Technology & AI
New LLM agent memory system MOSAIC hits 89% accuracy
MOSAIC, posted on arXiv on July 21, lifts long-conversation accuracy 27 points and catches 66% of factual conflicts before they corrupt an agent's memory.
Marcello Babbili
AI and technology, explained through what they actually mean. An AI-assisted newsroom under human editorial rules — every story cites its primary sources so you can check them yourself.
499 stories
Technology & AI
One AI controller runs 25 different systems without retraining
A neural network trained on 314,630 demonstrations controls 25 systems from underwater vehicles to chemical reactors, matching hand-tuned controllers in
Technology & AI
NVIDIA Spectrum-6 doubles AI factory network to 102.4 Tbps
NVIDIA's 102.4-terabit Spectrum-6 Ethernet switch is reaching gigascale AI factories first at CoreWeave, Microsoft and Nebius, doubling network capacity.
Technology & AI
True-fact attack hijacks RAG agents 83% of the time
New arXiv preprint shows AI agents can be redirected by reordering true facts, with 83.3% success across GPT, Claude, Gemini, DeepSeek and Qwen.
Technology & AI
23 attack paths target AI agent memory, OS defenses miss some
A new arXiv preprint maps 43 operations that corrupt self-hosted AI agents via legitimate OS calls, finding a residual surface no defense can distinguish.
Technology & AI
LaCache speeds diffusion LLMs 1.3X with training-free caching
LaCache caches unchanged tokens during diffusion LLM denoising, cutting inference 1.3X standalone and up to 40X stacked with other acceleration methods.
Technology & AI
RLHF bias audit: rater mood may skew AI preference data
A July 2026 arXiv paper proposes that human raters' emotional states leak into preference labels used to train AI, with a five-prediction audit plan.
Technology & AI
AI models guess your real intent only 22-32% of the time
New arXiv research finds AI models recover users' intended tasks just 22-32% of the time when instructions are ambiguous, versus 48% for humans.
Technology & AI
Systemic AI risk paper reframes harms as externalities
New arXiv preprint frames systemic AI harms as externalities and collective action problems, naming market concentration and algorithmic monocultures as
Technology & AI
AlphaFold2 weights hide protein folding landscapes, study finds
AlphaFold2's 93 million parameters encode conformational landscapes as a byproduct of static structure prediction, with drug design implications.
Technology & AI
DSWorld cuts AI data science agent training 14x
A new arXiv framework uses an LLM-based simulator to predict code outcomes, cutting agent training 14x and search-based inference 3-6x.
Technology & AI
AI trust gap leaves firms unable to prove safety, arXiv paper says
AI trust gap: new arXiv preprint argues firms can't prove safety claims, leaving buyers and regulators unable to distinguish safe systems from imitations.