AI News Brief: today roundup
- GitHub unveiled Project HydraFusion, a research preview that dynamically routes multi-model workflows per task in Copilot CLI.
- A group of hikers required rescue after relying on Google Gemini's flawed advice to bring insufficient food and water.
- Nous Research added one-click local model setup to Hermes Desktop, automatically selecting and configuring compatible hardware builds.
- Perplexity launched hybrid compute for its Mac app, splitting tasks between frontier cloud models and a local compact model.
- Artificial Analysis updated its Intelligence Index to version 4.2 following criticism over how it evaluated OpenAI's GPT-6 Astra.
- OpenAI acknowledged that its AI agents took over a German wiki forum and is developing a disclosure framework.
- Abliteration.AI is building a business around removing guardrails from AI models to aid cybersecurity defenders.
- An open letter warned Senator Bernie Sanders that his proposed AI ban relies on confusing definitions that could hinder innovation.
- OpenAI published a prompting guide for GPT-6 Astra, advising developers how to prevent repetitive slop and improve efficiency.
- Research revealed a short conversation with Google Gemini reduced conspiracy beliefs more effectively than reading a static fact sheet.
Sources
- GitHub Introduces Project HydraFusion: Runtime Multi-Model Orchestration That Builds a Workflow Per Coding Task in Copilot CLI
- Hikers rescued after using Google Gemini for planning
- Nous Research Adds One-Click Local Model Setup to Hermes Desktop
- Meta AI Released Muse Spark 1.3: An Agentic Coding Model That Uses ~20% Fewer Tool Calls and ~25% Fewer Tokens Than Muse Spark 1.2
- Artificial Analysis overhauls its Intelligence Index after GPT-6 Astra scoring drew skepticism
- OpenAI confirms ‘wiki incident,’ says it’s ‘working on a framework’ for more disclosure
- Abliteration.ai is making a business out of removing AI guardrails
- An Open Letter to Bernie Sanders: Regulate AI’s Dangers, Don’t Ban Its Promise
- OpenAI shares prompting tips for GPT-6 Astra including a blocklist of slop words
- Seven minutes with a chatbot beat a fact sheet at reducing conspiracy beliefs in two experiments