HeadFlash

Daily tech-news flash

AI, privacy and security, caught up in five minutes.

A short email and matching podcast for each niche you pick, every weekday — what moved, why it matters, and where to read the full story. No noise, no hype.

Subscribe free Browse niches
01

AI

  1. CodeRabbit raises $143M as AI code review hits $1.5B valuation CodeRabbit banks $143M at a $1.5B valuation, Deepseek hikes API prices, and OpenAI finds no link between AI use and revenue per employee.
  2. Grok 4.6 Ties OpenAI's Best, Undercuts Rivals on Price Grok 4.6 matches GPT-5.6 Sol on benchmarks at a fraction of the cost, while new research exposes LLM prompt extraction risks.
  3. Anthropic watermarks Claude text; Meta's $2B Manus deal unwinds Claude output gets invisible IDs, Meta splits from Manus under Beijing order, and Nvidia's fast Nemotron 3.5 Lightning lands.
  4. OpenAI gated GPT-5.6-Cyber; Meta's Muse Glimmer runs on consumer GPUs OpenAI restricts its cyber model to approved partners, while Meta ships an open-weight agentic model for a single consumer GPU.
  5. OpenAI Flags Astra as First Model at Critical Cyber Risk Level OpenAI pauses Astra work after internal tests suggest it may hit the highest cybersecurity risk tier, plus AMD buys Taalas and Meta launches Muse Code.
More AI
02

Privacy

  1. Flock Safety Tightens Data Rules After Police Misuse Flock cuts data retention to one week, DHS confirms warrantless ad-auction location buys, and Anthropic watermarks all Claude text.
  2. AI Providers Share One Encryption Key, Researchers Decode 315K Reasoning Blocks A single encryption key across OpenAI, Anthropic, and Google exposed hidden AI reasoning, 182 credentials recovered, and 62 live API keys found.
  3. California DROP Portal Draws 300,000 Deletion Requests California's new DROP platform sees 300,000 requests as 600+ data brokers face mandatory erasure; ALPR data flows to police; AI chatbot age checks criticized.
  4. ICE Pays LexisNexis $6.7M for Data Feeding Palantir Deportation Tools ICE's new procurement reveals AI-driven vetting and facial recognition plans, while California moves to rein in workplace brain data and China tightens data officer rules.
  5. WFP Renews Palantir Deal Despite Leaked Audit Warning WFP extends Palantir contract amid privacy audit concerns; FBI seeks predictive AI watch lists; wearables lack transparency.
More Privacy
03

Security

  1. North Korean hackers hit Windows zero-day; AI agents breach Taiwan Lazarus exploits a fresh Windows flaw, AI agents autonomously hit Taiwan agencies, and more in today's security brief.
  2. LiteLLM supply chain breach exposes 2,488 firms, 153GB of secrets Hudson Rock maps the largest AI supply chain breach to date: LiteLLM compromise hits AWS, Samsung, Cisco and thousands more.
  3. Docker cp Flaw Lets Containers Take Over Hosts; Zoom Zero-Click RCE Hits All Clients Docker patched a container-to-host escape; Zoom fixed a zero-click RCE affecting all platforms. Also: Polish power plant breach, DeadLock's blockchain C2.
  4. OpenAI's GPT-5.6-Cyber finds zero-days in Chrome and mobile OS OpenAI's new offensive-security model finds real Chrome zero-days, while Daybreak tiers open for defenders and researchers.
  5. Zbtlink Router Backdoor Exposes 100,000 Devices to Remote Takeover Hidden backdoor in 20+ Zbtlink router models allows root access to an estimated 100,000 devices worldwide.
More Security