Japan's Rapidus and Tenstorrent Partner to Develop 2nm AI Chips


Rene Haas, CEO of Arm, stated in a BBC interview that the most disruptive application of AI is in healthcare, with the potential to help humanity conquer cancer within this generation's lifetime and shorten drug development time. However, he emphasized that the industry must first address the issue of insufficient chip capacity to provide sufficient computing power for AI.
On Monday, the UN High Commissioner for Human Rights, Michelle Bachelet, warned that advanced artificial intelligence could pose an existential risk to humanity. She called for 'ironclad safeguards' to ensure AI safety and pledged to urge relevant companies to reduce risks. The remarks were made as she prepared to begin her second term in Geneva, stating that she shares the concerns of the industry regarding the threats posed by AI.
According to reports, Anthropic has decided to abandon the acquisition of AI startup Decart. The two parties had been negotiating a deal worth around $6 billion, but Anthropic withdrew after completing due diligence. The company may seek other collaborations in the future. Anthropic is increasing its investment in computing power to develop new products, while Decart's software can improve chip utilization efficiency and reduce the cost of AI training operations, which may have been the motivation for the acquisition, but the deal ultimately did not go through.
The Supreme People's Court issued the first judicial document on AI-related cases titled "Opinions on Legally Adjudicating AI-related Dispute Cases", consisting of 5 parts and 24 articles. It focuses on hot topics such as AI face-swapping, big data price discrimination, counterfeit celebrity live-stream selling, autonomous driving, and intellectual property. It clearly defines three basic principles and liability regulations, clarifies legal boundaries and responsibilities, and safeguards the healthy and orderly development of the AI industry.
WeChat AI open-sourced WeMM-Embedding, a multimodal embedding model in 2B, 4B, and 9B versions. It supports text, images, video, visual documents, and arbitrary mixed multimodal inputs. Its multimodal isomorphic core enables deep understanding of text and visual information, improving cross-modal retrieval.....