Updated daily708 verified stories
Research News tracker.
Material AI papers, studies, datasets, evaluations, and scientific advances with direct links to the underlying research.
- 30stories this week
- 708sources cited
- 708verified timeline entries
Verified sourcing
Every story links to the strongest available evidence: original sources when available, otherwise clearly attributed reporting.
Plain English
What happened, why it matters, and what to watch — without the jargon.
No filler
When the signal is thin, we publish nothing rather than padding the feed.
More stories
9 storiesInnovation
DirEAG paper proposes a better way to calibrate AI confidence in math answers
A new arXiv paper introduces DirEAG, a method that combines confidence reports from multiple prompts into calibrated evidence about possible answers, including the possibility that none is correct.arxiv.orgInnovation
Study says LLM persona evaluations change when question frames shift
An arXiv preprint reports that GPT-4o’s simulated persona patterns depend partly on how personality questions are ordered and aligned, suggesting that single aggregate scores may miss important behavior.arxiv.orgInnovation
AI search method reports more efficient discovery of viable neutrino flavor models
A new reinforcement-learning framework uses symbolic failure certificates to steer searches through more than 10^26 theoretical neutrino flavor models, the arXiv preprint reports.arxiv.orgInnovation
Survey maps the unresolved path to faster multimodal AI generation
A new arXiv survey and empirical study finds that diffusion-based parallel drafting remains largely unexplored for multimodal AI, despite reported speedups in text generation.arxiv.orgInnovation
Consilience proposes calibrated communication control for multi-agent AI reasoning
An arXiv preprint introduces Consilience, a framework that selects and certifies how language-model agents communicate when each holds only part of the evidence. The authors report higher decision accuracy and greater communication efficiency than fixed or unstructured discussion protocols on hidden-profile tasks.arxiv.orgInnovation
Preprint proposes an auditable framework for finance-focused LLM analytics
A new arXiv preprint argues that enterprise finance systems using large language models should be judged not only by answer accuracy but also by whether every claim can be traced to authoritative evidence. Its proposed framework improved citation traceability in a 145-question evaluation, while producing…arxiv.orgInnovation
Survey maps how AI agents operate through command-line environments
A new survey proposes a common framework for studying AI agents that act through terminals, arguing that model quality alone does not determine their behavior.arxiv.orgInnovation
VortexChat uses an AI agent to automate integrated photonic design
An arXiv preprint introduces VortexChat, a framework that uses a large language model to coordinate photonic-device generation, refinement and electromagnetic simulation. The authors report that it autonomously designed and helped fabricate a terahertz vortex-beam multiplexer whose measured performance matched…arxiv.orgInnovation
Weighted Memory Tree reports gains for long-horizon LLM agents
A new preprint proposes a hierarchical memory system that decides which parts of an AI agent’s execution history should remain active. On GAIA-Text, the authors report higher accuracy, lower prompt-token use and less persistence of poisoned information than linear memory.arxiv.org
One useful briefing each week
Keep up with AI without living in the feed.
Get the week’s verified AI news, original data, useful tools, learning picks, and fresh AI jobs.
Reach people who are learning AI
Hiring an AI professional or launching a useful AI product? Put it in front of people who came here to learn and act.
Post an AI jobSubmit an AI tool