RC RANDOM CHAOS

The Wire

Curated cybersecurity and tech news — AI-summarized, source attributed.

aipolicy

Anthropic to Watermark Claude Output Under EU AI Act Transparency Rules

Anthropic has signed the EU AI Act's Article 50(2) Code of Practice on transparency, committing to mark content its models generate. Claude models launched in t

via Hacker News ·
aiopen-source

antirez's h3.c brings native MiniMax-H3 video generation to Apple Silicon

h3.c is a from-scratch C and Metal implementation that runs the MiniMax-H3 generative model directly on Apple Silicon, targeting M3 Max and M5 Max machines. Bey

via Hacker News ·
privacypolicy

British NGOs Push Digital ID Laws Across 21 US States, Investigation Finds

An investigation by the outlet Effort alleges that five UK-based NGOs and their American affiliates are running a coordinated campaign to bring digital ID and a

via Hacker News ·
open-sourcetech-culture

CHICKEN Scheme 6.0 Goes UTF-8 and R7RS, Breaking a Decade of Compatibility

The CHICKEN Scheme compiler, which translates the Lisp dialect to portable C, has shipped its 6.0 release with two structural changes that reshape how the langu

via Hacker News ·
aitech-culture

Claude Nudges a 60-Year-Old Riemann Bound From 41.6% to 67.2%

An unreleased research build of Anthropic's Claude, prompted to 'take a real stab' at the Riemann hypothesis, failed at the headline problem but produced a genu

via Hacker News ·
policytech-culture

Dutch consumer group launches class action over PlayStation Store game prices

A Dutch mass-claim organization (massaschadeconsument.nl) is recruiting PlayStation owners for a collective lawsuit against Sony, arguing the company abuses a c

via Hacker News ·

Encrypted LLM Reasoning Traces Can Be Decoded, Leaking Secrets and Enabling Distillation

Researchers have shown that the encrypted chain-of-thought blocks major providers return to API clients aren't as private as they look. Anthropic, OpenAI, and G

via Hacker News ·
privacypolicy

France Bans Cold-Calling, Flips to Opt-In Consent With Fines Up to €375K Per Call

Starting August 11, 2026, France prohibits unsolicited telemarketing outright, replacing its long-standing opt-out model with a strict opt-in regime. Outbound s

via Hacker News ·
aitech-culture

Humanizing AI output is the wrong layer — render at the boundary, not mid-task

A wave of viral prompt tricks — agent instructions like "talk to me like I have ADHD" or "respond only in Simplified Technical English" — tries to tame the verb

via Hacker News ·
tech-cultureopen-source

Japan's 1954 Parametron: the ferrite-core logic that skipped tubes and transistors

In 1954, University of Tokyo graduate student Eiichi Goto invented the parametron, a logic element built from ferrite cores that stored a bit in the phase of a

via Hacker News ·
tech-cultureprivacy

Mining AIS Vessel Data Turns Up 50,000 Boat Names Full of Puns and Pop Culture

A data journalist pulled roughly 50,000 named vessels from NOAA's Marine Cadastre broadcast-point dataset, which captures Automatic Identification System (AIS)

via Hacker News ·
aiopen-source

Mojo Hits 1.0, Locking In a Stable Foundation for Modular's AI Stack

Modular has shipped Mojo 1.0, roughly two years after the language's 2023 debut. The headline commitment is stability: after a fast-moving development cycle tha

via Hacker News ·