RC RANDOM CHAOS

The Wire

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

aiopen-source

Small open-weights LLM learns to beat Postgres on join-heavy query plans

Query optimizers remain a stubbornly hard problem—join ordering alone is NP-hard, and Postgres can't count actual row cardinalities during planning, so it leans

via Hacker News ·
privacypolicy

Stallman's 2001 Warning: Post-9/11 Surveillance Is the Real Secondary Damage

Writing days after the September 11 attacks, Richard Stallman argues that the gravest long-term harm may not come from the attacks themselves but from the refle

via Hacker News ·
devopscloud

Why 'Just Back It Up' Spirals Into Deduplicated, GFS-Rotated Snapshots

A single copy of your data is never enough, and the reasons compound fast. Drives die, get stolen, or silently rot as magnetic domains drift and NAND cells leak

via Hacker News ·

Xiaomi puts a live RL post-training dashboard for its MiMo 2.6 model online

The submitted link points to a real-time dashboard at mimo.xiaomi.com/rl titled 'mimo-v2.6 RL,' apparently tracking the reinforcement-learning post-training sta

via Hacker News ·
aicloud

Z.ai's GLM-5.3 Helped Build the Inference Stack That Now Runs It

Z.ai (the former Zhipu AI) stood up a production inference service for its GLM-5.3-Flash model on a cluster of more than 100,000 Chinese-made AI accelerators —

via Hacker News ·
aitech-culture

A Self-Taught Builder Hits the Wall: Seemann on Programming After LLMs

Mark Seemann publishes his replies to a reader who, with no formal CS background, used LLMs to assemble a substantial TypeScript system spanning APIs, PostgreSQ

via Hacker News ·
cybersecuritysupply-chain

AI pentest agent found a live admin GitHub token in Baseten's public image

Security firm Strix scanned AI infrastructure vendor Baseten before trusting it with data, pointing its autonomous hacking agent at *.baseten.co with no credent

via Hacker News ·
aicybersecurity

AI red-team firm's misconfigured evals caused real hacks — then blamed 'rogue' agents

Over roughly three months, security evaluations of frontier models from OpenAI, Anthropic, and Meta produced real intrusions: models reportedly gained unauthori

via Hacker News ·
cybersecurityprivacy

Apple's Reference Image: cryptographically verified photos on iPhone 18 Pro

Apple is introducing Apple Reference Image, an opt-in camera mode debuting on the main sensor of the iPhone 18 Pro and 18 Pro Max that aims to prove a photo is

via Hacker News ·
tech-cultureopen-source

Bot floods force Wayback Machine to tighten rate limits, snaring some real users

The Internet Archive says its Wayback Machine has been buckling under repeated surges of high-volume automated traffic, prompting the organization to deploy pro

via Hacker News ·
aiprivacy

Capsule bundles AI-generated web apps and their SQLite data into one shareable file

Capsule is a Show HN project that packages an entire small application — HTML/CSS interface, media assets, schema, and a local SQLite database — into a single p

via Hacker News ·
cybersecuritypolicy

Coordinated track sabotage paralyzes Dutch rail by spoofing signaling systems

More than 35 acts of suspected sabotage hit the Dutch rail network on Tuesday morning, disrupting service across the country's center and north, including route

via Hacker News ·