AI Briefing
GeekNews news
The latest AI and developer news from GeekNews, with short summaries.
Visit website →'pi-subagents' extension released for delegating subagents in the Pi coding agent
An extension has been released that lets the Pi coding agent delegate tasks to sub-sessions, enabling parallel processing and review from diverse perspectives.
Ornith-1.0, a Self-Improving Open-Source Model for Agentic Coding, Released
Ornith-1.0, based on Gemma 4 and Qwen 3.5, has been released as a self-improving open-source model for agentic coding.
Tidal's AI Policy
Tidal has introduced a new policy banning monetization of AI-generated music to protect artists.
How to Calculate Compute-Adjusted LTV
This explains a new economic metric that supplements SaaS's existing LTV metric by reflecting the variable inference costs of AI products.
Go Micro - Agent Harness for Go
A framework for building AI agents and orchestrating services in Go has been released.
The History of Memory Prices from 1960 to 2026
A dataset visualizing historical price trends for major memory and storage types like DRAM, NAND, and HBM, along with the cost structure of AI accelerators.
Asian AI Startups Launch Alternatives to Anthropic Mythos
With U.S. export controls limiting access to Anthropic's models, Asian companies are launching alternative models to capture the market.
DSpark: Speculative Decoding Framework Combining Semi-Autoregressive Generation and Confidence Scheduling Unveiled
DSpark unveiled a speculative decoding framework that combines semi-autoregressive generation with confidence scheduling to solve the sharp acceptance rate decline problem of parallel drafters.
Open-Source Project Management Tool 'Paca' Unveiled to Integrate AI Agents as Team Members
The open-source project management platform 'Paca' has been unveiled, incorporating AI agents as equal members of Scrum teams.
An Open Letter on Akrites, a Joint Effort to Protect Open Source Security
Major IT companies have launched Akrites, an allied security project, to counter AI-driven vulnerability attacks.
Trump Administration Expands Vetting of Companies Seeking Access to Cutting-Edge AI Technology
The Trump administration is tightening regulation on Silicon Valley by expanding its vetting of companies seeking access to cutting-edge AI technology.
auth.md — An Open Protocol for Agents to Register on Behalf of Users
auth.md, an open protocol that helps AI agents sign up for services on behalf of users, has been released.
Meta Suspends Employee Tracking Program After Internal Data Leak
Meta temporarily suspended a program after an internal security incident exposed employees' keystroke data collected for AI training.
Krea 2: Technical Report on the Open-Weight 12B Image Model
Krea has released Krea 2, an open-weight, 12B-parameter image generation model optimized for creative exploration and control.
Show GN: How to Make an AI Agent Your Own
Introducing Monet, which provides local-based structured memory to maintain context for coding agents.
Agentic Testing - The Role of Agents in the E2E Test Stack
An analysis of how agent-based E2E testing differs from traditional deterministic testing and how it can be integrated into existing stacks.
ax-grep Improves Token Usage 3x with Accessibility Tree-Based Web Research
ax-grep reproduces the browser's accessibility tree to improve agent web search token usage by 3x and reduce memory usage by 15x.
Claude Code Releases Full-Sentence LLM Proofreading Skill to Overcome Limits of grep-Based Correction
A proofreading skill that has the LLM directly read every sentence has been released to fix grep-based correction errors in Claude Code.
Manipulating AI Search Using Reddit Is Too Easy
A study has found that simply manipulating short texts on UGC sites like Reddit can steer AI agents' answers toward spam.
opencodex — A local proxy for using any LLM with Codex
It's a local proxy tool that converts the OpenAI Codex protocol into other LLM APIs.