A new locally runnable Cyber-OSINT model suite is being announced: a Mixture-of-Experts (MoE) system with a 26B-parameter total size, 4B active parameters, and an extended 262K-token context window, fine-tuned on about 6,500 OSINT/CTI-style instructions. It is supervised-fine-tuned specifically for cyber threat intelligence and investigative workflows, and is presented with task-level capabilities such as threat-actor attribution, IOC pivoting, geolocation extraction, and Admiralty-style source reliability grading. The architecture and training emphasis aim to embed operational tradecraft and analytic heuristics into model outputs rather than rely on a static system prompt.
A complementary lightweight offering is a 7B cyber model engineered to run on an 8 GB GPU, described as a true fine-tune (not just a system-prompt wrapper) using post-training cyber/DevOps datasets for both offensive and defensive use cases. That smaller model supports a 32K native context and can be extended to ~131K tokens with YaRN. Both models are published to Hugging Face repositories (references include DeepHat/DeepHa… and terrorswift/REDCELL-26B-A4B-OSINT-Cyber-APEX-GGUF), suggesting immediate availability for local experimentation by practitioners who need long-context, task-focused OSINT/CTI tooling.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
OpenAI and Synopsys have announced GPT-Synopsys, a frontier intelligence platform designed to revolutionize chip design. The collaboration aims to enhance AI-driven automation and optimization in semiconductor development. (news.synopsys.com)
The US government's new AI chatbot has refrained from debunking claims made by Trump. The chatbot is currently experiencing connection issues that prevent full functionality. (techxplore.com)
OpenAI reduced the usage allowance for the $200 plan from 20x to 10x the ChatGPT Plus limit starting October 30, 2026. A new Pro 500 tier offers 25x the usage of Plus for users needing more capacity. (news.ycombinator.com)
Vectorless, reasoning-based retrieval-augmented generation (RAG) is introduced by VectifyAI to improve AI reasoning without relying on vector embeddings. The approach enhances reasoning capabilities in AI models, as detailed in their GitHub repository. (github.com)
The video features interviews with AI CEOs in 2026 discussing the future of artificial intelligence and its impact on industries. It highlights advancements in AI leadership and technological integration. (youtube.com)
A GitHub project involving 'torturing' large language models has sparked a heated debate about AI consciousness and model welfare. Experts agree that LLMs are not sentient, but some AI safety advocates worry about their 'mental health' as models become more advanced. (404media.co)
Today's best Hacker News stories, summarized and screenshotted, one email a day.