November 2026
86 stories so far · page 3 of 5Tuesday 17 November
15:06Language modelsIIIOpen source
A team at MBZUAI has published six language models for agent work, from phone-sized to server-sized, together with the data, code and checkpoints used to build them. The largest is competitive with open rivals on tool use, but trails closed models and has weak defences against hidden instructions.
14:31Safety & evaluationIIIOpen source
A preprint finds that two of four recent open base models write out benchmark test items, answers included, from memory. Zeroing those items drops one model from first to last.
14:21Safety & evaluationIIBenchmarked
When a tool fails, AI agents often switch to a backup that causes damage, even after reading the facts that should stop them. A preprint from the Hong Kong University of Science and Technology tests a cheap check that pauses the agent once and asks it to look again.
13:33Language modelsIIIOpen source
NVIDIA has retrained Google's small Gemma 4 model to follow instructions and use tools better, while keeping replies short. It fits in about 4 GB of memory and comes with a helper that speeds up replies.
Monday 16 November
18:31AgentsIIIOpen source
A Princeton and Snowflake team reports an AI agent setup that broke none of its written rules across 207 benchmark tasks. It also finished more tasks than four existing guardrail methods. The work is a preprint and has not been peer reviewed.
18:29Safety & evaluationIIIOpen source
A new preprint finds that when a strong AI agent hands file reading to a weaker helper, a planted fake record gets through far more often. The system ends up about as easy to fool as its weakest member.
18:27Safety & evaluationILab only
A peer-reviewed paper argues that copyright treats an author as someone who can answer for a work. On that reading, current AI systems fall short, but people who steer AI closely may still qualify.
18:26Vision & multimodalIIBenchmarked
A small machine learning model learns weather forecasting and data blending straight from 18 kinds of instrument. Near the ground it beats the main US ensemble for up to 12 hours, but it trails operational systems higher up.
18:25Safety & evaluationIIIOpen source
A preprint finds that teams of the same Claude model make less safe choices than one agent working alone. Naming a safety lead and writing down the safety rules closed most of the gap.
18:24AgentsIIIOpen source
A preprint reports that Claude Code, using only free software, wrote factory-scheduling programs. On the same computer, these came closer to the best possible result than the paid solvers CP Optimizer and Hexaly. The advantage held on large, regular test problems, but not on every problem type.
18:23AgentsIIIOpen source
A preprint tests dozens of AI agents on cancellations and deletions. Whether a search tool lists a second matching record, or only counts it, decides whether most of them stop.
17:35CodeIIIOpen source
A new open leaderboard, PerfArena, times the correct code of 13 language models in seven programming languages. Passing the tests turns out to say little about how fast or energy-hungry the code is.
16:15RoboticsIIIOpen source
Researchers trained a 27-billion-parameter open-weight reasoning model to control robot arms using only simulated practice. On a real two-armed robot it had never seen, it succeeded in 18 of 30 trials. The work is a preprint and has not been peer reviewed.
13:51EfficiencyIIIOpen source
A 7B model running on one graphics card answered about half of the questions sent to a costly AI agent. In the authors' tests this cut list-price spending by a fifth to a third, with no measurable drop in accuracy. The work is a preprint and has not been peer reviewed.
10:22EfficiencyIIBenchmarked
A preprint adapts off-the-shelf speech recognition models to turn recorded brain activity into text on an iPhone. It uses a small fraction of the memory that current research systems need. It has not yet been tried with a live user.
10:21CodeIIIOpen source
A new benchmark from Amazon researchers chains 20 coding requests together. Ten agents complete about three quarters of the single requests but only about half of the chains, mostly because they build on mistakes they never notice.
Sunday 15 November
08:00In industryIIBenchmarked
A five-week study of 404 engineers at JetBrains finds that an engineer's AI token bill says little about how much code they commit. The paper is a preprint and has not been peer reviewed.
08:00Safety & evaluationIIBenchmarked
An audit of five open medical image-reading models finds that a single accuracy score can hide total failure on one diagnosis. Two models never flagged a pneumothorax, yet explained their wrong answers fluently.
08:00Vision & multimodalIIBenchmarked
A preprint trains an MRI reconstruction model mostly on fake scanner data made from everyday photographs. On scans from other hospitals and machines, it beat methods trained on real patient scans.
08:00Safety & evaluationIIBenchmarked
A preprint shows that votes, prompts and decoy models can lift a language model far up a crowd-voted ranking. The tests were simulations on public LMArena data, not attacks on the live site.
All months