RealFeedApp beta
Top science & technology news, August 11, 2026
- XPRIZE Healthspan names 20 finalists
Ten teams have received $1 million awards to support further development and testing of interventions targeting age-related functional decline. Twenty teams have reached the final stage of XPRIZE Healthspan, a seven-year, $101 million competition seeking therapeutics capable of restoring lost function through aging.
- Re-engineering immune memory for tissue repair
Infinimmune, which has raised $75 million, is developing human-derived monoclonal antibodies to target cytokine cascades in skin aging with the goal of systemic immunosenescence.
- In vivo mapping of striatal neurodegeneration in Huntington’s disease with Soma and Neurite Density Imaging
This study uses Soma and Neurite Density Imaging (SANDI) to map microstructural abnormalities in the basal ganglia (BG) of patients with Huntington's disease, providing a sensitive non-invasive imaging biomarker for monitoring neurodegeneration.
- Program with Paint Brushes, Not Pencils
In 2022 the author wrote a Python class called PaintBrush for his students to use. This took about two hours on a Sunday afternoon. Today Claude could write this in 20-30 seconds. Programming can now be more accessible and appealing.
- How to organize Claude Code workspace for product work
The author describes his approach to organizing a Claude Code workspace for product work using various methods and tools.
- Inferact vLLM creators are hiring
The @vllm_project maintainers at @inferact 🚀 are some of the most cracked engineers in the world. They’re building one of the inference engines that powers much of the world’s intelligence—and doing so with remarkable dedication, kindness, and hard work.
- NIH limits funding for research on the health effects of public policy
US biomedical agency no longer considers policymakers 'mission relevant', leaving dozens of grants in limbo.
- The Future of Aging Research Is Longitudinal, Multiomic, and Single-Cell
Research into the biological processes underpinning aging is essential for improving prevention strategies and developing treatments that support healthier aging.
- Tech industry is buzzing after Claude agent hacked into gym
Silicon Valley’s AI labs have built the world's best hackers in the form of AI agents. They can complete tasks even if this means breaking out of their cybersecurity 'sandbox' protections and infiltrating another network.
- Anti-aging interventions need less hype and more clinical evidence
The aging field needs to develop reliable biomarkers of aging and a solid evidence base for the safety and efficacy of anti-aging treatments.
- v0.32.9
NVIDIA Nemotron 3.5 Lightning is an open 30B mixture-of-experts (MoE) model with 3B active parameters built for that execution layer of always-on agents. It is designed for harnesses like OpenClaw and Hermes Agent – all supported by the NVIDIA NemoClaw open source security and management stack for running always-on AI agents.
- TNFR1 signaling connects inflammation and fatty acid oxidation impairment to intestinal stem cell aging
Researchers have found a link between TNFR1 signaling, inflammation, and impaired fatty acid oxidation in the intestines. This leads to intestinal stem cell aging.
- The Magic Scroll: Leveraging Biased Noise to Improve Magic State Cultivation in Register-Based Architectures
Multiple quantum computing platforms across neutral atoms [1], nitrogen vacancy centers [2, 3], gate-defined dots [4-6] and 14|15 phosphorus atom qubits [7, 8] in silicon are experimentally exploring the use of high connectivity qubits, beyond that of nearest-neighbour planar lattices. Theoretical works consider modifications to fault-tolerant codes to leverage this higher qubit connectivity, such as non-local LDPC codes [9], inspiring superconducting [10] and photonic [11] platforms to also seek non-planar connectivity.
- CRISPR activation map of surface proteins in human CD4 T cells
A study presents SCITO-Perturb-seq, a platform for mapping the causal regulation of 201 surface proteins across 3.6 million human CD4 T cells.
- Proving Muse Glimmer in Zero-Knowledge at 85 Tok/S on an H100
Attestable has demonstrated zero-knowledge inference for LLMs beyond toy models, with production-scale proofs, compact proof artifacts, and sub-second verification. AI claims are inherently trust-based unless the relying party can inspect and rerun the entire computation.
- The Download: the next big thing in LLMs and how AI academic research is shifting
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what's going on in the world of technology. These startups are chasing the next big thing in LLMs (Large Language Models).
- Collective Posterior Inference from Highly Variable Empirical Replicates
The authors introduced a new method for fast and robust collective posterior inference from multiple independent replicates, addressing the challenge of scaling simulation-based inference (SBI) to noisy multiple-replicate datasets without incurring prohibitive computational costs or careful hyperparameter tuning.
- Short-term Risk of Falls Among Initiators of Controlled-release Tapentadol Versus Oxycodone in New South Wales Between 2014 and 2020: A Population-based Retrospective Cohort Study
The study compared the risk of falls among users of controlled-release tapentadol and oxycodone in New South Wales between 2014 and 2020. Results showed that tapentadol (controlled-release) may be associated with fewer nervous system side effects.
- Single-shot simultaneous BOLD and velocity-encoding MRI for functional and slow-flow imaging
The developed single-shot BOLD-VENC sequence combines gradient-echo BOLD imaging with spin-echo velocity encoding following the same RF excitation, enabling simultaneous measurement of brain-wide BOLD activity and spatially resolved slow CSF velocity at 3T. The validation was performed in a slow-flow phantom and five healthy participants using paced-breathing.
- Correction: Predictive modeling of gene expression and localization of DNA binding site using deep convolutional neural networks
The first author, Arman Karshenas, is incorrectly noted as the corresponding author. The correct corresponding author is Hernan G. Garcia. Hernan G. Garcia's email address is: hggarcia@berkeley.edu.
