Circulating tumor DNA (ctDNA) is emerging as a promising biomarker for postoperative monitoring of cancer patients. Precise estimation of circulating tumor fraction is crucial for evaluating treatment effects and timely detection of disease recurrence.
An analysis of what models know and how it tells us about how they were trained. By scoring them on niche facts, we can approximate how many parameters models like GPT-5 and Opus have using 'Incompressible Knowledge Probes'. By measuring how the models break down tokens, we can reveal facts about the datasets mixtures they used to train the model (or at least the tokenizer) using 'Data Mixture Inference'.
Muse Glimmer is now available on Ollama. It's a 30B multimodal model purpose-built for agent workloads that run locally. With Ollama, you can use Muse Glimmer to power coding agent applications such as Claude Code, Codex, Pi and more.
AI creates new forms of resource dilemmas, reminiscent of traditional commons tragedies.
AI company Recursive Language Models predicts that in 2026, they will use recursive language models to manage extremely long contexts and implement complex changes in large codebases.
The authors explore the connection between heart rate fluctuations and brain structure, finding that low-frequency heart rate variations are linked to specific aspects of brain function. They propose using these data to create a physiological fingerprint that corresponds to brain architecture.
A new study reveals that aging hematopoietic stem cells trigger a form of innate immune memory, known as trained immunity, which fuels the chronic low-grade inflammation that underlies age-related physiological decline. SIRT3, a mitochondrial deacetylase whose expression wanes with age, emerges as a key molecular brake on this process.
Biogen has completed the acquisition of RayThera Inc., a privately held biotechnology company focused on discovering and developing small-molecule therapies in immunology. The transaction adds multiple assets to Biogen's immunology portfolio, with potential to treat immune-mediated conditions across various indications.
Meta has released Muse Glimmer, the first open model from Meta Superintelligence Labs. It's a dense 30B parameter model built for local agentic and coding workflows. Available via Unsloth and Unsloth Dynamic quants for best performance.
We investigated the impact of sex on age-related proteome changes in ex vivo-degenerated young and aged murine peripheral nerves, revealing differences that suggest intrinsic nerve changes or altered systemic environment.
The author is ambivalent about the progress of GenAI and its impact on his career but appreciates its potential for personalized learning. He seeks advice on how to use LLMs to enhance advanced technical knowledge.
The Meta Muse Glimmer technology allows users to see and interact with virtual objects created based on their thoughts.
Nature published a study that reveals key roles of lysosomal dysfunction in heart aging.
Experts from Longevity.Technology have unpacked what actually keeps us healthy as we age. It's not just about protein intake, but also physical activity and lifestyle habits.
Study aims to use DNA methylation biomarkers to capture residual biological risk beyond existing clinical risk-factor models like PREVENT.
The perception of pain varies both within and between individuals, even when sensory input remains constant. Understanding how the brain encodes these intra- and inter-individual variations is central to elucidating the neural mechanisms of pain and to developing reliable brain-based markers for clinical use.
We develop RareCapsNet, a rare cell identification technique through capsule network in large single-cell RNA-seq data. The goal is to leverage the landmark advantages of capsule networks in single-cell domain by identifying rare, poorly represented cell populations.
Study finds that age and plasma p-tau217 levels are associated with clinical progression, predicting cognitive decline in memory clinic participants.
Without strong standards and smart oversight, artificial intelligence risks eroding the foundations of higher education.
In a third phase trial, obinutuzumab β significantly reduced relapse risk by 93.1% compared to placebo in patients with AQP4-IgG+ NMOSD, improved clinical and MRI outcomes, showed manageable safety, and demonstrated sustained efficacy without relapses during extended follow-up.