What's my take on Lin Junyang's departure?

Recently, Qwen has been getting an incredible amount of attention, then it encountered the global stock market crash, and then the departure of the 32-year-old P10 Lin Junyang, a young talent from Qwen, happened. The overall market is a mess, and Alibaba's stock price is under greater pressure than other China concept stocks. Friends in the community have been messaging me privately and commenting on my posts, asking me to share my thoughts on this matter. First, I'd like to recommend everyone to take a look at Lan Xi's take on Weibo, I think it makes a lot of sense.

As for my view, first, the loss of technical talent is definitely a regret; second, his departure was very dignified and not ugly, the core team also sent farewell wishes. Qwen, at least up to now, has been a very outstanding and united open-source model team, and it has also allowed foreigners on X, Hugging Face, and GitHub to understand Alibaba's domestic large model, earning a good reputation for it; third, many people talk about Alibaba's corporate culture this and that, but the reality is that this corporate culture created and mass-produced talents like Lin Junyang, so how do you explain that? Soil that produces talents in abundance, no matter how you look at it, is good soil.

Wu Ma sent an internal letter within Alibaba this morning, the original text is as follows:

"Dear Tongyi Lab colleagues:

The company has decided to approve Lin Junyang's resignation and thanks him for his past contributions in his role. Jing Ren will continue to lead the Tongyi Lab to advance the follow-up work. At the same time, the company will establish a basic model support team, coordinated by me, Jing Ren, and Fan Yu, to mobilize group resources to support the construction of basic models.

Technological development either advances or retreats. Developing foundational large models is our key strategy for the future. While continuing to adhere to the open-source model strategy, we will persistently increase R&D investment in the AI field and intensify efforts to attract outstanding talents. Let's work together."

The capital market reaction actually isn't that big, it's still more about the up and down of the overall China concept stock market. In fact, everyone also knows that the departure of one person or even a few people won't cause earth-shaking changes to the entire development of Qwen. How many OpenAI founders left? They all left, but that didn't stop its rapid advance.

Many people simply attribute Lin Junyang's departure to the irreconcilable conflict between technical idealism and commercialization development, which actually also misreads the current actual development of Alibaba AI, at least it's incomplete, or fails to grasp the main contradiction.

I believe the most important point behind the real logic of this change lies in: the upgrade of Qwen's strategic status, and the reconstruction of talent density.

Currently, Qwen's positioning has been comprehensively elevated from a simple foundational large model project to Alibaba's core overall strategy for the future, feeding back into almost all businesses. When a project is upgraded to the group's number one project, the challenges it faces, the problems to consider, and the resources required are completely different from those of a lab. At this stage, the company's management must make choices, knowing that: facing the increasingly fierce and brutal AI arms race at the global forefront, if Alibaba wants to continue moving forward, it must recruit more world-class technical experts to enhance the talent density of the foundational model team.

This process will inevitably involve the reorganization of the existing team structure and adjustments to Lin Junyang's own scope of authority and responsibility. For a young technical backbone who led Qwen's technology from 0 to 1, facing changes and dilution of his own authority and responsibility, finding it difficult to accept and thus resigning is a logical personal choice in the context of a fast-developing workplace and enterprise. If you put yourself in his shoes, you can understand.

And from a shareholder's perspective, for Alibaba, this might not be a bad thing. The competition in AI large models is not at a stage where a team of a few geeks can clear the game. If Alibaba AI is to be pushed to the next peak, increasing talent density is the only path. Having developed to this scale, organizational growing pains and adjustments are perfectly normal. A few technical geeks alone cannot make it big; having only geek geniuses without commercial vision, cases of company collapse are everywhere. You can look at the example of General Magic, where prodigies gathered, and ideals fell.

As a listed company, while its core e-commerce business is under pressure, Alibaba still invests tens and hundreds of billions in Capex into the AI field every year. Wall Street and shareholders have limited patience. The previous continuous climb in Alibaba's stock price was their joy, optimism, and cheers, but they weren't cheering for a purely technical open-source research lab; they were cheering for how the Qwen large model would feed back into Alibaba's e-commerce and cloud businesses, bringing future cash flow.

As a shareholder, I send blessings for Lin Junyang's departure; also as a shareholder, I believe that for Alibaba's grand AI undertaking and its prospects, choosing to increase talent density rather than sacrificing the whole picture for a few talents precisely shows Alibaba's clarity.