Issue an additional one million US dollars in stock options, MiniMax defines the talent profile

The smoke of the computing power competition has not yet cleared, and a fierce battle for talent has begun in the AI industry. Tech giants in Silicon Valley across the ocean are offering annual salaries in the hundreds of millions to poach top researchers. Meanwhile, the competition for AI talent in China is also heating up.

The latest company to make a significant move is MiniMax. The company revealed at a recent all - hands meeting that it has officially launched a long - term stock option incentive plan covering all functions of the company.

This plan is not only for core algorithm R & D personnel but also extends to all job sequences such as engineering, product, and operations, and even to newly recruited interns. That is to say, as long as they make outstanding contributions, regardless of job type and seniority, employees have the opportunity to receive stock option incentives. According to exclusive information obtained by 36Kr, this mechanism will exist as the company's talent system in the long term, encouraging every employee to make bold breakthroughs.

Facing this global competition for AI talent, MiniMax has come up with a different solution. Compared with the prevalent practice in Silicon Valley of poaching high - profile individuals with high salaries, MiniMax's all - employee incentive is more like an organizational innovation in the AI era: transforming individual creativity into collective potential through a talent incentive system.

From Computing Power Stacking to Talent Competition

For a long time after ChatGPT amazed the world, the development of the AI industry has been in a narrative of "brute force creates miracles": whoever can get enough capital support and hoard enough computing power can train a more powerful large - scale model. However, since DeepSeek optimized the training cost under the same inference performance with an all - local employee team and a unique technical path, "innovative talent" has once again become the focus of the global AI narrative.

As the economist Joseph Schumpeter proposed, the source of innovation lies in people's "creative destruction". On the road to AGI, the scarcity and irreplaceability of talent far exceed that of computing power. Only by stimulating the non - linear creativity of individual talents can we break through the limitations of computing power stacking.

The fierce talent war is playing out in Silicon Valley against this backdrop. Among them, Meta has been the most aggressive.

In June this year, dissatisfied with the performance of its Llama model, Mark Zuckerberg re - formed a "super - intelligent" team and poached a large number of employees from companies such as OpenAI, Google DeepMind, Anthropic, and Apple. This strategy of recruiting AI talent with large sums of money has caused a stir in the entire Silicon Valley.

Other tech giants are also not willing to lag behind. Microsoft, Google, etc. are also recruiting for their AI teams through various means. In July this year, in order to poach the co - founders and some senior researchers of the startup company Windsurf, Google spent $2.4 billion to acquire the entire company.

The smoke of the talent war is also spreading in China. Different from the "overt" poaching practices in Silicon Valley, domestic companies are cultivating talent through talent incentive methods and maximizing the creativity of talent. MiniMax is one of them.

MiniMax's booth at the WAIC 2025 Conference

More importantly, different from most companies that emphasize the "heroic" drive of core technical personnel, MiniMax has extended its talent incentive plan to all positions. Thus, MiniMax's talent incentive mechanism seems to have gone beyond the simple scope of talent poaching. It is more like an organizational innovation, attempting to pool individual capabilities into the potential of the entire organization.

From Individual Innovation to Organizational Innovation

Innovation is never a single - point breakthrough. It is the process of gradually pooling individual inspiration and actions into the collective strength of the organization through appropriate mechanisms and a common mission. Through its all - employee incentive mechanism, MiniMax aims to start from "individual innovation" and extend this creativity to the entire organization, promoting the formation of real "organizational innovation".

However, to make this innovation a normal state of the organization rather than sporadic flashes, a more fundamental question must be answered: What individual traits are worth being continuously amplified?

MiniMax has provided a definition. A person close to MiniMax told 36Kr that the concept of "talent double helix" implemented within MiniMax emphasizes two types of interacting common capabilities: First - Principles Thinking + Deep Curiosity.

First - Principles Thinking emphasizes the ability to discover "real problems". This also coincides with a common view after generative AI tools have penetrated into work and life scenarios: in the AI era, the ability to ask questions is more important than the ability to answer them. MiniMax aims to truly implement this theory in its organizational form.

MiniMax's office environment

In today's world where AI can replace most execution tasks, a position is no longer the boundary of ability. Instead, it is about identifying the real problems that determine the direction of the organization in a complex environment. This ability can come from algorithm scientists, engineers, product managers, or even interns.

Like building a rocket, a large - scale model itself is a complex system engineering. For example, in the OpenAI organization, roles from data experts to algorithm and engineering optimization, and even product managers play a leading role in the entire R & D process.

The insider pointed out that the above - mentioned idea is also the underlying logic of MiniMax's all - employee incentive design. In this way, anyone who can propose key propositions will be recognized.

What MiniMax emphasizes as "Deep Curiosity" refers to the driving force to continuously explore the unknown and the courage to constantly experiment and break through in the face of uncertainty.

They are the "natives" of the AGI era. They do not see AI as a threat but as the most powerful assistant to satisfy their curiosity and explore a broader world.

MiniMax alumni association, chatting with top university talent about technical details

MiniMax's internal concept holds that the path to AGI may go beyond current human imagination. Therefore, in this process, more important than existing experience is the continuous desire to explore, the courage to face the unknown, and the willingness to accept failure. The company regards this "Deep Curiosity" as an important force to push individuals and the organization to break through cognitive boundaries and achieve non - linear growth." The insider summarized.

Like the double - helix structure of DNA, these two traits are intertwined, jointly constructing sustainable creativity and competitiveness in the AGI era. In fact, it is such a group of people who are driving this round of AI development.

Ilya Sutskever, the co - founder and former chief scientist of OpenAI, is a typical example with First - Principles Thinking and Deep Curiosity. During his doctoral studies, he participated as a core member in the development of AlexNet by Alex Krizhevsky and his tutor Geoffrey Hinton, and they achieved an overwhelming victory in the 2012 ImageNet competition. This achievement is generally regarded as a key turning point for deep learning in computer vision, which has triggered a global upsurge in artificial intelligence research.

After founding OpenAI, he continued to start from first principles, advocating exploring general intelligence through paths such as "predicting the next token" and "information compression". This technical judgment has had an important impact on the direction selection of the GPT series and ultimately led to the phenomenal birth of ChatGPT.

These bold explorations are inseparable from the drive of Deep Curiosity. From computer vision to deep learning and then to general intelligence, Sutskever has been constantly exploring and questioning. Under the interaction of these two traits, Sutskever has undoubtedly become one of the most representative figures in the generative AI era.

Through this all - employee incentive measure, what MiniMax aims to achieve is to gather all the minds with "essential insights" and "deep curiosity" in different chains of AGI exploration, incorporate individual innovations in the algorithm, engineering, and product chains into an integrated mechanism, and make every exploration and innovation have the possibility of being recognized and amplified.

When the innovations of different roles are continuously superimposed, innovation will emerge at the organizational level, and ultimately converge into a continuous driving force for the entire company towards AGI.

In the Chinese AI startup landscape, MiniMax's uniqueness lies in that: it is regarded by external observers as a "Silicon Valley - style Pure Play AI company". Pure Play means not doing addition across industries but focusing on AI itself, forming a closed - loop from the model to the underlying products. At the same time, it is also considered by the outside world to have a strong will to survive, taking a pragmatic and radical approach to continuously break through in technical paths and quickly verify in commercialization.

The video shows the first and last frames of MiniMax's video model launch, supporting features such as the strongest complex instruction following.

From a more macroscopic perspective, every technological leap is accompanied by an update of the concept of talent. The industrial era emphasized execution, the Internet era valued speed and scale, while in the AI era, the most scarce are the ability to define problems and cross - border imagination. As large - scale models gradually take over repetitive execution and optimization, human value will be more concentrated on proposing problems worth solving and creating new combinations. In other words, the talent portrait of the next era is no longer the superposition of seniority and identity, but explorers who dare to face the unknown, stay curious, and have essential insights. This is not only the "talent double helix" emphasized by MiniMax but also the common call for future talent in the entire industry's journey towards AGI.