How Polaris Office's metric-compatible font generator dodges font licensing pain and what an adjacent LLM-as-Judge rubric thread reveals about production AI evaluation
How the public-apis curated free API list is being rediscovered as the inventory layer for the MCP catalog era
A walk through Microsoft's six-paper LLMLingua line from EMNLP 2023 through CoLM 2025 covering prompt compression KV-Cache optimization long-context acceleration and why production adoption lags the research
A digest of three arxiv papers surfacing within days of each other read along three axes — cooperation persistence and structure
How OpenAI runs Realtime voice traffic on Kubernetes by separating a stateless global UDP relay from stateful WebRTC transceivers
OpenAI Codex moves inside ChatGPT plans, unifying CLI, app, IDE, web, and a new Python SDK under one ChatGPT account — with policy clauses and workflow recommendations
How the EKS community kept container escape at bay for six days while AWS finalized an AMI patch for the algif_aead Copy-Fail vulnerability
gc-tree is an above-the-repo global context store for AI coding tools — switch context lanes like Git branches and inject only ~4% of total context per query
Simon Willison ran his SVG pelican prompt through 21 quantized variants of IBM Granite 4.1 3B and found 1.2GB and 6.34GB land at the same wall — what an informal benchmark reveals about quantization curves
KakaoTalk has restricted raw AI-generated emojis since 2023-09. A YouTube deep-dive spells out the loophole that still works — AI for concept and dialogue, manual editing for the final image.
