OpenAI introduces GPT-6 Sol and Luna, two new frontier AI models designed to bring advanced intelligence to daily tasks with varying balances of capability and cost.
OpenAI details how GPT-6 improves prompt caching with higher cache hit rates, new diagnostics, and explicit controls, leading to reduced latency and operational costs.
Hugging Face's Transformers library now integrates with llama.cpp quantization, enabling more efficient and accessible execution of large language models.
Simon Willison releases llm-typesafe, a new plugin for his LLM tool, adding support for TypeSafe AI's Jev model to enable structured yes/no, choice, and scoring questions.
This article explores the 'brighter than white' effect seen on HDR displays in Chrome 137+ and Safari 26+, discussing its visual impact and potential as a technological gimmick.
Simon Willison shares a quote from @therealcornpop highlighting how AI-generated content often lacks a definitive voice, genuine opinions, and exhibits predictable 'AI-isms'.
pnpm's latest minor release, 12.6, brings significant developer experience enhancements including automatic dependency deduplication and the ability to move node_modules directories.
Hugging Face highlights how UK AISI and EvalEval are collaborating to enhance the reproducibility of benchmark results in AI research, a critical step for reliable model evaluation.
Marie and Chris from CodePen discuss the history and technology behind CodePen Challenges, including their custom-built CMS and the benefits for the community.
Nathan Lambert hosts JS Denain of Epoch AI to discuss critical topics like AI's recursive self-improvement (RSI), the US-China AI gap, and the concept of 'jaggedness' in AI development.
A case study reveals how Parallel's agents leveraged GPT-6 Astra to research and synthesize labor-market data, cutting both time and cost by 50% compared to previous models.
OpenAI details its priorities and principles for rigorous, secure, and independent third-party assessments of frontier AI models and their safety safeguards.
