Technology & AI
ACCC welcomes mobile coverage map standard and warns telcos on misleading claims
Telcos must now publish comparable 4G and 5G coverage maps, but broad geographical claims still sit outside the rules.
Technology & AI
35 stories
Technology & AI
Why scattered data breaks AI agents — and why masked modelling beats contrastive learning
A new ICLR 2026 poster finds Masked Image Modeling outperforms contrastive learning when training AI across heterogeneous devices. Here's what changes.
Technology & AI
First unified cyber incident response taxonomy maps 457 sources across 25 years
Researchers have distilled 457 papers into the first unified map of the human, tech and organisational factors that make or break a breach response.
Technology & AI
Controllable AI drivers: how ‘soft gates’ let you steer virtual agents without breaking physics
New CNeVA framework dials per-channel behaviour in traffic sims — no frozen cars, no reward hacking.
Technology & AI
A 90.9% backdoor catch rate: why 'boring' constraints could fix AI coding agents
Old-school engineering constraints let a tiny AI reviewer catch nine in ten backdoors, using fewer tokens than agentic scaffolding.
Technology & AI
This Two-Signal Audit Spots ‘Abliterated’ Open-Weight AI Checkpoints With 95% Accuracy
Researchers can spot stripped safety layers in open-weight AI with 95% accuracy, but the fix has limits.
Technology & AI
Can We Map the Wiring Inside AI Agents? A New Preprint Says Yes
AgentFlow proposes framework-agnostic dependency graphs for auditing agent code, though its 238 flagged risks and superiority claims remain unverified by peer review.
Technology & AI
HaloGuard 1.0 promises elite AI safety in a lightweight open-weights package
The 0.8B-parameter constitutional classifier reportedly outperforms 27B rivals on seven benchmarks, though the paper awaits peer review
Technology & AI
LIB-TRAP: When Standard Cell Libraries Become the Attack Vector
An unreviewed arXiv paper argues that hardware trojans could hide in the building blocks of chip design, shifting supply-chain risk for fabless firms and their customers.
Technology & AI
New Transformer architecture targets extreme events in time-series forecasting
A non-peer-reviewed arXiv preprint introduces Exformer, a three-channel attention mechanism for hydrologic data. The claims are intriguing—and await independent verification.
Technology & AI
EvoVuln and the $50 smart contract audit: too good to be unverified?
An unpeer-reviewed arXiv paper claims procedural knowledge evolution can outperform larger AI models on vulnerability detection with minimal data.