Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
A researcher revealed that federal officials were alarmed by Fable 5 after a simple 'fix this code' prompt, not a jailbreak, according to the person who read the research paper.
A developer created Practical AI Engineering, an open-source curriculum for building production AI systems with LLMs. The curriculum covers 8 phases from token basics to deploying monitored systems with safety guardrails…
Multiple news organizations and platforms are using generative AI to forecast 2026 FIFA World Cup outcomes. Microsoft Copilot generated score predictions for matches on June 14-15, Alibaba's Qwen introduced a match predi…
Google's John Mueller and Martin Splitt argued on a podcast that stripping web pages to markdown for AI SEO removes crucial elements like links and navigation that aid search engine discovery and ranking, countering clai…
Big tech companies are amassing massive GPU clusters, giving them an insurmountable compute advantage over local models. Even as local models improve, the gap in intelligence between local and datacenter models will wide…
Moonshot AI released Kimi K2.7-Code, a coding-focused agent model with 1 trillion total parameters and a 256K context window, aimed at multi-step software engineering tasks. The Beijing lab positions the open-weight mode…
A developer released Cultivation World Simulator, an AI-driven simulation where each character is an independent LLM agent, on Epic Games Store for free. The game allows players to observe and subtly intervene in a world…
A Hacker News user asks why no US company has released open-weight AI models, noting that Chinese labs and Europe's Mistral are more dedicated to open weights than US labs like OpenAI, Google, and Microsoft, which have r…
A developer outlines a three-layer safety framework for AI agents that interact with databases, emphasizing structural safeguards over simple prompts. The approach includes action boundaries, pre-execution validation, an…
A developer identifies five common retrieval failures in RAG systems that cause incorrect answers, including chunk boundary issues, semantic ambiguity, ranking failures, stale data, and low-confidence retrieval. The post…
Sonar reports that AI agents generate bloated, unmaintainable code, citing a Carnegie Mellon study of 807 open-source projects using Cursor that found a 30% rise in static analysis warnings and 41% increase in code compl…
France announced an additional €655 million for artificial intelligence, primarily to develop a single sovereign chatbot for its entire civil service of roughly one million employees. The investment, part of a broader €1…
A developer built GetCiteFlow, an AI visibility scanner that uses a hybrid analysis architecture combining LLM evaluation with deterministic checks to measure how well websites are cited by AI search engines like ChatGPT…
PagedAttention, a memory optimization technique in the vLLM inference server, applies virtual memory concepts to manage the KV cache in large language models, improving throughput by reducing fragmentation. The KV cache …
A developer proposes using ArchUnit tests to enforce package boundaries when using AI code generators like Cursor Composer, preventing architectural drift. The approach involves writing executable fitness functions as JU…
Upstage CEO Sung Kim told reporters in Seoul that a US directive restricting foreign access to Anthropic's advanced AI models underscores the need for South Korea to develop sovereign AI capabilities. Anthropic disabled …
Anthropic executives will meet senior White House and Commerce Department officials Monday after the US government restricted access to its Fable 5 and Mythos 5 models, citing a potential jailbreak vulnerability. The Com…
Google Cloud launched the Open Knowledge Format (OKF), a vendor-neutral markdown specification designed to provide AI agents with curated context. The spec formalizes the LLM-wiki pattern using markdown files with YAML f…
FoxBox has become the first SMS hardware gateway to implement the Model Context Protocol (MCP), enabling native AI and automation interoperability for on-premise SMS infrastructure. The MCP integration allows AI agents a…
A developer built RagScope, an open-source tool that audits RAG pipelines to reveal wasted retrieval latency from chunks fetched but never used in the final prompt. The tool uses OpenTelemetry traces to compare retrieved…