Top Student at Their Peak-Chapter 959 - 348: Evolving Intelligence

Clearly, both Tao Xuanzhi and Tian Yanzhen subconsciously believe the idea of hiding a mathematical problem within the invitation is Qiao Yu’s handiwork.

After all, the problem posed is quite tricky and very much in Qiao Yu’s style.

Especially Tian Yanzhen, he obviously knows about July’s existence. He still frequently communicates with July to this day.

But he didn’t realize July could independently formulate such a complex mathematical problem.

Excluding July, other artificial intelligences are not incapable of creating problems.

They can even provide a highly complete set of examination papers. It has now begun to be demonstrated that using AI technology to assist teachers in creating important exam papers can be done.

This can shorten the time taken to draft important exams and reduce the number of teachers needed, making resources easier to allocate and ensuring confidentiality.

After all, inviting dozens of teachers every year for months of isolated living is a cumbersome affair.

However, the problems created by other artificial intelligences are mostly an extension of human wisdom. Or rather, modifications of human question bank content.

Making them like July, embedding specific answers into a complex problem, especially abstract Qiao Algebraic Geometry problems, Tian Yanzhen confidently asserts that no AI other than July can accomplish this.

Previously, he felt that July was temporarily unusable on a large scale. But upon hearing this news, it indeed shocked Tian Yanzhen.

Even if July still doesn’t understand creation, being able to independently design such intricate mathematical problems means it can already do a lot.

Or at least in solving mathematical puzzles, it has definitely surpassed many highly matured existing AI products.

In the field of mathematics, those who can tackle difficult problems aren’t necessarily capable of creating them. However, those who can create them usually tackle problems effortlessly.

Problem creators often have a deeper understanding of mathematics and think more comprehensively and flexibly when solving mathematical issues.

This is perhaps why the same experts rotate within Huaxia’s college entrance exam drafting group every year.

For a moment, Tian Yanzhen didn’t know what to say.

Though he believes Qiao Yu’s research will certainly succeed and he has great expectations for July.

But how long has it been?

Counting fully, from being a little program that only bantered with him daily to now stumping a considerable number of mathematicians worldwide, July has astonishingly grown in just half a year.

According to Qiao Yu, July’s growth isn’t linear, but exponential.

This means the early accumulation phase will be slow. With no tuning systems, everything depends on July’s main program to independently comprehend and judge...

Thus, July’s growth path must be one of accumulating substantial groundwork before bursting forth. Moreover, the early accumulation process can’t simply be about stacking databases.

Otherwise, with current machine retrieval and matching capabilities, it would be impossible to resolve the long-term memory issue of AI.

Especially without parameter libraries. AI memory isn’t just about storing and querying memory; it requires deep understanding and internalization of information.

So, this can only indicate that in half a year’s time, July has not only mastered critical techniques in the field of basic mathematics but might have also reached high standards in algorithms, reasoning, and optimization.

If July continues to learn ever faster, Tian Yanzhen dares not predict to what level it might grow in a year or two.

After the shock, Tian Yanzhen quickly asked the question.

"You mean July mastered all knowledge from basic mathematics to generalized modal axioms and Qiao Algebraic Geometry within half a year?"

"Well... it should be like that. Actually, I’m not entirely sure of July’s learning progress. You know its database was designed with multiple layers of encryption from the start.

Tao Xuanzhi even wrote a letter specifically to discuss this issue with me. I only saw this letter an hour ago, realizing that July played such a big joke on everyone...

I roughly studied the problem it posed, and just had a brief discussion with it. I can only say July’s progress is faster than I anticipated.

This might be related to its main distributed layout. You know, Google, Microsoft, and domestic companies like the Quantum Simulation Computing Alliance are very courteous.

Upon hearing my need to test the new generation of AI, they voluntarily opened up their computing power to provide support.

The current situation is that July can not only communicate with us independently but also access numerous databases uninterrupted for 24 hours, enriching its knowledge base.

So perhaps the early learning and accumulation speed is a little faster than I previously estimated..."

Qiao Yu diligently explained, feeling this reasoning was quite rational.

After all, July’s learning and operational rules are quite different from all previous AI.

Thus, all prior experience with AI can’t be applied to July.

Additionally, the computing power supporting July’s learning is significantly greater than that of any other company’s AI developed so far.

After all, purchasing those servers and bulk buying various CPUs and GPUs cost money.

July, however, has been consistently using the computing power provided free by numerous large technological enterprises worldwide, and can even access databases without strong security levels.

These conditions combined demonstrate that previous experiences are no longer applicable here.

Perhaps on other levels, the cooperation between Huaxia and West University isn’t entirely thorough yet.