Neura News

AI News

News reporting focused on AI and machine learning, covering the companies behind these technologies, their real-world applications, and the ethical concerns they raise. This includes areas like generative AI (large language models, text-to-image and video), speech tech, and predictive analytics.

Latest News

16 articles
Research

Modular Pretraining: A New Approach to Containing Dangerous AI Knowledge

Researchers at Anthropic and AE Studio have introduced Gradient Routed Auxiliary Modules (GRAM), a method that isolates dangerous knowledge in large language models into switchable modules during training. This approach allows operators to control access to sensitive content, potentially reducing risks of misuse. Preliminary experiments show promise across models up to 5B parameters, but the method has not yet been applied to production-scale systems.

Aug 1712 minNeura News
Research

LittleLearner Models Trained Only on K-5 Curriculum Show Skills Are Elicited, Not Acquired

Researchers released LittleLearner, a family of language models trained from scratch on a strictly filtered K-5 elementary school curriculum, to answer whether capabilities beyond training data can be elicited or acquired through scaling, post-training, and in-context learning. The answer is largely no: scaling, post-training, and in-context learning amplify what the curriculum taught, but none meaningfully improve out-of-scope performance. The pretraining filter sets the effective capability ceiling, providing a controlled sandbox for studying knowledge acquisition and RL.

Aug 165 minNeura News
AI Models

Former OpenAI employee launches data startup, predicts $100 billion shift in AI training strategy

Andrew Ho, a former OpenAI employee, has founded a startup focused on producing high-quality training datasets, predicting that AI labs will spend over $100 billion on targeted data collection. He argues that the current scaling approach for large language models is failing to achieve genuine generalization, especially in specialized fields like bioinformatics. The article explores the broader debate on AI specialization versus versatility, with researchers from Cambridge and Google DeepMind supporting the view that current models are hitting a ceiling on creative problem-solving.

Jul 304 minNeura News
AI Models

Petals Lets Users Run Large AI Models at Home Like BitTorrent

Petals is a decentralized platform that allows users to run large language models such as Llama 3.1, Mixtral, Falcon, and BLOOM on consumer-grade hardware by sharing computational resources in a peer-to-peer network. Users load only a portion of a model and join a network of others serving the remaining parts, enabling inference speeds of up to 6 tokens per second for Llama 2 70B and 4 tokens per second for Falcon 180B. The platform supports fine-tuning, custom sampling methods, and access to hidden states, combining the convenience of an API with the flexibility of PyTorch and Hugging Face Transformers.

Jul 232 minNeura News
AI Tools

AWS details observability strategy for SageMaker AI LLM inference

AWS published a technical guide on comprehensive observability for large language models deployed on Amazon SageMaker AI. The approach separates monitoring into infrastructure quantity and LLM quality dimensions, using CloudWatch and Amazon Managed Grafana to visualize GPU utilization, cost, and response quality signals. The solution includes alert thresholds for metrics such as safety scores and composite quality scores, evaluated by an LLM-as-judge setup.

May 304 minNeura News
Research

Fake AI citations in biomedical papers surge twelvefold since 2023

A new audit of 2.5 million biomedical papers found fabricated references have increased more than twelvefold since 2023. Researchers at Columbia University and other institutions published the study in The Lancet, linking the rise to widespread use of language models like ChatGPT. The fake citations are hard to detect and pose special risks because they appear frequently in review articles that inform clinical guidelines.

May 263 minNeura News
AI Models

Hassabis sees AI singularity soon; LeCun says LLMs not intelligent

At Google I/O 2026, DeepMind co-founder Demis Hassabis declared humanity is entering the singularity, predicting AGI within five years. Yann LeCun of AMI Labs countered that current language models lack true intelligence, which he defines as solving new problems without prior training. Gemini co-lead Oriol Vinyals offered a middle ground, acknowledging models are strong in code and math but still missing experiential learning.

May 243 minNeura News
AI Models

NVIDIA Nemotron-Labs Diffusion Models Speed Up Text Generation

NVIDIA has released Nemotron-Labs Diffusion, a family of diffusion language models (DLM) that generate multiple tokens in parallel and iteratively refine them. Available in 3B, 8B, and 14B scales, the models support three inference modes: autoregressive, diffusion, and self-speculation. The 8B model achieves up to 6.4x higher tokens per forward pass compared to autoregressive models while improving average accuracy by 1.2% over Qwen3 8B.

May 234 minNeura News