OpenAI has opened orders for Codex Micro, a limited-edition desktop keypad built to monitor and control AI coding agents. Megan Morrone reports for Axios that the device is a collaboration with Work Louder, a boutique hardware maker known for customizable keyboards for developers and designers. The small macro pad sits next to a regular keyboard. …
Cursor, the coding tool used across much of the Fortune 500, is quietly building a general-purpose AI agent for people who never touch code. Eloise Jones reports for TechTimes that the project, internally codenamed Sand, began rolling out to Cursor employees in late June, running on infrastructure leased from SpaceXAI. Sand is designed to answer …
Craig Mod talks with Dan Shipper for Every’s YouTube channel about rebuilding his entire software stack with AI models such as Anthropic’s Claude, while drawing a hard line at using the technology to write. Mod, known for the newsletters Roden and Ridgeline and books including Kissa by Kissa, used to pay around $67,000 a year …
Canva has launched Canva Code 2.0, a major update to its AI coding tool that lets anyone build interactive websites and apps from plain-language prompts, then edit the result like a normal Canva design. Michael Nuñez reports for VentureBeat that the feature is now open to all 265 million monthly users, including those on Canva’s …
A new open-source framework called Agent-as-a-Router promises to make AI model selection smarter and cheaper. Ben Dickson reports for VentureBeat that the concrete implementation, called ACRouter, beats both static routing systems and the common strategy of defaulting to premium models like Claude Opus. Companies use model routing to send simple tasks to cheap, fast models …
Anthropic has found that its Claude chatbot consistently expresses different values depending on which model version users choose and which language they speak. Jason Nelson reports for Decrypt on a new study in which Anthropic researchers analyzed 309,815 anonymized conversations involving subjective tasks such as giving advice or feedback. The company distilled more than 3,300 …
Three major AI developers released new models within days of each other this week, and all of them lead with the same pitch: lower cost per task. Lorelei Smillie and Rachel Metz report for Bloomberg that OpenAI’s GPT-5.6, SpaceXAI’s Grok 4.5 and Meta’s Muse Spark 1.1 all promise more output for fewer tokens, the units …
A year after Meta overhauled its AI division following the troubled Llama 4 launch, Max Kan and his co-authors write for SemiAnalysis that Meta Superintelligence Labs (MSL) may be closing the gap on OpenAI and Anthropic. The authors argue Meta is the only company positioned to excel at all three ingredients of frontier AI: data, …
Journalists working with AI tools risk losing certain foundational skills while developing new ones, according to a new study published in Journalism Practice. Andrew Deck reports for Nieman Journalism Lab that researcher Shangyuan Wu of the National University of Singapore interviewed 14 working journalists in Singapore to explore this shift. The study builds on last …
Meta has shut down an AI image feature that let users manipulate photos of people by simply tagging their public Instagram account, just days after launching it. Corbin Bolies reports for Variety that the company faced immediate backlash over the tool’s design. The feature was part of Meta’s new Muse Image model, introduced on Tuesday. …