DeepSeek's V4.1-Flash uses a new encoder-decoder design to make long contexts cheaper to run
One set of parameters processes the prompt and a larger one generates the answer, and DeepSeek says the design cuts the memory for cached context to a quarter of V4-Flash's. On September 10 DeepSeek released V4.1-Flash, the...
AI agents account for little of the commerce on Coinbase’s x402 payment protocol
x402 was designed for payments by AI agents, but their share of commercial spending is just 0.6% to 7.5%, TRM...
AlphaFold helps biologists find a cellular off switch tied to cancer-cell movement
The switch, a protein called Avl9, comes from a family long thought to act as on switches, and the findings suggest some...
IBM and NASA release an open-source AI model for mapping ice and craters on the Moon
Researchers can reuse one pre-trained system for each new lunar mapping task instead of building a separate model for every geologic...
A 20-minute training course improves people's ability to detect AI-generated faces
A 15-to-20-minute training session built at the University of Southampton moved more than 600 participants from worse-than-chance...