Qiao Ze took over for Li Jiangao and explained, "The end goal is an intelligent data model similar to ChatGPT. Li and I have a short-term objective to introduce a causality model into an intelligent data model like ChatGPT through a theoretically devised Group Theory decoupling framework, reducing the model's reliance on big data.
Our long-term goal is to create a mathematical theoretical foundation to construct an AI logic platform at the base level, into which related digital technologies can be integrated on demand."
Xu Dajiang was stunned.
Launching such a massive project right off the bat?
Isn't that terrifying?
50,000 yuan? Are you kidding me?
Even 500 million yuan wouldn't be enough, would it?
The general public might not be aware of how expensive it is to implement something like ChatGPT, but the academic community has some understanding.
After all, this thing is too hot right now!
Industry insiders have already done their calculations very clearly.
To train a model that rivals ChatGPT, one would need to invest approximately 1 billion US dollars per year just on hardware like graphics cards and CPUs.
And that's only the investment in hardware, not including various labor costs.
The most crucial aspect is the need for algorithm Continue reading more chapters on NovelHub for the best experience.
Since OpenAI launched its first generation of generative pre-trained models, GPT-1, in 2016, it took six years to make ChatGPT popular, during which Microsoft continuously injected funds into the project.
Alright, let's say that Qiao Ze, this genius youngster, really does come up with an algorithm framework more powerful than ChatGPT to conduct a dimensional reduction attack on it. But where will the enormous computing power come from? Who will provide the vast database needed to train the model?
Even if he sold himself, Xu Dajiang...
He wouldn't be able to gather enough money to play this game!
Isn't this exactly the kind of project that those big internet companies should be considering?
As for the long-term goals...
If such an AI logic-based foundational platform could indeed be constructed, withstand scrutiny, and be accepted, Xu Dajiang would not find it strange at all if Bill Gates were to crawl to Xilin University of Technology just to bow to Qiao Ze.
So at this moment, Xu Dajiang was trembling his lips, not knowing what to say.
Xilin University of Technology really didn't have the clout to attract such a formidable mogul to invest in such an ambitious project.
Not to mention that the project's initiators were not academicians, Yangtze River Scholars, young scientists, or outstanding youths, but just an associate professor and a freshman who had just started college.
Qiao Ze looked at Xu Dajiang somewhat puzzled.
That's because the dean's face, which had always worn a smile, had suddenly become very strange.
He was just introducing the project's concept; was there a need for such a reaction?
Fortunately, Li Jian Gao could probably sympathize with Xu Dajiang's feelings at the moment. He said with a bitter smile, "Qiao Ze, have you ever thought about how much this project is going to cost? How much manpower and resources are needed, and can we accomplish it just by ourselves?"
Qiao Ze came to a realization, then looked at Li Jian Gao and said, "It doesn't need a lot of money. I've checked the costs online, renting servers, computing power, and other expenses—it should be about one or two million yuan. We don't need to actually create a ChatGPT for mass promotion. Besides, we don't need that much money for the initial start-up."
"The short-term goal is just to verify that this framework is feasible, to ensure the service quality of end-to-end while optimizing the model to reduce the scale of parameters. It's enough to prove that the algorithm framework developed according to our theory has advantages. I'm not interested in the rest of the perfection work."
"Moreover, I believe that as long as there is sufficient evidence proving this, we will likely not lack funds for our long-term planning. I've seen many similar cases online. There will be people rushing to pay for it. Most importantly, our papers will definitely not be blocked again, right?"