with a zero-trust verification pipeline that treats the LLM’s parametric memory as hostile.” 👀 🔗 https://t.co/nFZBhpfUKq https://t.co/HYoRr2wYhf
LinkedIn Scams: An Ecosystem of Identity Theft and Fraud
LinkedIn scams go beyond fake posts or “I’m grateful” stories; they form an ecosystem of identity abuse, data harvesting, and financial fraud.
![Using A VPN Could Subject You To Government Surveillance, Senators Warn [Roundup]](https://hixhlmpcokxhartfkpyi.supabase.co/storage/v1/render/image/public/images/thumbnails/39247618a8a8ee27b3202d48d0cd6919.webp?width=1200&resize=cover&quality=75)
Using A VPN Could Subject You To Government Surveillance, Senators Warn [Roundup]
U.S. senators have cautioned that using commercial VPNs, many of which are foreign‑owned, could inadvertently waive privacy protections and expose Americans to government surveillance. The Department of Homeland Security notes billions are spent each year on such services, highlighting potential...
Funniest/Most Insightful Comments Of The Week At Techdirt
Techdirt’s weekly roundup highlighted the most insightful and funniest reader comments, ranging from a security analysis of the White House’s new app to sharp legal critiques of the Murthy ruling. The editorial spotlight featured a deep dive into Virginia’s controversial...
Alcatraz Raises $50 Million Series B for AI‑Powered Facial Authentication in Commercial Buildings
Alcatraz, the Cupertino‑based AI security startup, announced a $50 million Series B led by BlackPeak Capital, Cogito Capital and Taiwania Capital. The funding will accelerate its Rock™ facial authentication system, which replaces badges and PINs with privacy‑first AI, across data centers,...
Anthropic’s Claude Code Leak: Why the Instinct to Fire Someone Is the Lazy Response
Anthropic unintentionally released nearly 2,000 Claude Code source files, which were quickly copied and viewed 29 million times online. The company framed the incident as a human‑error packaging issue and confirmed no employee was dismissed. CEO Boris Cherny emphasized that the...
SECURITY AFFAIRS MALWARE NEWSLETTER ROUND 91
The Security Affairs Malware Newsletter Round 91 aggregates the latest high‑impact malware research, spotlighting a new macOS infostealer called Infiniti that leverages ClickFix and Python/Nuitka, and a WebSocket‑based pivoting implant named RoadK1ll. It also details a series of supply‑chain compromises of...
AI Is Making Crypto's Security Problem Even Worse, Ledger CTO Warns
Ledger’s CTO Charles Guillemet warned that artificial intelligence is dramatically reducing the cost and speed of crypto hacks, turning what was once a high‑skill, expensive endeavor into a near‑zero‑cost operation. Over the past year, crypto attacks have siphoned roughly $1.4 billion,...
Descope Leads 2026 CIAM Rankings as Enterprises Prioritize No‑Code Identity
Descope was crowned the best overall Customer Identity and Access Management (CIAM) platform for 2026, with more than 1,000 production customers including GoFundMe and Databricks. Its no‑code workflow builder, adaptive multi‑factor authentication and passwordless options are reshaping how enterprises secure...
Mikko Hyppönen Demonstrates Live Drone Hacking at Black Hat 2025, Flagging New Threat Vector
Mikko Hyppönen, a 35‑year cybersecurity veteran, hijacked a drone onstage at Black Hat 2025 in Las Vegas, turning the conference into a live showcase of emerging aerial attack techniques. The demonstration highlighted the gap between mature device security and the still‑raw field...
AI Health Advice Raises Privacy Risks Without Clear Standards
More people are turning to AI for health advice, but it comes with trade-offs. Uploading medical data can deliver useful insights, yet it also raises serious privacy concerns and risks around how that data is stored and used. The lack of...
Researchers Didn’t Want to Glamorize Cybercrims. So They Roasted Them
Researchers at Trellix launched the Dark Web Roast, a meme‑filled blog that mocks cybercrime groups to counteract industry glamorization. The series lampoons a ransomware crew’s content‑calendar extortion tactics, an exploit developer pricing a Cisco RCE bug at $70,000, and undervalued...
Drift Hack Exposes DeFi Audits Gaps and Possible NK Threat
Another week, another DeFi exploit 🫠 @omeragoldberg joined me to unpack the Drift Protocol hack: ⁉️ What went wrong? 👀 How the attack resembles the Mango DAO and Resolv exploits 🤔 Why was Circle so slow to react? ⚠️Are North Korean state actors behind the...
The Hidden Cost of Hybrid: Data Risk and Compliance Gaps in Financial Services
Hybrid working has become the default model for UK financial services, but it is fragmenting data governance and exposing firms to hidden compliance risks. The spread of personal devices, unsecured networks, and shadow‑IT tools makes it difficult to maintain audit...
Brit Flight Attendant Locked Up in Dubai After Sharing Photo of Airport Drone Attack in Group Chatt
A 25‑year‑old British flight attendant with FlyDubai was arrested and placed in Al Barsha pre‑trial detention for posting a photo of the March 7 Iranian drone strike on Dubai International Airport in a WhatsApp group. He is charged under Dubai’s 2021 cybercrime...
CBP Facility Codes Sure Seem to Have Leaked via Online Flashcards
A public Quizlet flashcard set titled “USBP Review” inadvertently disclosed four‑digit gate codes and internal procedural details for a U.S. Customs and Border Protection (CBP) facility in Kingsville, Texas. The set was taken down after WIRED alerted the platform, and...
Seeking Tools that Scan Repos for Malicious Dependencies
What are vendors that offer scanning of PRs or repos to protect against malicious dependencies? I know of Sonar (Advanced Security), Socket .dev, JFrog. What else do you know of or use and what does it do? (At some point, you want...
Meta Halts AI Training After Data Breach Exposes Secrets
Meta freezes AI data work after breach puts training secrets at risk https://t.co/WtjqyiuUAu via @thenextweb
Hid-Omg-Detect: Linux Driver In Development To Detect Malicious HID Devices
Zubeyr Almaho is spearheading a new Linux kernel driver called hid‑omg‑detect, designed to passively monitor and flag potentially malicious HID peripherals such as compromised keyboards and mice. The driver evaluates low‑level signals like keystroke‑timing entropy, immediate post‑enumeration typing, and suspicious...
User Seeks Help Recovering Hijacked X Account
Hey @nikitabier my friend JT’s X account got taken over. It was @tcmllc I can send over any details via DM Can you help us out?
Senators Warn VPN Use May Invite Government Surveillance
Using A VPN Could Subject You To Government Surveillance, Senators Warn [Roundup] - View from the Wing https://t.co/hJwQPvAsyq
Iranian Drone Strikes Hit AWS Data Centres in UAE and Bahrain, Trigger Regional Outages
Iranian Shahed drones struck two Amazon Web Services (AWS) data centres in the United Arab Emirates on March 1 and damaged a third site in Bahrain on April 1, prompting limited service disruptions across the Gulf. The attacks mark the first deliberate...
Iran’s Charming Kitten Exploits Insiders and Low‑tech Tricks
Iran’s Charming Kitten group relies on deception, insider access, and low-tech methods to steal trade secrets and compromise systems. https://t.co/8jKdiH2bzt
Valid Finding Reveals Overlooked Cookie Injection Requirement
The finding is valid but we need to have a cookie injection on the target or its subdomains but I noticed something the AI didn’t notice…yes AI with humans or spend a lot a lot of tokens.
Russia's VPN Blockade Triggers Nationwide Banking Collapse
Telegram founder Pavel Durov said Russia's Friday effort to block VPNs overloaded the country's traffic filters and knocked out banking apps, leaving cash as the only payment method for hours. The incident highlights how aggressive internet controls can destabilize critical...
UK Tightens Energy Cybersecurity After Poland Attack
UK defining stronger energy cybersecurity rules after Poland attack #energysky -- via pv magazine global: https://t.co/UAgNu4x2di
Hope Hyperliquid Is Conducting Emergency Security Review
I really hope Hyperliquid is in a war room right now assuming they’ve already been compromised and reviewing every last thing they’ve done for the last year and a half …
Post-Quantum Cryptography: Moving From Awareness to Execution
Google’s new whitepaper moves the anticipated quantum‑break date, or “Q‑Day,” to 2029 and urges enterprises to adopt post‑quantum cryptography (PQC). The tech giant highlights that elliptic‑curve encryption could be compromised with fewer qubits than previously thought, and it showcases concrete...
New Attack Exposes Massive Vulnerability—Review Now
I called the attack “chilling” only from what details we knew on Thursday. What actually happened is orders of magnitude more chilling. READ THE WHOLE THING. If you work in this industry, you or your protocol may be a sitting...
Evidence Suggests North Korea May Have Orchestrated Drift Attack
Are North Korean state actors behind the Drift Protocol attack? @omeragoldberg says the markings are there, while sharing what it would take to confirm the speculation 👇 https://t.co/8SgvDYy6Lx
Your Neighbor Just Got a Home Security System, but Should You Be Worried? ‘It’s Inherently a Little Creepy’ Says Surveillance...
Hilary Schneider, CEO of SimpliSafe, says rising privacy concerns are reshaping the home‑security market after a controversial Ring Super Bowl ad and backlash against AI‑powered license‑plate cameras. SimpliSafe differentiates itself by giving customers ownership of video footage, employing mechanical privacy...
Skull Vibrations Could Become Future Biometric Passwords
Vibrations in your skull may be your next password by Rutgers University @TechXplore_com Learn more: https://t.co/oRTnakuTjb #EmergingTech #Innovation #Tech #Technology https://t.co/qvg0gNKuw2
Allowlisting CRLs in Plants: Ongoing, Working Fine
I went through a process to allowlist CRLs in our plants. It's not complete but it's been fine.
The Hack That Exposed Syria’s Sweeping Security Failures
In early March 2026, several Syrian government accounts on X—including the presidency, central bank and ministries—were hijacked, posting pro‑Israel messages and explicit content. The breach was quickly contained, but investigators traced the takeover to shared credentials and a lack of...
Is That Image Actually Malware? Find Out
Image or Malware? Read until the end and answer in comments :) https://t.co/5nD545aoAi #BreakingNews https://t.co/Vvny6JzyBD
Why DDoS Mitigation Fails: 5 Gaps That Testing Reveals
Companies pour money into DDoS mitigation, yet outages persist because the tools are rarely tested under realistic attack conditions. Red Button’s simulations reveal that 68% of faults are severe, with an average DDoS Resilience Score of 3.0—far below the 4.5‑5.0 benchmark....
GStack Receives 14 Security Fixes, Half From Community
14 security bug fixes just landed for GStack, half of which were community PR's. https://t.co/98jmCzQ38i
ESP32-S3 Gets Post-Quantum Encryption with Aethyr Edge Node Open-Source Firmware
Aethyr Research has released open‑source firmware for ESP32‑S3 that adds post‑quantum encryption using ML‑KEM‑768, BLAKE3, and XChaCha20‑Poly1305. The firmware boots in 2.1 seconds and completes a full PQC handshake in 35 ms, with an 833 KB flash footprint and 157 KB free SRAM. It...
Houston, We Have a Protocol.
The episode dives into space cybersecurity, featuring Brandon Bailey of the Aerospace Corporation and Cass (Kaz) Vogel, Blue Origin’s Director of Cybersecurity Governance, Risk, and Compliance. They discuss the evolving threat landscape as humanity expands beyond Earth, emphasizing the need...
Android 17 Upgrades the Boot Chain for the Quantum Age: Google Is Embedding Post-Quantum Cryptography Deep Within the System
Google announced that Android 17 will embed post‑quantum cryptography (PQC) across its core security stack, including Verified Boot, Remote Attestation, and the Android Keystore. The implementation relies on NIST‑approved lattice‑based algorithms such as ML‑DSA‑65 and ML‑DSA‑87, with testing slated for the...
Mercor Cyberattack Forces Meta to Suspend AI Recruiting Partnership
AI recruiting platform Mercor confirmed a security breach linked to the open‑source LiteLLM tool, leading Meta to suspend its partnership. The incident, tied to extortion group Lapsus$, underscores growing cyber‑risk for HRTech firms that rely on third‑party AI infrastructure.