Rest of World

DeepSeek's R1 model challenged top U.S. AI systems last year, but OpenAI says the Chinese startup accessed its models to distill outputs and imitate performance https://lnkd.in/eruTV_qd

State aid explains only a fraction of BYD’s cost gap with Tesla; the real advantage is scale, lower overheads, and tighter control of the supply chain https://lnkd.in/eMTHm3KG

How cheap Chinese phones catapulted Kenya into the global digital economy https://lnkd.in/euQPqsjH

The Chinese government is using AI to bring traditional Chinese medicine into the future, with automated diagnostics, custom prescriptions, and international ambitions https://lnkd.in/e9hbXitJ

Dubai is betting big on experimental transport from U.S. startups, even though similar bets in the past collapsed without a single ride https://lnkd.in/ejZSHwvz

OpenAI, Meta, Google, Microsoft have invested billions in India. Mandatory licensing may force them to rethink AI deployment in one of the world's biggest markets https://lnkd.in/enWf9ctY

The next frontier in AI infrastructure may be orbit 🛸🪐🛰️ Without shared governance or ownership, space-based data centers could leave many countries producing data — but with little control over how it’s used https://lnkd.in/eXtGFJWn

Dubai is becoming a test bed for experimental transport. After failed bets on Hyperloop and robotaxis, the city is again backing technologies that don’t yet operate at scale https://lnkd.in/e4kngtBm

Millions of people around the world power today’s AI systems. Most never see where their work ends up. An investigation by Rest of World, the Bureau of Investigative Journalism, and the Pulitzer Center’s AI Accountability Network found that gig workers in Africa helped produce data for an Australian company, Appen, whose work included projects linked to a secretive U.S. military unit. Appen recruits a global workforce of more than a million people to transcribe audio, translate text, and label data used to train AI systems. Workers interviewed for this story — including some in refugee communities — described opaque projects, low pay, and little information about who the client was or how their work would be used. https://lnkd.in/eGNmQiuN

"For reassurance, experimentation, and low-pressure decisions, AI agents can feel helpful. For bigger choices, I still trust myself more." https://lnkd.in/edryCNgF