AI companies advertise ever-larger context windows for their language models, but a developer argues that the usable portion is far smaller than the marketed figures suggest. Garrit Franke writes in his blog Garrit’s Notes that effective model performance begins to degrade around 100,000 tokens, regardless of whether the advertised limit is 200,000, one million, or …
Huawei has unveiled HarmonyOS 7, a major update to its mobile operating system that connects users to more than 2,000 specialised AI agents. Iris Dengin reports for the South China Morning Post that the company positions the release as its entry into what it calls the “agent era.” Richard Yu Chengdong, chairman of Huawei’s Consumer …
Bot traffic has overtaken human traffic on the internet for the first time in history. Mark Tyson reports for Tom’s Hardware that Cloudflare CEO Matthew Prince publicly acknowledged the milestone, admitting he had not expected it to happen until 2027. “Welp, that happened faster than I predicted,” Prince wrote. According to Cloudflare data, bots now …
Poke, a startup that lets users interact with an AI agent via text message, has become the first third-party AI agent approved to run on Apple’s Messages for Business platform. Sarah Perez reports for TechCrunch that the approval marks a significant expansion of a platform previously reserved for businesses communicating with their own customers through …
Snowflake has introduced a two-layer context system designed to stop AI agents from returning contradictory answers when querying the same business data. Sean Michael Kerner reports for VentureBeat that the announcements came at Snowflake Summit 26 in San Francisco and center on two new products: Horizon Context and Cortex Sense. The core problem is that …
Microsoft has introduced a new security framework designed to contain AI agents running on Windows devices. Called Microsoft Execution Containers (MXC), the system lets developers and IT administrators define precisely what an AI agent is allowed to do — and enforces those rules at the level of the operating system kernel. Michael Nuñez reports for …
Microsoft Scout is an always-on AI agent that manages emails, calendar conflicts, and internal messages on behalf of office workers. Reece Rogers reports for WIRED that Microsoft announced Scout at its Build developer conference. Scout is built on top of OpenClaw, an AI tool that attracted early adopters in San Francisco at the start of …
OpenAI’s AI agent Codex has grown far beyond its origins as a tool for software developers. According to two new publications from OpenAI, the system now counts more than 5 million weekly active users and is increasingly used by professionals across industries to handle the routine, time-consuming work of modern office life. OpenAI writes and …
Odysseus is a new open-source project that lets users run a full-featured AI workspace on their own hardware. Developer pewdiepie-archdaemon publishes the project on GitHub as a self-hosted alternative to commercial AI interfaces such as ChatGPT and Claude, with a strong focus on privacy and local data storage. And, yes, that PewDiePie, the YouTube guy. …
A new generation of AI tools is reshaping how software is built and, according to their most devoted users, how work itself is done. Steven Levy reports for WIRED on the rapid rise of AI agents, tracing how two tools — Anthropic’s Claude Code and the open-source project OpenClaw — have ignited what many in …