Tech News
Show HN: Bor β Open-source policy management for Linux desktops
Bor is an open-source system for centralized Linux desktop management, featuring a Go agent and server that stream policies over mTLS/gRPC in real time.
July 2026 newsletter
The June edition of my sponsors-only monthly newsletter is out. If you are a sponsor (or if you start a sponsorship now) you can access it here.
Announcing the Agentic Catalog Experience in Amazon Quick
Amazon Quick introduces the Agentic Catalog Experience, an AI-powered workflow for data curators to discover upstream catalog assets in natural language and auto-create Datasets and Topics with inherited semantics.
Co-Designing AI Model Attention for Fast, Interactive Long-Context Inference
As agentic and long-context workloads become common, the context lengths increase and attention consumes a larger share of inference time.
Show HN: NixOS-DGX-Spark β Nix and NixOS on the DGX Spark
A new project enables running NixOS on NVIDIA DGX Spark and Asus GX10, with users reporting success running k3s and DeepSeek models.
GitHub Repos
open-multi-agent/open-multi-agent
TypeScript AI agent orchestration framework with dynamic workflows.
# TypeScript# agent framework# agent orchestrationMemPalace/mempalace
The best-benchmarked open-source AI memory system.
# Python# ai# chromadbcft0808/edict
ποΈ δΈηε ι¨εΆ Β· OpenClaw Multi-Agent Orchestration System β 9 specialized AI agents with real-time dashboard, model config, and full audit trails
# Python# ai agents# ai orchestrationResearch Papers
ExtractBench: A Benchmark for Schema-Guided Enterprise Document Extraction
ExtractBench is a new benchmark for evaluating how well AI agents extract structured data from enterprise documents according to user-defined schemas, measuring accuracy, completeness, grounding, and cost.
RecHarness: A Bandit-Routed Agentic Harness for Self-Evolving Recommender Systems
RecHarness is a system that uses a bandit algorithm to choose which part of a recommender model to modify, while an LLM generates the actual code changes. It shows stable improvements in offline and online tests, and the code is available.
TransMem: Transforming Hidden States into Memory for Large Language Models
TransMem is a new method that helps AI agents remember useful information from long conversations by converting hidden states into reusable memory, without reprocessing everything. It uses a lightweight network to apply these memories on the fly, making it faster and more...