South China Morning Post:
DeepSeek V4 Pro has 1.6T total parameters, its largest model by that metric, and V4 Flash has 284B parameters; both models have a context window of 1M tokens  —  The company says its cost-efficient new V4 model is competitive with top closed-source models from OpenAI and Google DeepMind

Reuters:
Huawei says its Ascend supernode based on the Ascend 950 AI chips will fully support DeepSeek V4, as DeepSeek launches a preview of its V4 model  —  Huawei Technologies said on Friday its Ascend supernode based on Ascend 950 AI chips would fully support Deepseek's V4 versions …

Bloomberg:
DeepSeek releases its new flagship models V4 Pro and V4 Flash in preview, saying V4 Pro trails the performance of state-of-the-art models by about 3 to 6 months  —  DeepSeek rolled out preview versions of a new flagship artificial intelligence model a year after upending Silicon Valley …

Reuters:
S-1 excerpts: SpaceX warns multiple investigations into sexually abusive AI imagery created with xAI's Grok may hurt the company's global market access  —  The multiple investigations into xAI's creation and dissemination of sexually abusive imagery may lead the company to lose access to certain markets …

Ben Thompson / Stratechery:
Q&A with Google Cloud CEO Thomas Kurian on building infrastructure for AI agents, balancing internal needs and the demands of customers like Anthropic, and more  —  An interview with Google Cloud CEO Thomas Kurian about Google's cloud priorities, enterprise agent platform, and Google's integration advantage.