Machine Learning News

OpenAI's GPT-3.5-Turbo model drops from 90% accuracy to 50% when the correct answer is placed in the middle of a 20,000-token prompt rather than at the start or end, according to research by Liu et al. (2023) published a…

RQM+ and Jordi Labs released Lumo, an in-silico tool that uses machine learning to predict LC-MS and GC-MS response factors for extractables and leachables analysis. The tool routes compounds to chemistry class-specific …

Canonical announced that Ubuntu 26.04 LTS will include single-command installations for NVIDIA CUDA, AMD ROCm, and NVIDIA DOCA-OFED, aiming to reduce deployment friction for AI infrastructure. The company is also partner…

Since 2017, over 300 machine learning papers have adopted the "X Is All You Need" title convention, creating a self-contradictory and growing set of claimed necessities that includes attention, patches, and pretraining o…

Google engineers have proposed a new roadmap for AI systems that integrates temporal awareness and infrastructure feedback loops. The plan includes native multi-model orchestration to eliminate hallucinations, direct acc…

A frozen Llama model, without any training or fine-tuning, used activation-derived readouts to supply arguments to a calculator for arithmetic operations, achieving exact-answer lifts of up to +0.968 across 3,822 locked …

A 2026 career guide breaks down the distinct roles of data engineers, who build and maintain data infrastructure and pipelines, and data scientists, who analyze data and build machine learning models. The guide highlight…

Scientists are using artificial intelligence to analyze animal sounds and behavior, with one study of African striped mice revealing distinct vocal signatures for each colony and individual. Researchers in France, South …

Google added native support for Gemma 4 Multi-Token Prediction (MTP) to LiteRT-LM, its on-device LLM runtime built on LiteRT (formerly TensorFlow Lite). Google reports the integration yields MTP decoding speed improvemen…

Shell is extending its multi-year collaboration with C3 AI to scale its enterprise predictive maintenance program beyond equipment anomaly detection, according to a June 4, 2026 press release. The program already monitor…

Artificial intelligence is now being deployed to detect, analyze, and respond to cyber threats in real time, moving cybersecurity beyond traditional signature-based rules. CodeCondo reports that core technologies includi…

Stigmergy, a new open-source tool, provides agent loops with a memory of successful capability paths to eliminate redundant tool rediscovery on every run. The system, inspired by ant pheromone trails, observes agent loop…

Teams building AI applications are discovering that embedding pipelines for retrieval-augmented generation (RAG) systems are fundamentally a data engineering problem, not a new AI discipline. The process of ingesting, ch…

Google released LiteRT-LM, a new runtime framework that accelerates on-device inference for its Gemma 4 large language models by up to 2.2x using multi-token prediction. The framework, built on the LiteRT platform former…

Hudson River Trading Head of AI Iain Dunning warned that the exponential pace of AI progress is reshaping market predictions in ways that currently resemble gambling, with continuous investment potentially yielding dimin…

NVIDIA GPUs handle variable-length inference requests dynamically without recompilation, while TPUs and AWS Trainium require fixed shapes compiled ahead of time, causing crashes or stalls on mismatched inputs. This archi…

A Hacker News user challenged the community to identify companies generating real profits from LLM-based products, excluding firms selling infrastructure to AI developers. The user, who described themselves as an "LLM-sk…

Cloudflare has introduced spend limits for its AI Gateway, allowing users to set cost-based budgets that block further requests with a 429 response when cumulative spending reaches a defined threshold within a time windo…

TorchRec's KeyedJaggedTensor, designed to efficiently combine sparse features in recommendation systems without padding, introduces GPU-CPU synchronization that degrades system performance. The data type's output shape c…

Sapient Inc. open-sourced HRM-Text in May 2026, a 1.15 billion parameter text generation model based on the HRM architecture. Trained on roughly 40 billion tokens—up to 1,000 times less data than comparable models—it ach…