84 stories from 15 sources
The week in AI and Technology
Click on a headline to see stories that shaped it.
Threads
6 threads
AI infrastructure demand remained powerful, with memory and chip investment supporting the boom, but Nvidia's server delay exposed supply-chain fragility. Growing opposition to data centers added power, water, land, and permitting constraints to the expansion.
Meta expanded Muse Image from generation inside its apps to prompts based on public Instagram handles, while making consent controls difficult to find. Backlash over public-account content, likeness, and harmful alterations then forced the company to disable the feature.
OpenAI moved from announcing GPT-5.6 to embedding it in Copilot and adding an agent that can complete extended, multi-step tasks. It also folded Atlas's browsing capabilities into existing products, favoring integrated workflows over a standalone browser.
AI companies and startups directed new capital and products toward enterprise integration, agents, and workplace automation rather than model quality alone. Rising token costs and growing open-source adoption showed that buyers are also weighing economics, deployment control, and customization.
OpenAI's product leadership shifted as Fidji Simo departed and Greg Brockman took direct control, while the safety organization also lost a senior leader. Apple's trade-secret lawsuit and the ensuing public accusations added legal and partnership risk to the company's AI-device ambitions.
AI-enabled abuse advanced from generating industrialized phishing campaigns to exploiting hallucinated software packages for malware distribution. Together, the cases show how language models can lower expertise requirements across separate stages of an attack.
Day’s Recap