Tuesday, February 24 · ~45s read

Hey 👋 Here's your signal for today:

🔥 Anthropic Accuses Three Chinese Labs of Stealing 16 Million Claude Responses
Anthropic says DeepSeek, Moonshot, and MiniMax used roughly 24,000 fraudulent accounts to extract over 16 million Claude outputs at scale — then used them to train competing models. The claim immediately sparked a fierce backlash on X, with critics pointing out the irony of an AI company crying foul over training data practices, given how large models (including Claude) were themselves trained on the open internet.

⚡ White House Launches a "Tech Corps" to Export American AI Abroad
The Biden-era Peace Corps is getting a Silicon Valley makeover: the White House announced a new Tech Corps that will recruit and deploy STEM graduates as volunteers to implement American AI solutions in developing countries. The move is explicitly framed as a counter to China's growing influence in the Global South, where open-weight Chinese models like DeepSeek and Qwen have gained traction by being cheap and locally deployable.

📌 UCSF Study: AI Beats Human Research Teams on Medical Data Analysis
A new study in Cell Reports Medicine found that generative AI could analyze large medical datasets — specifically for predicting preterm birth — far faster than traditional human research teams, and in some cases produce stronger results. A junior student pair using AI assistance completed in months what normally takes seasoned data science teams years, pointing to a potential breakthrough for clinical research bottlenecks.

That's your Tuesday signal. See you tomorrow.

— The AI Decoded Team

Get the full weekly deep-dive every Monday → getaidecoded.com/subscribe