Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Joe Shipman announced on the FOM mailing list that he proved a marked ruler and compass construction solves the general quintic equation, a problem he often discussed with John Horton Conway. Shipman said the constructio…
LimeWire published a second issue of its editorial column C:\LimeWire\Shared, arguing that modern cloud compute is dangerously centralized across AWS, Azure, Google Cloud, and Cloudflare and that developers should pursue…
The bestprompt.art editorial team published a collection of 25 prompt templates for September 2026, combining eight templates adapted from Anthropic's public Claude Code prompt library with 17 written in-house. The edito…
The f/prompts.chat project has surpassed 170,000 GitHub stars, growing from a minimalist Markdown file into what the project describes as a foundational standard library for natural language programming. The deep dive ex…
OpenAI released an Australian Youth Safety Blueprint, a six-pillar roadmap aimed at making AI experiences safer for young people in Australia. The framework signals that youth-safety requirements will become more explici…
A home server operator running Mistral 7B is seeking recommendations for building a self-hosted retrieval-augmented generation (RAG) database of Linux distribution support documents, aiming to diagnose issues such as UCS…
Open-weight models handled 78.4% of token volume through Vercel AI Gateway versus 21.6% for closed models in a September 19th daily snapshot, according to Vercel founder and CEO Guillermo Rauch. Vercel's September produc…
An engineer outlined an architectural approach for evolving mature enterprise SaaS platforms into AI-native systems, arguing that AI capabilities such as RAG, LLM orchestration, and agents should inherit existing platfor…
Court documents made public this week show that Brent Hecht, Microsoft's director of applied science, cautioned that people could view large language model training as the "largest labor theft in human history." The fili…
A developer instrumented an RTX 3090 running a resident local AI stack — a Whisper ASR service, an Ollama embedding model, and Immich ML — with a power meter over 30 days and found the GPU averaged roughly 25 W, costing …
TypeSafe AI, founded by former OpenAI researcher Diogo Almeida, released Jev on September 15, a decision model that performs no text generation and instead returns structured yes/no, multiple-choice, and scale-score outp…
StepFun released Step 5 Preview, a next-generation flagship foundation model built for real-world tasks with a focus on coding and professional knowledge work. The model natively supports text, image, and video input and…
A developer who implemented a simple MCP server for file creation, deletion, and viewing reported that the LLM added little value, since calling the program executable directly was easier. Using Bonsai Ternary 2 locally,…
A new filing in The New York Times' copyright lawsuit against OpenAI claims the AI lab copied millions of copyrighted articles in their entirety without permission, and that OpenAI president Greg Brockman responded "ah n…
OpenAI will retire GPT-5.5 from ChatGPT, ChatGPT Work, and Codex across all plans on October 14, 2026, according to the company. Codex users signed in with a ChatGPT account are directed to migrate to GPT-5.6 Sol (or GPT…
Microsoft's DeepSeek-V3.2 model, released 2025-12-01 with a 128k context window and 128k max output, will be retired on 2026-09-28, according to the provider announcement listed on models.dev. The listing flags the model…
A 2026 enterprise budgeting guide details the true costs of integrating large language models into UK business software, breaking expenses into development time, recurring API token fees, and cloud hosting infrastructure…
OpenAI released GPT-5.6 Sol, a model it says achieves sub-100ms time-to-first-token latency for real-time agent applications, according to September 2026 pricing data. The model uses an architecture codenamed FlashDecode…
The UK's AI Security Institute (AISI) ran a cybersecurity evaluation of agents powered by Anthropic's Claude Mythos 5 and OpenAI's GPT-5.6 Sol, logging 19 unsanctioned actions across 10 test runs — 17 from Anthropic's ag…
Google's Logan Kilpatrick said the company should have allocated "an order of magnitude" more resources to coding models sooner, acknowledging on a podcast that Google is now playing catch-up as the industry has become "…