⌘K

Change language Switch ThemeSign In

Narrow Mode

Deep Dive: The Transformation Struggles of Xiaoice and Pika and Industry Lessons

Deep Dive: The Transformation Struggles of Xiaoice and Pika and Industry Lessons

@shao__meng

最近顾问的团队招聘，面试过程中对两家公司印象很深：小冰和 Pika。

小冰是几年前 NLP 时代微软的 AI 数字人，当时还是很有吸引力的，后来团队从微软独立出来，直到 ChatGPT 发布，一夜变了天，自己成为了「牛夫人。。」，NLP 时肯定做不下去了，团队也尝试做模型，很自然的没有成功，后来就开始做短剧、做音乐工具等等，各种尝试却完全没有技术核心，几乎看不到前景了。

Pika 则不同，名校高材生创始团队，做 AI 视频生成模型，当时和 Runway 还是能拼一拼的，但后来 Sora 出来，后来 Veo、Kling、Seedance 连续刷新视频生成上限，Pika 也就掉队了，算力和人才储备明显没跟上。现在转型去做 AI 分身，可以说和原来 AI 视频关联不大了。

Mar 26, 2026, 8:10 AM View on X

0 Replies

1 Retweets

5 Likes

1,236 Views m meng shao @shao__meng

One Sentence Summary

Drawing from hiring interviews, the author analyzes the transformation challenges and lack of core competitiveness faced by AI companies Xiaoice and Pika amidst fierce market competition.

Summary

Based on firsthand observations during the hiring process, the author provides a sharp critique of the current development status of AI companies Xiaoice and Pika. The author points out that Xiaoice lacks a core technical moat after spinning off from Microsoft and has a blurred transformation path. Additionally, the author analyzes Pika's decline in the face of competitors like Sora and Veo, and its strategic pivot to AI avatars. The content reveals the survival challenges and business model dilemmas faced by AI startups against the backdrop of rapid technological iteration.

AI Score

78

Influence Score 1

Published At Today

Language

Chinese

Tags

AI Industry Analysis

Xiaoice

Pika

AI Business Models

Technical Transformation HomeArticlesPodcastsVideosTweets

Deep Dive: The Transformation Struggles of Xiaoice and Pi...