🛡️ AI–Cybersecurity Weekly – 15 Updates 1️⃣ #GPUHammer: New RowHammer Attack Variant Degrades AI Models on #NVIDIA GPUs – thehackernews.com/2025/07/gp… 2️⃣ #ZipSecurity Raises $13.5M to Deliver AI‑Driven Cybersecurity for Small Businesses – axios.com/2025/07/14/zip-13-… 3️⃣ #Google AI #BigSleep Stops Exploitation of Critical SQLite Vulnerability Before Hackers Act – thehackernews.com/2025/07/go… 4️⃣ #EmpiricalSecurity Raises $12M to Build Custom AI Security Models – siliconangle.com/2025/07/17/… 5️⃣ #CrowdStrike Adds GenAI Security Tools to AWS Marketplace – crowdstrike.com/en-us/press-… 6️⃣ #StrongestLayer Raises $5.2M to Tackle AI‑Powered Email Threats – strongestlayer.com/blog/stro… 7️⃣ @confident_sec Raises $4.2M for AI‑Focused Data Protection Platform – techcrunch.com/2025/07/17/co…‑4‑2m/ 8️⃣ #CERTUA Links #LameHug AI Malware to APT28 Targeting Defense Sector– thehackernews.com/2025/07/ce… 9️⃣ #ISC2: 30% of Cyber Teams Now Use AI Security Tools – isc2.org/Insights/2025/07/IS… 🔟 @get_abstracted and #SentinelOne Partner to Deliver AI‑Driven Security Operations – itsecuritywire.com/quick-byt… 1️⃣1️⃣ #Sysdig Executive Discusses Use of AI in Cloud Threat Response – securitybrief.com.au/story/e… 1️⃣2️⃣ #QuestSoftware Releases AI Tool to Support Active Directory Incident Handling – securitybrief.com.au/story/q… 1️⃣3️⃣ Cloud Security Alliance Releases AI Control Matrix for Governance and Risk – cloudsecurityalliance.org/pr… 1️⃣4️⃣ @Google #Gemini Flaw Enables AI Phishing via Prompt Injection – tomshardware.com/tech-indust… 1️⃣5️⃣ Fake AI Platforms Spread Malware to Web3 Developers – thehackernews.com/2025/07/en… #AIsecurity #LLMsecurity #CyberThreats #AISecurityTools #GPUSecurity #PromptInjection #AIvulnerabilities #SecureByDesign #CloudSecurity #StartupFunding #CyberOps #MalwareDetection #AIcompliance #AIphishing #Web3Security #AISOC #APT28 #ThreatDetection #GenAI #APISecurity #ZeroDay #RedTeaming #CSAMdetection #AImalware #AIgovernance #AdversarialML #CyberDefense #AIrisks #SecurityFunding #AIstartups

How efficient can #CSAMDetection really be ? Neural Hash or not, this is just a Locality Sensitive Hash, and there are bounds on false-negative rates for a given amount of noise. In high dimensions, IIRC, those bounds are not great... @ccanonne_ maybe ?