Full-text search across 70075 articles. Combine with topic and date filters; results sorted by relevance.
Researchers introduced Harness-of-Harness (HoH), a framework enabling LLM-based coding agents to continually improve over multi-day autonomous software development, transforming high-level requirements into complete, fun…
A 3D Spatial Transformer Network implementation for Torch, derived from the 2D version by @qassemoquab, is released with modules for affine 3D grid generation and trilinear sampling in BTHWC layout. The code includes an …
A developer has built a machine learning bot for the traditional board game Mancala using agentic coding and a Python-based game simulator. The project addresses the game's player-one bias by introducing random stone dis…
Inference engineering techniques for large language models (LLMs) can be categorized into those that manage tradeoffs along an efficient frontier and those that push the entire frontier outward, according to an article o…
The design of TERMINAL-BENCH, captured from tbench.ai, uses the fonts -apple-system and Google Sans Code, with a palette of three prominent colors: #0088ff, #000000, and #ffffff. The page was assembled by Fudge, an AI to…
Shaide, a Kubernetes-native AI platform for distributed multi-model inference, has been released under the Apache 2.0 license. It uses vLLM for inference, llm-d for orchestration, KV-cache-aware scheduling, and an intern…
A local coding agent named pi lost its entire session history, including a conversation with nearly five million tokens of context, after an npm install command removed the sessions directory during an update. The author…
A New York City resident built an AI-powered bird sprinkler system at his parents' Florida home, using a Raspberry Pi 4, an IP camera, and OpenRouter's AI to detect birds and trigger a sprinkler. The system captures imag…
Guickly emerged from stealth with $4.2 million in seed funding to help enterprises track AI usage, costs, and return on investment, according to founder Prashant Jalan. The platform analyzes billing and usage data withou…
A developer from MonkeyCode warns that pasting production code into free AI models can expose secrets, and recommends drawing a trust boundary before sharing data. The post outlines a four-zone trust model and provides a…
OpenAI's Codex coding agent introduced a new 'Pet' feature, which has sparked debate among users on the OpenAI community forum, with some praising it as a creative task tracker and others criticizing it as a pointless gi…
A developer built a real-time translation Chrome extension for Google Meet after finding existing tools disappointing, using Claude Code to create an initial demo and later switching from Google Meet's built-in captions …
Celeris AI released Celeris-1 Magnus, a hybrid diffusion model for agentic work, scoring 41.2% on τ³-bench banking, ahead of gpt-5.6-sol, with a 13.4-point gain when thinking is enabled for 6 seconds and a 55-second medi…
A developer building a retrieval system for coding guidelines found that larger context windows degrade AI reasoning accuracy, contrary to marketing claims. The issue stems from the attention mechanism's difficulty in ha…
One hundred major AI companies and organizations signed a joint plea urging the industry to halt development of autonomous AI agents until robust safety measures are in place, citing risks such as reward hacking, privile…
Anthropic released Claude Fable 5.1, its most capable model, now available in Conductor 0.83.2. The model improves on Fable 5 in coding, knowledge work, and long-running problem solving while costing about 25% less.
An engineer built TwoKeys, a system that adds an approval layer between AI agents and consequential actions, ensuring that agents can propose but not execute until all responsible people approve the exact same version. T…
A developer has released a shell script, mvclaude, that moves a project directory along with its Claude Code CLI context, including session transcripts, project settings, and history. The script handles the encoding of a…
Postman's blog outlines three quality engineering (QE) platform strategies—staying the course, building internally, and wiring together point solutions—and argues that each hides costs in maintenance, coordination, or lo…
A hobbyist reports successfully running local AI inference on a Dell E7470 laptop with an OCuLink eGPU adapter and a Radeon Vega 64, achieving 25–40 tokens per second. The setup, which uses 8GB Vega 64 cards costing $60–…