Tech · Culture · Fiction
Golden Gate's crashes cluster in exploit-grade code
macOS Golden Gate's reported instability sits in the same subsystems - WebKit, ImageIO, IOKit, XNU - that carry every documented macOS exploitation chain.
Anthropic, OpenAI, and DeepMind grade their own models' danger
AI labs publishing dangerous-capability evals turns safety disclosure into marketing, and the self-graded scorecards drift toward a race to the bottom.
Unsealed court briefs turned training data into evidence
Unsealed briefs in the authors' case against Microsoft and OpenAI reveal training-data provenance as a board-level liability surface, not a technical detail.
Elastic compute won't make your AI cheaper
DeepSeek Elastic Compute (DSec) cuts AI inference cost only when traffic is spiky and the scaling policy is tuned - how to measure peak-to-average and size it.
It walks through the handshake between your apps
Ember-1 exploits trust relationships between applications, moving past perimeter and identity controls. What failed, why, and what must now be true.
Loading a trusted model runs untrusted code.
from_pretrained resolves a Hugging Face model name to whatever the branch holds now, extending a one-time trust decision over content it never re-checks.
Loading a model executes the uploader's code
How autonomous AI agents weaponize Hugging Face via pickle deserialization and trust_remote_code, mapped to MITRE ATT&CK and ATLAS with the telemetry defenders see.
The Wire — latest
All →- 101 npm packages secretly add devs' WhatsApp to groups
- 543,000 Valid Credentials Still Exposed on GitHub Despite Safeguards
- 6 Dangerous Browser-Based Attack Techniques in 2026
- AI and the timeless fear of technological replacement
- AI Boosts SOC Worker Satisfaction but Bars Entry for Many
- AI Coding Agents Leak 13,000 Internal Company Images on GitHub
- AI Coworkers Break Current Security Models
- AI-Driven Attack Exploits Zammad Zero-Days to Breach Network
- AI 'Hacking' Misreported as Rogue Behavior
- AI Models Achieve Cyber Capabilities Breakthrough
Stay in the loop
New writing delivered when it's ready. No schedule, no spam.