Full-text search across 67394 articles. Combine with topic and date filters; results sorted by relevance.
Software engineer and blogger Naoki Shibuya warns that developers who blindly trust AI-generated code risk becoming 'meat proxies' who add no value, and argues that critical thinking is the only differentiator in AI-assi…
A new agent pattern called Code Mode is replacing traditional tool-calling in LLM agents, cutting context window usage by up to 99.4% in a demo. Instead of streaming every tool definition and intermediate result through …
A developer outlines a design methodology for building Effect-based TypeScript applications, emphasizing a graph-first approach where the program structure mirrors a call graph of data flow. The approach defines domain n…
Affaan Mustafa, winner of the Anthropic Hackathon, has released Everything Claude Code (ECC), an agent harness performance optimization system containing over 108 skills, 25 subagents, 57 commands, and automated hooks an…
Cloudflare's AI code reviewer has flagged nearly 250,000 deviations from engineering standards and blocked 16,000 merges over the past four months, while its spec reviewer agent evaluated close to 600 technical designs. …
Google Cloud added immutable revisions and percentage-based traffic splitting to Agent Runtime by July 6th, enabling developers to release new agent code without replacing the running engine, resource name, or endpoint. …
Formula 1® and AWS have built the Data Accelerator, an agentic AI solution on Amazon Bedrock AgentCore, reducing data source onboarding from up to 8 weeks to approximately 40 minutes of code generation plus hours of depl…
Elon Musk pledged on July 24 that X would open-source its entire codebase by August 2026, but as of August 3 the release is blocked on an internal security review with no firm completion date. The pledge follows a scanda…
Trail, a new Python SDK, captures AI agent activity as signed OpenTelemetry spans, adding agent-aware tool taxonomy, MCP injection flagging, and skill hashing to detect silent substitution, while exporting via OTLP to an…
Danny Arends, a long-time D language developer, has built DLLM, a minimal coding agent in D that runs directly on llama.cpp without Python or bindings, using D's ImportC feature to access the C API with zero FFI overhead…
A developer's guide to using AI agents effectively emphasizes treating them as agents rather than chatbots, focusing on spec-driven planning, executable acceptance criteria, and setting budgets and finish lines. The auth…
DeepSeek V4 Flash, a 284 billion parameter open model, improved from roughly 7% to 54% on the DeepSweep benchmark through post-training, and now requires about 168 GB of VRAM for 4-bit quantization or 110 GB for 3-bit, t…
NVIDIA released nvmath-python v1.0, a Python library that provides access to NVIDIA CUDA-X math libraries for high-performance core math operations on CPU, single GPU, or distributed multi-GPU, multi-node systems. The li…
An engineer found that Claude Code's AI agent could fill every text field in a web form but could not attach PNG images, because file-upload capability is gated differently depending on the surface. The agent's file_uplo…
Google DeepMind CEO Demis Hassabis has quietly withdrawn Google from the AI race against OpenAI and Anthropic, betting on world models that understand and simulate the real world rather than the recursive self-improvemen…
JetBrains Research has open-sourced KotlinLLM, an IntelliJ IDEA plugin that adds Smart macros to Kotlin/JVM projects, allowing developers to delegate runtime logic to an LLM while generating persistent Kotlin source code…
A developer built GamesMom, a platform with over 50 educational browser games, using AI agents to accelerate development. The AI helped with planning, repetitive coding, debugging, documentation, SEO, and testing, but th…
NVIDIA Labs released the Object-Oriented Agents (NOOA) research preview, an open-source Python framework that treats AI agents as single Python classes to improve performance and developer workflow. The architecture iden…
A new AI-driven scheduling system for dairy processing plants uses real-time sensor data from silos, pasteurizers, and fill lines to continuously update production sequences, reducing losses from stale clipboard readings…
A new training method called replacement-aware training produces sparse auto-encoders (SAEs) that retain language capabilities when used in a full replacement model of Gemma-2-2B, unlike standard SAEs which cause catastr…