Workflow
OpenAI首个开源大模型再延期、收购Windsurf失败;Manus “删号跑路”?百川联创离职,创始团队仅剩2人|AI周报
AI前线·2025-07-13 04:12

Group 1 - Manus has undergone significant layoffs, moving its headquarters to Singapore and hiring at high salaries, while clearing its domestic accounts on multiple platforms [1][2] - The company has reduced its workforce in China to about 120 employees, with over 40 core technical staff relocating to Singapore, while others face layoffs with compensation packages [2][3] - Manus is preparing for potential IPOs in Hong Kong and A-shares, with a higher probability for the latter due to recent strategic investments [6][7] Group 2 - The co-founder of Baichuan Intelligence, Xie Jian, is leaving the company amid a series of executive departures, including the commercialization head and others [7] - OpenAI has delayed the release of its first open-source AI model for further safety testing, and its acquisition of Windsurf has failed, leading to talent shifts towards Google DeepMind [8][10] - Alibaba's VP and former DingTalk CEO, Ye Jun, is set to leave the company after a series of strategic adjustments [12] Group 3 - Intel is facing large-scale layoffs, with CEO Pat Gelsinger admitting the company has fallen out of the top ten in the semiconductor industry, and its market value is currently at approximately $103.9 billion [13][14] - DeepSeek's usage has plummeted from 50% to 3% due to delays in updates and issues with data quality for training its new model [17][18] - The AI healthcare assistant app "Xiao He AI Doctor" has been launched by ByteDance, providing health consultations and report interpretations [32] Group 4 - The Kimi K2 model has been released and open-sourced, showcasing strong capabilities in code generation and general agent tasks [24][25] - The Grok-4 series AI model has been launched by xAI, claiming to outperform human graduate-level intelligence across various subjects [26][27] - Google has integrated the Veo 3 AI model into its Gemini application, allowing users to convert photos into short videos with audio [28]