
📡 Your radar for the AI era. Daily news on AI, cybersecurity, privacy, SIAM and agile, plus the certifications that grow your career. #EXINRadar
Utrecht, the Netherlands
Joined July 2009
- Tweets3.2K
- Following3.5K
- Followers3.1K
- Likes614
Italy’s data protection authority, the Garante, has fined Meta €805m, saying the Facebook app for Android accessed photos, videos and their metadata on Italian users’ phones without specific consent or proper notice.
thenextweb.com/news/italy-fi…
1
8
The number of ransomware attacks Recorded Future has seen this year already dwarfs the total for 2025, according to longtime ransomware researcher Allan Liska.
wired.com/story/beware-the-n…
1
1
23
Anthropic says many users hold back Claude Opus 5.5 with outdated prompting habits. In its tests, responses without phrases like “think hard” started faster, with no measurable drop in quality.
Users should remove these cues from prompts.
the-decoder.com/anthropic-sa…
19
Microsoft's November 2026 patches fix CVE-2026-124194, an exploited Outlook spoofing flaw.
Prioritize the Outlook fix: crafted message files can leak users' NTLMv2 hashes without an external-content warning.
computerworld.com/article/42…
110
NATO withdrew guidance to use Yandex Go, a Russian-owned taxi app, at its defense industry conference in Turkey.
Attendees should review transport choices amid concerns the app could expose movements to Russian intelligence.
therecord.media/nato-issues-…
70
Malaysia is probing whether Speedmatrix, cloud firm Megaspeed’s local unit, re-exported AI chips.
Firms re-exporting US-origin high-performance AI chips from Malaysia need trade permits under rules dating to July 2025.
thenextweb.com/news/malaysia…
60
OpenAI’s Daybreak pairs early-access AI web app penetration testing with proposed vulnerability fixes.
Engineers still need to review and test patches before merging.
helpnetsecurity.com/2026/11/…
38
ControlPlane researchers found the UK's Planning Inspectorate left Microsoft Entra ID credentials on GitHub for 18 months.
Teams should check public code for keys granting tenant-wide Microsoft 365 access.
itpro.com/business/public-se…
60
Trump’s AI advisers discussed AI cyberattack reporting for critical US sectors on Thursday, per Reuters.
The Cybersecurity Information Sharing Act of 2015 expires next month; review protections for sharing threat data.
thenextweb.com/news/trump-ai…
1
61
Mistral published a technical report on Mistral Large 4 (ML4), its coding-focused AI model.
IT teams sizing in-house deployments should account for a 606GB NVFP4 checkpoint running on four GB300 GPUs.
computerweekly.com/news/3666…
1
90
Ars Technica reports drone and missile strikes on data centers permanently deleted some Amazon Web Services customer data this year.
Cloud risk reviews need to consider permanent data loss from physical attacks.
arstechnica.com/gadgets/2026…
1
68
IBM released experimental Granite 4.1 3B and 8B previews with LLMON, a markup language separating instructions from data.
Teams can test static prompt-injection resistance; the previews are not for production.
huggingface.co/blog/ibm-gran…
40
Ireland’s privacy watchdog fined Tinder owner Match Group €11mn for incomplete data downloads, Politico reports.
Privacy teams should check whether access tools omit data categories or replace them with summaries.
thenextweb.com/news/tinder-m…
46
Dubai Electronic Security Center and ITU, the UN digital technologies agency, signed a deal to model major digital disruptions.
Cable and satellite disruptions are in scope, giving continuity teams risks to plan for.
gulfnews.com/technology/duba…
28
DOE is developing AI-powered tools through AI-FORTS, an energy cybersecurity program launched in 2024.
industrialcyber.co/utilities…
1
29
Microsoft reported only “limited recovery” from an Outlook email and Teams outage; UAE users were still affected Friday.
IT teams should plan for services to return gradually as users’ apps reconnect.
gulfnews.com/technology/micr…
1
86
The UK's NCSC launched Cyber Adversary Simulation, linking critical infrastructure and public sector organizations to approved red-teaming providers.
Teams need to define business impacts to avoid before testing.
itpro.com/security/ncsc-puts…
24
Microsoft researchers poisoned about 1% of a coding-agent training corpus to trigger unsafe commands from routine requests.
Teams should review training tool calls for actions unrelated to user requests.
microsoft.com/en-us/security…
1
1
57
SANS Institute’s 2026 State of ICS/OT Security Survey found only 15% of respondents could sustain safe manual operations beyond 48 hours.
OT teams should test production without centralized SCADA or DCS control.
industrialcyber.co/threats-a…
35
Research firm Challenger, Gray & Christmas attributes nearly all 300,252 October tech hiring plans to Amazon's seasonal jobs.
HR teams should not treat this largely warehouse hiring as permanent IT demand.
computerworld.com/article/42…
45