CodeQL 2.26 Now Scans Your AI Code for Prompt Injection
GitHub's CodeQL 2.26.0 detects prompt injection in JS/TS AI apps using OpenAI, Anthropic, and Google GenAI SDKs. Here's what changed and what to ...
Amazon Project Moonraker: What Developers Need to Know
Amazon's leaked Moonraker project aims to give Alexa compound task execution. Here's what the SDKs ...
Grok 4.5 Developer Guide: API, Benchmarks, and When to Use It
xAI's Grok 4.5 ranks #1 in agentic tool use at $2/$6 per million tokens. Here's ...
Oratomic’s $300M Bet: Fault-Tolerant Quantum at 20K Qubits
Oratomic raised $300M to build a fault-tolerant quantum computer with just 20K qubits, not one ...
LiteRT.js: Run AI Models in the Browser, No Server
Google's LiteRT.js runs .tflite models in the browser via WebAssembly — 3x faster than TensorFlow.js. ...
Hugging Face Kernels Are Now Signed Hub Artifacts
Hugging Face’s July 6 Kernels update introduces signed artifacts, a trusted publisher model, and a breaking repo-type change. Here is what your inference ...
Claude Apps Gateway: SSO and Spend Caps for Claude Code on AWS
Anthropic's Claude Apps Gateway is a self-hosted control plane that replaces scattered API credentials with ...
GPT-Live-1: OpenAI’s Full-Duplex Voice Model, Explained
OpenAI's GPT-Live-1 brings true full-duplex voice and a two-layer delegation architecture. Here's how it works ...
GLM-5.2: Open-Weight Model Beats GPT-5.5 at 1/6th Cost
Z.ai's GLM-5.2 beats GPT-5.5 on SWE-bench Pro, tops Design Arena, and delivers 1M-token context at ...
Codex App Server: Inside OpenAI’s Agent Harness
The Codex App Server is the bidirectional JSON-RPC layer that ties every Codex surface together ...