"MODEL"的搜尋結果
2026-04-16
08:12

中國的 GigaAI 憑藉製造數據優勢推進世界模型技術

中國的 AI 新創公司 GigaAI 正在推進世界模型(world model)技術,透過龐大的工業數據來強化實體 AI(physical AI),在表現上超越領先的美國模型。儘管它在資金與合作方面有所進展,但在安全部署與真實世界的複現方面的挑戰仍然至關重要。
展開
01:51

美團開源LongCat-Next:3B參數統一視覺理解、生成與語音

Meituan Longcat Team's open-source LongCat-Next is a multimodal model based on MoE architecture, integrating five capabilities including text, visual understanding, image generation, and speech. Its core design DiNA achieves unified task processing through discrete tokens, while the dNaViT used in the visual aspect enables excellent image generation performance. Compared with similar models, LongCat-Next demonstrates leading benchmark performance across various metrics, showcasing its advantages in multimodal understanding and generation.
展開
00:53

Mercedes-Benz Integrates Domestic AI into Maybach: Zhipu and Tsinghua Create On-Device Multimodal Large Model for S-Class

According to 1M AI News monitoring, Mercedes-Benz China announced that an edge-side multimodal large model jointly developed with Tsinghua University and Zhipu, an AI large model company listed on the Hong Kong Stock Exchange in January this year, will be deployed in the next-generation Maybach S-Class and the new S-Class sedan. The model is deployed in the rear-seat entertainment system, integrating natural language processing with multimodal understanding and generation capabilities in vision, audio, and other modalities. Combined with in-vehicle cameras and other perceptual hardware, it completes inference under local computing power conditions. According to Zhipu's official statement, Mercedes-Benz has become the first automotive brand to apply multimodal large models in rear-seat entertainment systems.
展開
08:34

Circle Stock Price Plummets 20% Due to Revenue Ban, New Regulations May Create Regulatory Moat for USDC

Circle Internet Group stock price plummeted 20%, with a market value loss of $4.6 billion, due to the draft "Digital Asset Market Transparency Act" prohibiting passive income provision to stablecoin holders. Despite short-term impacts on application restrictions, analysts believe the act will build regulatory moats for Circle, optimizing its business model. Additionally, Tether commits to accepting audits to enhance transparency. USDC's future performance will influence market trends.
展開
07:21

讓AI編程助手直接操控桌面:Zig原生CLI工具usecomputer開源,相容Claude Code和Codex

Open-source developer Tommaso De Rossi launched the desktop automation tool usecomputer, written based on the Zig language, supporting macOS and Linux, allowing AI programming assistants to control the mouse, keyboard, and take screenshots. The tool provides a variety of CLI commands, can automatically convert coordinates, adapt to the model's context window, and supports direct inline screenshots. This project was extended by De Rossi based on the browser automation tool Playwriter.
展開
07:08

Figure founder invests $100 million from personal funds to build "personal AI": former Apple iPhone Air design engineer joins to create a series of AI hardware

Figure CEO Brett Adcock announces the establishment of new AI lab Hark, aimed at developing the world's most advanced personal intelligence, integrating multimodal models with customized hardware. Hark has been operating in stealth for 8 months, with plans to release its first AI model in summer. The team comprises approximately 45 people, with plans to expand to 100, and has entered into a computing power collaboration with Nvidia. Adcock has previously founded multiple startups, while Figure continues to operate independently.
展開
07:02

人形機器人公司 Figure CEO 自掏 1 億美元創辦 AI 實驗室 Hark,前蘋果 iPhone Air 設計師加盟

The CEO of humanoid robot company Figure, Brett Adcock, announced the establishment of a new AI lab called Hark, dedicated to developing advanced personal intelligence that integrates multimodal models with custom hardware. Hark has been operating secretly for 8 months with an initial investment of $100 million, and plans to launch a new type of AI device with the first batch of models expected to be released this summer. The team has recruited talent from companies like Apple and Meta, and is collaborating with NVIDIA to leverage GPUs for model training.
展開