How AI is reshaping society, people, and work
13 articles
Loft Orbital's YAM-9 satellite ran Google's Gemma 3 vision-language model directly in orbit, analyzing Earth imagery in real time without downlinking raw pixels. It marks the first publicly confirmed execution of a VLM in space.
Nobel Prize winner John Jumper and three other senior DeepMind researchers are joining Anthropic and OpenAI, intensifying doubts about Google's long-term AI competitiveness.
NAVER and NVIDIA announced a full-stack AI factory partnership centered on South Korea's GAK Sejong data center, starting at 55 megawatts in early 2027 with a long-term target of 1 gigawatt. The infrastructure will power next-generation HyperCLOVA X models and NAVER's Seoul World Model.
Announced at Volcano Engine FORCE on June 23, Seedance 2.5 generates a native, un-stitched 30-second 4K video clip in a single pass, with support for up to 50 reference inputs and localized frame editing.
Meta debuted its own-brand AI smart glasses starting at $299 — cheaper than its Ray-Ban collab line and powered by Muse Spark, the first model from Meta's Superintelligence Labs.
John Jumper, who shared the 2024 Nobel Prize in Chemistry for his work on AlphaFold, is leaving Google DeepMind to join Anthropic. Coming days after Gemini co-lead Noam Shazeer departed for OpenAI, the back-to-back exits sent Alphabet shares down as much as 7.2%, erasing roughly $250B in market cap.
Google signed a deal to rent ~110,000 NVIDIA GPUs from SpaceX's Tennessee data center at $920 million per month through June 2029, mirroring Anthropic's parallel $1.25B/month agreement with the same facility.
At Build 2026, Microsoft unveiled the MAI (Microsoft AI) model family — seven purpose-built models spanning reasoning, coding, image, voice, and transcription — trained from scratch on commercially licensed data with no distillation from OpenAI or any third party.
The Federal Energy Regulatory Commission unanimously ordered six major U.S. grid operators to overhaul how they process AI data center interconnection requests, giving them 60 days to comply or face regulatory consequences.
ElevenLabs released Dubbing v2, an AI dubbing model that analyzes the original speaker's performance — emotion, pacing, intonation — and reproduces it faithfully across 90+ languages without going through a text transcription step.
JPMorgan Chase has moved its AI spending out of discretionary R&D and into core infrastructure alongside data centers and fraud-detection systems, committing $19.8B in total tech spending for 2026.
OpenAI shut down its AI video generation service Sora on April 26, 2026, just six months after its public launch. The service reportedly cost $15 million per day to operate while generating only $2.1 million in total revenue.
NVIDIA began mass production of its Blackwell Ultra B300 chip on June 12. The GPU ships with 288 GB of HBM3e memory and targets 50x higher throughput per megawatt versus Hopper — purpose-built for the agentic AI era's inference and reasoning demands.