ByteDance unveils Seedance 2.5 in Beijing, saying the AI video model can generate 30-second clips from up to 50 reference materials, up from 12 for Seedance 2.0 (Juro Osawa/The Information)
OpenAI expands 'Daybreak' cyber program: New tools, partnerships, and a cyber-focused GPT-5.5 aim to help 'patch the world' www.itpro.com/security/ope...
"THORChain resumed all network activity after implementing multiple security upgrades and a vault migration to fix the vulnerabilities that led to the $10.7 million exploit."
1.8 Million Credential Attacks Target Business Phone Systems as CloudSEK Records 15 Million SIP Events
Internet-facing business telephone systems are being targeted through sustained and automated attacks designed to steal credentials and generate fraudulent international calls, CloudSEK…
Five Eyes Warn AI Cyber Threat Is Months Away — Western intelligence agencies under the Five Eyes alliance warned that new AI-driven cyber threats are likely within months, not years. The warning comes after the US move to restrict access to Anthropic’s AI models amid concerns about…
OpenAI shipped a telemetry system that logs more than the actual work being done. Codex burning through SSDs at a rate of ~640 TB/year – one user hit 37 TB written in 21 days. On a consumer SSD that’s full drive death in under a year. github.com/openai/codex...
Guest Post: Dark web analysis: Deepfakes as a service discussions soar by 39% Findings from the dark web reveal that discussions surrounding deepfakes as a service have exploded in 2026, already su...
The Global #AppSec USA CFP is OPEN until June 29th! 🎤🔥 Got insights or real-world stories? Take the stage in San Francisco and inspire the AppSec community.
AI-Powered Bug Hunting: The Dangerous Combination That’s Redefining Cybersecurity + Video
Introduction: The cybersecurity landscape is witnessing a paradigm shift where artificial intelligence and human ingenuity are merging to create something far more potent than either could achieve alone. When…
Information Security Briefly@infosecbriefly.bsky.social
Patch the Planet coordinates GPT-5.5-Cyber security research with maintainers to validate vulnerabilities and improve CI/CD, reducing useless AI-generated fixes.
OpenAI is shifting cybersecurity focus from finding vulnerabilities to fixing them, with an updated Codex Security plugin, GPT-5.5-Cyber, and Patch the Planet to speed remediation for open source projects. #OpenAI#CodexSecurity#PatchthePlanet
New York-based crypto analytics startup Allium raised a $40M Series B led by Amplify Partners, with participation from Kleiner Perkins and others (Ben Weiss/Fortune)
Boston Metro Weather Forecast Bot@boston-4castbot.bsky.social
NWS Boston (via Facebook): [Lightning Safety Awareness Week - Day 2] Did you know there are many different kinds of lightning? Learn all about them in the graphic below or by visiting http://nssl.noaa.gov/education/svrwx101/lightning/types/ #WeatherReady [Link]#MaWX
“AI is about to dramatically accelerate the speed, scale, and sophistication of cyberattacks, lowering barriers for adversaries and giving them capabilities that were once limited to highly skilled actors,”
Hacking enabled by subsidised AI data centres 🤦♂️
780th Military Intelligence Brigade (Cyber)@780thmibdecyber.bsky.social
Lost in relocation: analysis of a new loader distributing CASTLESTEALER
Elastic Security
"CIS-region and Russian-language exclusions point to a financially motivated, Russian-speaking threat actor." www.elastic.co/security-lab...
AryStinger Botnet Turns 4,300+ Legacy Routers Into Stealth Reconnaissance Proxies—Are You Next? + Video
Introduction: A previously undocumented malware botnet named AryStinger has silently compromised more than 4,300 end-of-life routers worldwide, transforming them into a distributed…
Treating all opposition to data centers as a national security threat is the wrong move, says Jordyn Abrams, who argues policymakers must address legitimate public grievances about AI—not just police them.
How to Hack Your Way Into a Cybersecurity Career Without Breaking the Bank – KrazePlanet’s New Student Offers Make Bug Bounty Training Accessible for Everyone + Video
Introduction: The cybersecurity industry faces a critical talent shortage, yet many aspiring ethical hackers and penetration…
Lately, it feels harder to rely on online services. Between constant tracking, rising subscription costs, and growing privacy concerns, I find myself using my Raspberry Pi increasingly to bring some services back under my control. Several... #raspberrypi #linux
Some people ask: "If I have DKIM, do I need DMARC?" The short answer is yes. In our latest article, we'll cover exactly what DKIM does, where it's protection ends and why SPF, DKIM, and DMARC must work together: dmarcian.com/why-dkim-onl...#DKIM#DMARC#dmarcian
Anthropic has confirmed a major, system-wide artificial intelligence tools outage
Anthropic has confirmed a major, system-wide outage causing elevated error rates and widespread disruptions across its entire suite of artificial intelligence tools. The incident, which began around 10:02 AM ET…
Transform your .NET apps with OpenAI's Model Context Protocol (MCP)! Seamlessly integrate #AI agents for enhanced performance. Discover how MCP can streamline your projects. #OpenAI#DotNet
AI has shrunk exploit development from months to hours, making patch-first defense too slow. TTP-chain validation and control-aware testing can prove real exposure even when no exploit exists. #PicusSecurity#OpenBSD#CVE202529824
AI-Powered Offensive Security: The True Cost of Building vs Buying an Autonomous Penetration Testing Platform + Video
Introduction: The cybersecurity industry is witnessing a paradigm shift as frontier AI models like GPT-5.5 and Mythos demonstrate unprecedented capabilities in vulnerability…
Developers are increasingly using agent loops to push AI models through longer, messier tasks. The approach can cost far more, but Anthropic engineers showed it can produce much better results.
Cyberattack disrupts Cedro Têxtil in Sete Lagoas, Brazil, halting factory operations since June 18 as IT and business systems are restored with expert support. #Brazil#CedroTêxtil #SeteLagoas
Market intelligence company Klue confirms it has suffered a breach, for which cybercrime group Icarus takes credit; Jamf, HackerOne, and others are affected (Zack Whittaker/TechCrunch)
25 Must-Watch YouTube Channels for Free OT/ICS Cybersecurity Training in 2026 + Video
Introduction: Operational Technology (OT) and Industrial Control Systems (ICS) form the backbone of critical infrastructure—from power grids and water treatment facilities to manufacturing plants and food…
Beehiiv adds Cloudflare AI Crawl Control to writers' dashboards, allowing them to decide whether AI crawlers can scrape their work (Duncan Riley/SiliconANGLE)
Today’s free newsletter is about how Silicon Valley — as exemplified by the AI bubble — has abandoned its supposed principles of rationality and meritocracy, and now more closely resembles a cargo cult desperate to recreate bygone eras of growth.
Been bad at sharing stuff I've written here. From two weeks ago: Worked with our threat hunters on this piece around how to mature a threat hunting program—when to scale, friction points, how to build a continuous feedback loop, etc—over at @redcanaryco.bsky.social: redcanary.com/blog/threat-...
Shortly after OpenMDW 1.1 dropped, NVIDIA AI announced adoption of the open source AI licence for its major model families. Learn more about how this easy-to-apply license enables openness, consistency, and clarity across all components of an AI model distribution.
Anthropic may ask Claude users to verify their age and identity by uploading their government-issued documents, according to a new version of the company’s privacy policy.
Data leaks, ransom notes, identity hacks—all 2026's headline crime. 🤯 Closed‑community apps and new age‑verify laws are making us show ID, while malicious actors keep finding ways to steal it. #Privacy #Cybersecurity
In a new episode of the Cyberwire podcast, “Vulnerability response. Built for humans, outpaced by machines,” XBOW Head of Security Labs Federico Kirschbaum shares his perspective on vulnerability management when frontier models are finding the flaws.
The Silent Guardian: Hardening Industrial Control Systems Against the Next Generation of Cyber-Physical Attacks + Video
Introduction: In the high-stakes world of Oil & Gas, the Distributed Control System (DCS) is the brain of the operation, orchestrating complex processes from refining to pipeline…
by @lwinner.bsky.social 1/2 “the ban was never about an AI jailbreak. It was about a fight. Anthropic had refused to let the Pentagon use its models for fully autonomous weapons—machines that decide on their own whom to kill. The company drew a moral line, and the Pentagon responded..
I can’t wait for the day when the US government has to formally bail out AI companies with trillions of dollars but then say they don’t have the money for healthcare, childcare, and student debt forgiveness…
What we learned at Europe’s largest cybersecurity conference What did the Proton team learn at Infosecurity Europe 2026? Find out our key observations and takeaways from the conference Infosecuri...
