Llama

Ollaya is a project being discussed on Hacker News, described as 'Ollama for open-source, Jev-style decision models'. The framing suggests a tool that makes decision-making models as easy to run locally as Ollama made large language models, though the single post title gives little detail. With 537 likes and a high rank, commenters appear interested in the analogy to Ollama, but the posts collected do not explain what the tool actually does or why it is generating attention.

A developer has published a parameter-by-parameter walkthrough of their llama.cpp setup for running the Qwen 3 8B model locally, explaining what each command-line flag does and how the options are tuned for maximum performance on their hardware. The guide is aimed at people running AI models on their own machines, where cryptic command-line options often make local inference setups hard to understand and reproduce.

DarkMailr version 1.1.0 has been released, bringing a complete redesign and several new features. The update introduces a black-and-white interface, live Ollama model selection, persistent form state, and reset and copy buttons. It also adds separate first and family name fields and an improved prompt for generating phishing emails, a capability that is likely to draw scrutiny from security professionals.

A developer is testing Bespoke Nimble, a small decision model run locally through Ollama. The model takes evidence, a question and a set of allowed answers at request time, meaning the same model can handle many classification tasks without retraining. The author is comparing it with another model called Jev in a write-up, and interest centres on whether compact local models can replace task-specific trained classifiers.

Meta's AI assistant Muse has entered a partnership with Shopify to expand AI-powered purchasing and store management support, starting in North America. The tie-up will let merchants use agentic AI tools to assist customers with buying decisions and run online stores more efficiently. The news is being shared in Japanese business and tech circles, with attention on how Meta's AI, built on its Llama models, will compete in the fast-growing AI commerce space.

Spanish directing duo Los Javis appeared on laSexta's 'Cara al show' with Marc Giró, where they revealed which actress they would like to work with, saying she has 'never called us'. The interview comes as they enjoy the success of their latest production, 'La bola negra'.

Meta is developing a new AI agent called Muse that ties together its smart glasses and metaverse ambitions, alongside continued work on Llama 4 and the use of secure virtual machines. The project is attributed to Mark Zuckerberg's broader push to merge Meta's hardware and AI efforts. The report, circulating on cybersecurity channels, also touches on the security architecture behind these systems.