Reflection released Beam, its first open-weight model, a mixture-of-experts system that activates 23 billion of its 501 billion parameters per token and aims to match GLM 5.2 on coding and reasoning w…
Deepseek and Huawei have released open-source programming tools for Huawei's Ascend AI chips, centered on TileLang, a language designed to offer a simpler programming model than Nvidia's CUDA. The par…
An anonymous developer built Fukurama, an AI frontend designed for iterative, branching work rather than linear chat. The tool lets users branch off from a conversation, work through problems in isola…
Pen.dev launched as a WebGL-based design canvas that lets users import any part of any app and design with AI agents directly on the canvas. The tool supports ChatGPT, Claude, Grok, Qwen, Gemini, Kimi…
The Trump administration has few effective options to slow China's rise as an AI superpower, according to Bloomberg, as it weighs export controls against Beijing's retaliatory leverage over rare earth…
Deepseek released V4.1-Flash, a multimodal model with 552 billion parameters that cuts KV cache memory to a quarter of its predecessor's requirement, with only 16 billion parameters active per token. …
Google Analytics 4 has introduced a native AI Assistant channel in its Default Channel Group, effective May 13, 2026, giving qualifying visits a dedicated medium, channel grouping and campaign value. …
A Medium blog post by an unnamed author promotes OpenCode, an open-source coding tool, as a free alternative to Anthropic's Claude, claiming it offers unlimited searches and a large context window usi…
XBin, an open-source project at xbin.dev, introduces a self-modifying sandbox system that separates code and runtime environments to let AI agents build and manage mini-apps with strict permissions, e…
Open-source and free AI models are becoming nearly as capable as frontier American models, with local models like Ornith 1.5 9B running on consumer hardware and free API models from Nvidia and others …
A developer used an AI agent to reverse engineer every peripheral within arm's reach, gaining full control over devices such as a microphone, webcam, and key light, and documented the process. Another…
In the latest round of LLM benchmarks, Deepseek v4 models scored 93 points, placing them in the middle of the Tier A pack, with Flash 0731 completing the test in 43 minutes and Pro 0813 in 48 minutes,…
Chinese AI company Deepseek released Deepseek-V4-Flash-Vision-Exp, an experimental multimodal model that adds image understanding to its text capabilities, and on Deepseek's internal benchmarks it nea…
Frontier AI labs Anthropic and OpenAI are pursuing a lock-in strategy to retain users and recoup billions in training costs, evolving from model-level to organizational workflow-level integration, wit…
Deepseek has moved its flagship V4-Pro out of the testing phase and released its agent software, Harness v0.1, under the MIT license. API prices are going up at the same time, with cache hits jumping …
Lethe, a portable identity layer for AI applications, launched in solo beta on Hacker News, enabling users to transfer context and understanding across different LLMs such as ChatGPT, Deepseek, Grok, …
A developer has released an open-source, 100% native Swift harness for macOS, claiming it is the first non-Electron alternative to tools like Cline, Codex, and Claude Code. The project, licensed under…
Deepseek released V4 Flash "0731," scoring 50 on the Artificial Analysis Intelligence Index, ten points above the previous version and one point behind OpenAI's GPT-5.6 Luna, while costing about 60 pe…
A Hacker News user asks the community for advice on building a cheap AI coding agent setup using a BYOK open source client to avoid vendor lock-in, specifically inquiring about Deepseek and Kimi model…
Sriram Krishnan, former Senior White House AI Policy Advisor, warns that American frontier models are losing the open-source race to China, citing the rapid release of Chinese models like Kimi K3, Dee…