News and context for the technology shaping everyday life.

OpenAI’s debut hardware is a $230 mini-keyboard that surfaces the status of multiple Codex threads at a glance, underscoring how agentic AI is pushing demand for human oversight tools.

Google’s court-backed remedies in the Epic case could shift Android app distribution in the U.S., changing how developers reach users and how much control Play Store holds over the ecosystem.

Apple says a former engineer used a security bug to access confidential files after leaving for OpenAI, turning a personnel dispute into a wider test of AI trade-secret protections.

The search startup has reportedly built a model-agnostic coding agent, codenamed Teammate, that its engineers have been using internally since May.

Apple’s court fight over the EU’s Digital Markets Act hit a setback as judges backed the bloc’s designation of the App Store and iOS as key services subject to competition obligations.

The French startup, backed by notable AI and startup figures, is releasing ZML/LLMD as a free way to explore faster inference across Nvidia, AMD, Google TPU, Apple Metal and Intel Arc hardware.

A new open-source project from a former Vanta employee proposes append-only, hash-chained records for AI agents so customers can verify what happened at runtime.

If you fancy a Windows 11 layout, but want to use Linux, you're in luck.

The EU is dealing with two different “Chat Control” tracks at once: a temporary scanning regime that expired and is now being revived, and a separate permanent proposal that is still stuck in negotiations.

A blog post on Beagle SCM argues that the best way to use large language models in developer workflows may be to surround them with deterministic tools that steadily take over repetitive tasks.