AI Products36氪

Is a Chinese Version of Muse Worth BAT's Bet?

36氪Oct 14, 20261 min read

AI Summary

After Meta's personal AI agent Muse went viral, Chinese tech firms such as ByteDance and Tencent rushed to launch their own versions. However, China's fragmented app ecosystem makes it hard to replicate a general-purpose agent like Muse, so companies are more likely to embed agents into their existing service systems. Trust and profitability challenges remain.

AI Insights

China's tech giants are unlikely to replicate Meta's Muse as a general-purpose agent due to fragmented app ecosystems, instead embedding AI agents into existing services where data permissions are more controllable. The race is less about building a clone and more about securing position in the emerging AI super-entry point, while trust, security vulnerabilities, and unclear profit models remain the key barriers to mass adoption.

Analysis written by AI from the original article. Check it against the source.

Translated text

Meta's personal AI agent Muse reached 5 million downloads within three weeks of launch, prompting OpenAI, Google and Anthropic to follow. In China, ByteDance is developing a personal assistant project codenamed 'Spell' under its Doubao team, while Tencent launched the cloud assistant LightVela and is testing Handy Bot as a standalone app. However, Muse uses an independent cloud virtual machine architecture to autonomously handle general tasks, whereas BAT firms are embedding agents into their existing ecosystems with more restrained permissions. This is largely because China's leading apps are segmented: an agent cannot easily read WeChat chats, pull Douyin recommendations and place orders on Taobao at once. Alibaba packages its app ecosystem as skills callable by its own agents, ByteDance's ArkClaw deeply integrates Feishu and Douyin plugins, and Tencent is building SkillHub. Personal agents log in, pay and make decisions on users' behalf, so liability for delegated authority and security are central issues. Muse disclosed a zero-day vulnerability two weeks after launch, and only 8% of US users said they would give Meta their passwords. Profitability is also uncertain, since multi-step agent tasks raise CPU, browser and API costs per task. Personal agents remain some distance from mass adoption, and the firms that balance security, privacy, openness and user experience are most likely to win the next internet gateway.

Related News

AI Products36氪

Geeks Build Cabbage-Priced Biped Robot for Under 3,000 Yuan, with Old Phones as Key Contributors

UCLA's RoMeLa lab has developed PhoneBot, a 1.8 kg biped robot with 13 degrees of freedom and a materials cost of about $400 (roughly 2,677 yuan). The robot is powered by a smartphone that serves as its sensor suite, onboard computer, and interface, and by using used older phones, total cost stays under 3,000 yuan.

AI Products36氪

Alibaba, ByteDance and Meta Ignite Personal Agent Race: No Table Without a Good Model

In late September, Alibaba announced at its Yunqi Conference that Qwen (Tongyi Qianwen) would accelerate personal agent development, and on October 13 it unveiled 'Qwen Go!', a personal assistant in limited invitation testing. ByteDance (Doubao), Tencent and others have also launched personal agent projects, sparked by Meta's Muse. On the model side, Meta's Muse Wonder and Alibaba's QwenGyre reinforcement-learning framework aim to strengthen long-running agent capabilities.

AI Products36氪

Tech Giants at Home and Abroad Scramble for the Muse Entrance, But the Personal Agent Remains a False Proposition

A month after Meta launched the personal agent Muse, touted as the most captivating product since ChatGPT, ByteDance, Tencent, Alibaba, and Ant Group have rushed out similar offerings. However, Muse's daily net new users fell 62.4% from the September peak, and analysis suggests the model struggles to fit the domestic Chinese market in both product logic and business model.