Super God-Level Top Student

"Oh, let him in." After a simple instruction, Satya Nadella shifted a bit, making his sitting posture more formal, while contemplating why Kevin Scott unexpectedly wanted to see him.

Today, he indeed had a phone call with the CTO of Microsoft. The main point was to understand the development of Microsoft's own AI technology.

Unlike Apple's system-level interaction with OpenAI, Microsoft's real strategic focus is on AI application products. The collaboration with OpenAI is also to strengthen the fundamental market position of its system.

The reason he specifically talked to Sam Altman was because of the challenges Microsoft faced. If a series of defensive actions were not implemented, with OpenAI's increasing indecisiveness, the pressure on Microsoft in the systems market would evidently increase.

Of course, this can also be seen as pressure from the shareholders' meeting. Especially given the backdrop of a 17% decline in Surface sales in the first quarter of 2024, many shareholders naturally understand the situation concerning OpenAI.

Therefore, rather than saying Satya Nadella's high-profile meeting with Sam Altman was to pressure OpenAI, it's better to say it was a gesture. A company as big as Microsoft is unlikely to bet entirely on a company it cannot control on such a future-critical path.

In fact, Microsoft has also been developing an often-overlooked AI track — lightweight AI small models.

For instance, it took only a year to update from Phi-1 to Phi-3. This technology corresponds to three models, namely Phi-3-mini with 3.8 billion parameters; Phi-3-small with 7 billion parameters; and Phi-3-medium with 14 billion parameters.

These technologies are directly embedded for local operation. Windows Copilot Runtime includes a set of APIs supported by over 40 on-device AI models provided by Windows. The only problem is that in many open-source tests, many of Phi-3's intelligent capabilities are not ideal.

This is probably also why Sam Altman is not very interested in talking to Satya Nadella.

These large companies are all somewhat duplicative.

Even back then, when Sam Altman almost got kicked out of the OpenAI he founded, it was with Microsoft's support that he regained the company, ousting the previous board and rebuilding it. However, from Sam Altman's perspective, the main reason is that this founder and CEO has fundamental support at the grassroots level.

Over ninety percent of employees were willing to threaten resignation to support the two founders, enough to bankrupt all the board's plans at that time.

But who could have thought that things would be unpredictable? Two years after that incident, Satya Nadella would start to worry about the threat posed by a former junior to his flagship products.

...

"Hello, Kevin, I guess you must have some good news to find me at this time?"

When he saw Kevin Scott's figure walking into the office, Satya Nadella turned his chair and asked with a smile.

The obviously not too fit Chief Technology Officer shrugged, directly sat opposite Satya Nadella, and then said with a face full of hesitation, "In fact, I can't quite determine if this counts as good news for Microsoft; yes, I'm not sure."

"Oh? What is it?" This answer instead piqued Satya Nadella's interest, and he clasped his hands, asking seriously.

"Here's the thing, this afternoon I was with... sigh, never mind, in the afternoon I received an email discussing some technical matters. It pointed out some technical, um, flaws in our intelligent model... okay, these aren't important.

The key point is that the other party said they could provide an opportunity for our Asia-Pacific Research Center to collaborate with the Huaxia Xilin Institute of Mathematics and participate in the operation system design work for the future-oriented biosupercomputing research project."

Obviously, Kevin Scott's ability to articulate directly wasn't too strong, but the meaning was conveyed clearly enough.

"Xilin Institute of Mathematics?" Satya Nadella frowned, then looked across with a probing expression.

"Where Qiao Ze works; they're leading a biosupercomputing project and have already reached a phase success," Kevin Scott explained.

Hearing Qiao Ze's name, Satya Nadella immediately understood.

Although Qiao Ze is a mathematician, he is closely related to the technology field.

When he met Sam Altman, he even used Huaxia's artificial intelligence system to threaten him.

Moreover, Qiao Ze happens to be an AI framework developer capable of putting pressure on OpenAI.

Of course, besides these, Qiao Ze's achievements are legendary even here. Let's put it this way, for Satya Nadella, if Qiao Ze were willing to work for Microsoft, he wouldn't mind letting the person in front of him leave immediately, then give Qiao Ze both the CTO and scientist positions.

"Biosupercomputer? Can you verify the other party's identity?" Satya Nadella leaned back slightly, looking quite intrigued at Kevin Scott.

"Well, the other party claims to be the chief designer of the biosupercomputing project and a core figure at the Xilin Institute of Mathematics. However, they refuse to disclose other details for now. But I can judge from some technical documentation the other party provided that this identity should be genuine. And if we agree to the collaboration, we would directly sign a contract with Xilin and the Mathematics Research Institute."

Kevin Scott explained with some embarrassment.

He could tell that Satya Nadella was somewhat dissatisfied with this answer, but there was nothing he could do; he wanted to ascertain the other party's identity, but the initiative in this matter wasn't his.

In fact, the other party's words implied that they have many options, and Microsoft is not even the first choice.

"Alright, then what do you think, if this collaboration can really be facilitated, what benefits do we stand to gain?" Satya Nadella asked in a different way.

This question finally brought a vibrant look to Kevin Scott.

"Change, Mr. Nadella, an unprecedented change. The other party painted a picture of a future world for me. If all this can be realized and Microsoft participates, at least another glorious half-century could be expected," Kevin Scott said assertively.

"Explain in detail." Satya Nadella propped up his chin.

"Simply put, if the biosupercomputing project is successful, together with high-speed mobile internet, it can transform many existing hardware ecosystems. The initial stage is the revolutionary transformation of cloud computing and storage, including AI training functionalities.

You can imagine biological chips that can fold and deconstruct under the action of catalytic enzymes as needed, adapting to computational tasks. Coupled with storage speed and density far exceeding electronic chips, it can make large clustered supercomputing become history, further freeing us from dependency on hardware companies like Intel, NVIDIA, and AMD.

Especially, the support biosupercomputing offers to artificial intelligence is disruptive. The fit between the two is nearly perfect. In addition, with biological systems being inherently energy-efficient, their heat emission is several orders of magnitude lower than electronic products—we wouldn't need to spend enormous costs to place servers underwater to cool clusters.

Even in a more distant future, the other party proposes that no local electronic device would require the large hardware support currently necessary. CPUs, graphics cards, hard drives—none would be needed. Just a local memory and network connection would suffice to delegate computing and storage tasks to widely deployed biological servers. You know what this means.

Of course, these are all visions of development. More immediate benefits for us would be the chance to incorporate the Qunzhi Framework into Microsoft's systems if we secure this cooperation opportunity. We can integrate causally decoupled intelligent models into local callable memory.

It doesn't require support from cloud-based large models, yet the user experience it delivers will be a qualitative leap. With the help of this intelligent model, our NPU can assist customers in handling virtually all complex issues. I believe even a Mac combined with OpenAI wouldn't have such an experience." Kevin Scott finished almost in one breath.

Clearly, he was hoping to achieve this collaboration, or perhaps the technical details provided by the other party truly convinced him.

Well, Satya Nadella admits, he himself was almost swayed by that final reason. As for those before it, though enticing, Satya Nadella did not think he would be around to see that time.

But with OpenAI being unpredictable, a more advanced AI technology alternative indeed sounds like a good choice.

The only problem is that he vaguely remembers that the school where the Xilin Mathematics Research Institute is located seemed to have been entered into some blacklist.