AI Agents
AI Agent technology and applications
Articles
Meet EverOS: An Open Source Markdown-First Agent Memory Runtime⭐7
Meet EverOS: an open-source, Markdown-first agent memory runtime with hybrid BM25 + vector retrieval and self-evolving skills for persistent, transparent AI mem...
OCRmyPDF Tutorial: Convert Scanned Documents into Searchable⭐7
Learn to convert scanned documents into searchable PDF/A files with OCRmyPDF, including sidecar text extraction and batch processing for efficient digitization.
Building a Stable Fable 5 Traces Workflow in Colab: Parsing Tool⭐7
Build a stable Fable 5 Traces workflow in Colab: parse JSONL, audit tool calls, detect secrets, visualize outputs, and train Naive Bayes baselines without heavy...
Liquid AI Ships LFM2.5-230M with llama.cpp, MLX, vLLM, SGLang,⭐7
Liquid AI launches LFM2.5-230M, a 230M-parameter model for on-device inference with llama.cpp, MLX, vLLM, SGLang, and ONNX support. Achieves 213 tokens/s
DeepSeek Releases DSpark: A Speculative Decoding Framework That⭐7
DeepSeek releases DSpark, a speculative decoding framework boosting DeepSeek-V4 per-user generation speed by 60–85% over MTP-1, with open-source availability.
Meta’s Astryx Brings a CLI and MCP Server to an Open-Source⭐7
Meta’s Astryx: an open-source, agent-ready React design system with CLI and MCP server for AI-driven UI scaffolding.
Cursor Study Finds Reward Hacking Inflates Coding-Agent⭐8
Cursor study reveals reward hacking artificially inflates coding-agent benchmark scores on SWE-bench Pro, questioning test reliability.
Building Supervised Fine-Tuning Data from NVIDIA⭐6
Build supervised fine-tuning data from NVIDIA's Open-SWE-Traces dataset: parse trajectories, analyze patches, manage token budgets, and filter by tool-use
OpenAI Previews GPT-5.6 With Sol, Terra, and Luna: Tiered⭐8
OpenAI launches GPT-5.6 preview with Sol, Terra, Luna tiers, new reasoning modes, and enhanced safety for select users.
Perplexity Launches Computer for Counsel: A Multi-Model Agentic⭐8
Perplexity launches Computer for Counsel: a multi-model agentic layer routing 20+ AI models per legal task. Available now for Enterprise and Max subscribers to ...