Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Anthropic CEO Dario Amodei refused a Pentagon ultimatum to remove usage restrictions on the Claude AI model, which prohibited mass domestic surveillance and fully autonomous weapons. The US government retaliated by banni…
Flama, a Python web framework, now offers native support for building Model Context Protocol (MCP) servers, enabling AI agents to call functions, read data, and use prompt templates through simple decorators on Python fu…
A developer built an automated extraction and chunking pipeline that converts any documentation site into clean, structured markdown ready for vector stores. The pipeline, available as the RAG Docs Extractor on Apify, re…
LlamaIndex and Pinecone have become a reliable combination for production RAG systems, with LlamaIndex handling orchestration and Pinecone managing vector storage. A 2024 report indicates over 60% of AI pilot projects st…
Apple announced a major revamp of the Shortcuts app in iOS 27, integrating Apple Intelligence to allow users to create automations using natural language instead of navigating complex menus. The update aims to make the p…
A developer created apte, an async-first test framework for Python that unifies unit tests and LLM evals in the same engine with shared fixtures. The framework uses typed dependencies with Annotated and Use, supports nat…
Retrieval-augmented generation (RAG) pipelines are failing in production because most teams focus on the retrieval and generation layers while neglecting data quality, according to experts. Mayank Bhola of TestMu AI and …
A developer participating in the Gemma 4 challenge built Dresscode, an AI stylist that uses computer vision to digitize wardrobes and LLM function calling to recommend weather-appropriate outfits. The project demonstrate…
Anthropic accused Alibaba of using nearly 25,000 fake accounts to extract capabilities from its Claude AI models, generating over 28.8 million exchanges. The alleged campaign, described as the largest known attack of its…
Florida has filed a lawsuit against OpenAI and its CEO Sam Altman, alleging the company concealed risks associated with its ChatGPT product from users.
A developer built a C# LLM event parser called EventParser and tested it using Promptfoo's LLM-as-a-judge evaluation. The project separates the prompt file from the code, allowing Promptfoo to test the same prompt used i…
Meta has replaced roughly half of all human moderation requests with large language models in 2025 and plans to push that share above 90% for some content types by year-end, a shift employees warn is too fast and lacks o…
A single H200 GPU with 141GB HBM3e cannot comfortably run DeepSeek V4 Flash (284B total, 13B active parameters) due to VRAM constraints, even with 2TB system RAM for offloading. The model requires an 8-GPU H200 node for …
Canopy Wave announced a guide on using the world's first unlimited token plan with OpenCode, an open-source AI coding assistant, to help developers manage intensive coding sessions and agent workflows.
Kreuzcrawl v0.3.0 ships with 14 language bindings, a tiered WAF-aware dispatch engine, and reduces peak streaming memory from ~2.5 GB to ~20 MB. The release also enables SSRF defense by default across all outbound calls …
Florida Republican Anna Paulina Luna was caught using AI to draft a bill amendment after text including 'Claude responded:' appeared in the document. Luna initially claimed AI was used only for spellcheck and grammar, th…
Ora Computing launched an automated LLM compression engine that reduces model size by up to 70% with minimal accuracy loss, enabling deployment on edge devices, on-prem servers, or cloud infrastructure. The technology cu…
A practitioner's field guide details five distinct prompt injection attack classes that evade simple keyword filters, including instruction smuggling through metadata, cross-document poisoning in RAG systems, indirect in…
A developer used AI agentic coding to build a system that crawls a billion web pages in 24 hours, writing less than 4% of the code by hand. The project achieved high throughput at scale, demonstrating that vibe coding ca…
A developer released cc-preview, an open-source macOS tool that extracts and displays images pasted into Claude Code sessions, replacing placeholder text with a native preview interface. The tool uses Hammerspoon to prov…