Notes from the trenches on AI engineering, LLM apps, and the full-stack work that holds it all together.
pxpipe is a local proxy that renders dense system prompts and old history as PNGs before they hit Claude Code. Real workloads report 59 to 70 percent lower bills, with a lossy catch you need to understand.
Maxime Rivest's open-source Riddle app sends handwritten pages to a vision LLM and animates replies on e-ink. No chat box. Just pen, pause, and flowing script.
Shanghai AI Lab's 35B MoE agent reaches trillion-parameter benchmark territory by scaling trajectory length to 45K tokens and distilling six domain teachers. Here's what agent-horizon scaling actually means.
Anthropic launched Claude Science in beta on June 30: a macOS and Linux desktop app with 60+ database connectors, live code execution, HPC orchestration, and full provenance on every artifact. Here's what it actually does.
Anthropic redeployed Claude Fable 5 globally on July 1 with a new cybersecurity classifier, Opus 4.8 fallbacks, and a cross-industry jailbreak severity framework. Here's what actually changed for developers.
Nous Research's Judgment Release adds completion contracts, a coding verification evidence ledger, selectable Mixture-of-Agents, and a zero P0/P1 backlog sweep. Here's what changed for production agent workflows.
LangBot is an open-source, production-grade platform for deploying AI agents across Discord, Slack, Telegram, WeChat, Lark, DingTalk, and more. Here's how it wires LLMs, RAG, and n8n workflows into real chat channels.
Nemotron-Labs-TwoTower splits a 30B Nemotron backbone into a frozen context tower and a trainable diffusion denoiser, hitting 2.42x generation throughput with open weights on Hugging Face.