Former Google video generation expert joins TikTok owner ByteDance amid generative AI race

A key contributor to Google’s VideoPoet video generation technology has joined TikTok owner ByteDance, as Big Tech companies on both sides of the Pacific race to advance generative artificial intelligence (AI) tools after OpenAI’s Sora stunned the world with its capabilities.

Jiang Lu, based in San Jose, California, also an adjunct professor at the Carnegie Mellon University (CMU), has joined ByteDance’s intelligent creation team, according to his profile on the company’s internal staff messaging platform seen by the Post. Chinese tech blog Jazzyear first reported the news on Thursday.

Jiang reports to Yang Jianchao, head of intelligent creation technology, who in turn reports to Zhu Wenjia, who leads research on large language models at ByteDance, according to the profile.

Do you have questions about the biggest topics and trends from around the world? Get the answers with SCMP Knowledge, our new platform of curated content with explainers, FAQs, analyses and infographics brought to you by our award-winning team.

The company did not immediately respond to a request for comment sent outside business hours on Sunday. Jiang did not respond to an emailed inquiry.

Jiang joined Google in 2017, working on content creation and deep learning for various services, including Google Ads, YouTube and the self-driving unit Waymo. Last year, his primary focus was VideoPoet, a large language model for video generation that was released by Google in December.

Before that, he was an intern at Yahoo and Microsoft in San Francisco and Beijing, respectively. In 2017, Jiang graduated from CMU with a Doctorate of Philosophy in AI. He completed a master’s degree at Free University of Brussels after graduating from China’s Xi’an Jiaotong University, according to his curriculum vitae.

OpenAI’s text-to-video model Sora, released on February 16, opened a new front in AI competition as tech giants allocate more resources to the next generation of applications.

A video created by Open AI’s text-to-video Sora tool plays on a monitor in Washington, February 16, 2024. Photo: AFP alt=A video created by Open AI’s text-to-video Sora tool plays on a monitor in Washington, February 16, 2024. Photo: AFP>

China’s Big Tech firms have been working on similar technologies, including ByteDance’s PixelDance introduced in November 2023, and Tencent Holdings’ VideoCrafter, but neither can compare to Sora’s capabilities in video generation.

In early February, ByteDance unveiled Boximator, a video motion control tool to assist in video generation, but the company said the technology was still in its infancy and not ready for general release. “It still has a big gap with leading video generation models in terms of image quality, fidelity and duration,” said a company representative.

In January, ByteDance chief executive Liang Rubo berated employees for being too slow to react to the emergence of new technologies such as generative AI. Liang said in an internal meeting that employees did not start discussing ChatGPT, OpenAI’s chat bot launched in November 2022, until the following year.

This article originally appeared in the South China Morning Post (SCMP), the most authoritative voice reporting on China and Asia for more than a century. For more SCMP stories, please explore the SCMP app or visit the SCMP’s Facebook and Twitter pages. Copyright © 2024 South China Morning Post Publishers Ltd. All rights reserved.

文章来源：South China Morning Post

TKFFF合作，请扫码联系