THE DAILY BRIEF

The Manhattan District Attorney's Office seized 12 websites operated by a single firm that collectively targeted approximately 1,200 victims with deepfake content, marking the largest-ever legal action against harmful deepfake platforms. The sites targeted customers of major AI companies including OpenAI, Anthropic, and Meta.

► This demonstrates law enforcement is moving aggressively against deepfake infrastructure, setting precedent for how governments will prosecute AI-enabled fraud at scale.

Surveys by Epoch AI and Ipsos show the share of U.S. adults using AI almost daily more than doubled between March and August 2026, rising from 8 percent to 19 percent, while weekly-only usage fell from 17 percent to 10 percent. AI is rapidly becoming embedded in everyday workflows.

► This velocity of adoption means consumer-facing AI tools are moving past novelty into necessity, reshaping baseline expectations for productivity software.

OpenAI used its own large language models to help design its Jalapeño chip, demonstrating a closed-loop approach where AI systems accelerate hardware development. The story generated significant technical interest on Hacker News with 116 points and 83 comments.

► Vertically integrating AI model development with custom silicon design could give OpenAI significant cost and performance advantages over competitors relying on third-party chips.

Runway is developing technology to stream AI-generated video in real time as users prompt it, rather than requiring users to wait for finished clips. The approach builds on GWM-1, its world model that generates video frame by frame, with potential applications in robotics and autonomous driving.

► Shifting from batch processing to live streaming fundamentally changes the user experience and opens new use cases like interactive robotics control where latency matters.