Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
NVIDIA's H100, H200, and B200 GPUs offer different trade-offs for AI workloads, and choosing among them depends on factors like model size, memory requirements, and whether the use case is training or inference. The H100…
Two July 2026 papers demonstrate that AI models improve by compressing information rather than accumulating it. The Chinese Academy of Sciences' Institute of Automation released PhiZero, a world model that compresses vid…
A developer's optimization campaign on an RTX 3080 Ti cut a GPT-2-small-class transformer's forward pass from 78.2ms to 1.60ms, a 49× speedup, beating torch.compile's 1.72ms and reaching 136,000 tokens/second. The single…
A technical explainer from tensor.khalilli.ai demonstrates that the floating-point arithmetic error causing 0.1 + 0.2 to equal 0.30000000000000004 and training loss to become NaN at step 40,000 stems from the finite numb…
Sail Research has open-sourced HTDYM (How To Deploy Your Model), an internal performance modeling infrastructure that helps determine the optimal hardware and sharding configuration for serving large language models, aim…
NVIDIA released NemotronLabs VoiceChat 11B, an open 11B end-to-end speech-to-speech model for real-time full-duplex conversation, achieving 448 ms smooth turn-taking latency on Full-Duplex-Bench 1.0 and a user-interrupti…
A developer incorporated patches from a reader named Madhu that enable llambda.lisp, a Lisp-based inference project, to run on Linux. The patch uses mmap to load model weights into the Lisp address space outside the heap…
Neural scaling laws are empirical statistical laws in machine learning that describe how neural network performance changes with model size, training dataset size, and training cost, with some models also scaling inferen…
An ICML 2026 paper submitted July 9 reports that amplifying the weight-space difference between Qwen3-VL instruct and reasoning checkpoints made planted secrets or unintended behaviors surface up to 10 times more often t…
A developer built LoanAI, a real-time credit risk assessment platform that predicts loan default probabilities using machine learning models. The system, built with Flask, Bootstrap 5, and Python, offers instant risk sco…
An AI model adapted from image inpainting can reconstruct missing atoms in crystal structures with a reported 97 percent success rate, addressing gaps that X-rays cannot detect, such as hydrogen atoms. The model, develop…
This week's top developer tools include TensorFlow Quantum 2.0, PyTorch Lightning 1.8, Azure AI SDK for Go 1.0, CLIMA, and Django AI, each offering new capabilities for machine learning and AI integration. A developer al…
Google DeepMind's WeatherNext Cyclones model delivered an average lead-time advantage of a day or more over leading operational tropical-cyclone models for track, intensity, and wind-radius forecasts, according to a Natu…
A developer seeking advice on experimenting with retrieval-augmented generation (RAG) pipelines asked the community how they compare configurations of retrievers, chunking strategies, embeddings, rerankers, and LLMs, and…
IREE, an open-source compiler from the IREE project, now compiles machine learning models to native RISC-V CPU code with support for the RISC-V Vector extension (RVV), hand-written microkernels, and data-tiling. The proj…
A developer shared a beginner-friendly tutorial on loading datasets into Jupyter Notebooks using the Pandas library in Python. The guide demonstrates importing Pandas, using pd.read_csv() to load local or remote CSV file…
A developer announced the start of a learning and teaching journey on DEV, centered on building a Fake News Detection System using Python, Machine Learning, NLP, and Django. The project involves concepts like data prepro…
DeepMind's WeatherNext AI model can predict deadly cyclones one day further in advance than existing methods, generating a 15-day forecast in under a minute on Google's custom Tensor Processing Unit chips. In tests again…
GPT-5.6 and Claude Fable have settled a 25-year-old open problem in wireless communications, proving that a simple polynomial-time algorithm can achieve maximum likelihood detection in MIMO systems at the same signal-to-…
A developer who completed Codeboxx's AI for Developers certification built a full-stack operations platform for Rocket Elevators that monitors 45,000+ elevators across Canada in real-time. The system integrates predictiv…