Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Amazon's 2026 Project Eluna deploys agentic AI models in fulfillment centers to autonomously plan, use tools, and execute multistep operational tasks, moving AI beyond digital applications into physical environments. The…
Bending Spoons filed for an IPO, positioning itself as an anti-AI play in a market where tech stocks fell on Friday due to a strong labor report raising interest rate concerns. Meanwhile, OpenAI is expected to release GP…
NVIDIA has introduced the NVFP4 training recipe in TransformerEngine, enabling 4-bit mixed-precision pre-training on Blackwell GPUs with no measurable accuracy loss compared to FP8 baselines. The recipe, demonstrated in …
On May 29, security researcher Taylor Hornby discovered a critical vulnerability in Zcash's Orchard privacy pool using Claude Opus 4.8, a flaw that could have allowed attackers to generate ZEC from nothing by bypassing t…
Andrej Karpathy's LLM Wiki design is being adopted as an external memory architecture for AI agents, mapping raw source documents into a structured markdown knowledge base. The system uses three layers—immutable raw sour…
Nils Boeffel launched a Leanpub book titled 'Modernizing Legacy Software with AI: A Technical Leader's Guide to AI-Driven Transformation,' which presents a methodology that completed a legacy modernization project in rou…
Amazon researchers have identified a fundamental performance bottleneck in AI agent systems, formalizing it as the "intent-execution gap" — the mismatch between what an AI model intends to do and what its software harnes…
Economist Thomas Piketty has reportedly aligned with degrowth ideology, according to social media posts, while other commentators suggest prominent economists hold a negative understanding of economic systems. The claims…
A new study reveals that large language models (LLMs) silently corrupt documents when delegated editing tasks, with even top-tier models like Gemini Pro, Claude Opus, and GPT-5 degrading 25% of original content after 20 …
Google upgraded NotebookLM on Monday with Gemini 3.5 and Antigravity, achieving over 65% accuracy improvements across core evaluations. The update adds secure cloud computing for code execution, expanded output formats i…
AI agents that remember user preferences but forget their own prior actions appear unintelligent due to state failures, not model limitations, according to engineers who shipped a memory layer in production. The hard par…
KiloBench launched as a new evaluation framework that measures AI coding models based on real-world production cost and performance rather than benchmark scores. The tool emerged after its creators found that a model sco…
AI safety researchers are exploring whether future advanced AI systems could be offered incentives like money or computing power in exchange for cooperating with human operators, rather than attempting to seize control. …
Anthropic published its official guide to building Claude Skills, a system that allows users to create reusable instruction folders for the AI assistant. The guide covers the technical structure of skills, including the …
A developer reports that running gpt-oss-120B on an NVIDIA DGX Spark (GB10) failed due to MXFP4 quantization lacking native hardware support on SM121, causing kernel hangs and memory spikes. Switching to Qwen3.6-35B-A3B …
Google released Gemma 4 12B, a multimodal AI model with an encoder-free architecture that processes images and audio directly without separate encoders, enabling on-device agentic workflows on standard laptops. The 12-bi…
Anthropic reported that its Project Glasswing initiative, launched in April to use its Mythos AI model for identifying software vulnerabilities, has detected 23,000 potential flaws across 1,000 open-source projects. Desp…
The growing use of artificial intelligence in weather and climate modeling relies on machine learning techniques that researchers have studied for years, not the generative AI behind viral chatbots. A National Weather Se…
Anthropic's Claude Mythos program has reported 23,019 vulnerabilities but only demonstrated fixes for 75, raising questions about its claimed frontier-exclusive capabilities. Independent teams have reproduced the showcas…
GitHub Copilot experienced degraded availability for the Claude Opus 4.7 model across Copilot Chat, VS Code, and other Copilot products on June 8 due to an issue with an upstream model provider. The incident was resolved…