AI Briefing
GeekNews news
The latest AI and developer news from GeekNews, with short summaries.
Visit website →Looking Productive at Work
This analyzes how generative AI separates output from actual competence, leading to 'AI slop' and declining judgment within organizations.
Y Combinator's OpenAI Stake (0.6%?)
It has been raised that Y Combinator holds approximately 0.6% equity in OpenAI, and that this warrants disclosure of interest.
Zuckerberg "Personally Approved and Actively Encouraged" Meta's Copyright Infringement
Meta and Zuckerberg have been sued over alleged copyright infringement in training Llama.
Today I made the difficult decision to reduce Coinbase's size by about 14%
Coinbase is cutting 14% of its workforce and reorganizing into an AI-native organization in response to AI-driven changes in work productivity.
Computer use is 45x more expensive than a structured API
For the same task, the computer use approach was 45 times more expensive than the structured API.
Three Inverse Laws of AI
It presents three principles that humans must follow in order to use generative AI safely.
Gemma 4 MTP Was Hidden, Then Community Dug It Up, Then Google Belatedly Added a Workaround
Google excluded Gemma 4's MTP feature from the public release, then began supporting it via an auxiliary model after the community discovered it.
When everyone has AI but the company still learns nothing
For individual AI productivity to translate into organizational performance, a new agent-centric operating system and workflow changes are needed.
Accelerating Gemma 4: Faster Inference with a Multi-Token Prediction Drafter
Google released an MTP drafter for Gemma 4 that boosts inference speed by up to 3x.
Google Chrome silently installed a 4GB AI model on devices without consent
Evidence suggests Chrome quietly downloads a 4GB Gemini Nano onto hardware-eligible devices.
Bun (JS runtime) is being vibe-ported from Zig to Rust
An experiment using AI agents for Bun's Zig-to-Rust port has been revealed.
Agent Skills
Agent Skills has been released as workflow scaffolding for AI coding agents.
Show GN: Memex - a local RAG MCP server that infers semantic relationships between notes and auto-injects them into Claude's context
It's a RAG server that uses local embeddings and MCP to identify semantic relationships between notes and automatically inject context into Claude.
How LLMs Distort Our Written Language
LLM editing changed the conclusions, stances, and voice of writing, distorting its meaning.
Understand-Anything - A Plugin That Turns Code/Knowledge Bases into Interactive Knowledge Graphs
A plugin that turns code and knowledge bases into interactive knowledge graphs has been released.
What It Actually Takes to Run Code on a €200 Million Supercomputer
It explains the architecture of Spain's MareNostrum V supercomputer and how it actually operates using SLURM.
Using an "Underdrawing" for Accurate Text and Numbers
It proposes a method of first creating an underdrawing with SVG/HTML to improve the text and number accuracy of image generation models.
Redis array: a short story of a long development process
A Redis developer revealed a professional workflow that used AI to design and implement a new Array data type.
Blue Ocean Opportunities in the Agent Economy
As AI shifts from chat-based interfaces to autonomous execution agents, the SaaS and investment landscape is expected to be fundamentally reshaped.
How OpenAI delivers large-scale, low-latency voice AI
OpenAI redesigned its WebRTC architecture to achieve large-scale, low-latency voice AI.