Chapter 166 - 124: God’s Draft
However, this open attitude toward cutting-edge fields also created headaches for editors like Bob.
On the one hand, the underlying foundation of artificial intelligence is neural networks. This led countless ambitious people to see the field as fertile ground for churning out junk papers.
They would often just take a few classic neural network architectures, make some minor, inconsequential tweaks, and then haphazardly arrange them in new combinations. They’d tack on a mathematical proof that looked profound but was actually utter nonsense, and finally, run the model on a few overused, Entry-Level datasets to get some results. And just like that, another hot-off-the-press junk paper was born.
This problem grew even more rampant after large-model technology exploded in popularity. The JAMS submission inbox was flooded with this kind of "academic garbage" almost daily.
On the other hand, there were indeed top scholars with exceptionally deep mathematical expertise. These were the ones proposing truly groundbreaking improvements to AI algorithms, or even entirely new algorithmic frameworks, from a more theoretical, mathematical standpoint.
But these types of papers often employed a vast array of abstruse, obscure, and sometimes even novel mathematical tools and methods.
Across the entire world, you could count the number of people qualified and capable of reviewing a manuscript of this caliber on one hand.
And every last one of these heavyweights in the world of mathematics was an incredibly busy person, swamped with endless responsibilities.
Even with someone as famously approachable as Terence Tao, getting him to agree to review a paper was a matter of luck.
A delay of a single day could mean he was gone on a business trip for a month. Bob’s current predicament was a common one.
The problem was, aside from Terence Tao, Bob couldn’t think of a single other candidate off the top of his head.
The crucial point was that the paper he held was fundamentally different from those shoddily produced junk papers!
This had already been unanimously confirmed during the initial screening by the most senior editors on the JAMS editorial staff.
Even though they had put their heads together and studied the paper for the better part of a day, they could only make sense of a small fraction of it. But that was enough for them to determine one thing: the author of this paper possessed a frighteningly high level of mathematical skill!
That was why the editor-in-chief had specifically instructed Bob to get this paper’s review process completed before the next issue went to press.
So, Bob could only steel himself and make one last effort.
"Professor Tao, I promise you, the contents of this paper will absolutely capture your interest!" Bob’s voice held a note of pleading.
And that was no exaggeration.
For him, with his level of mathematics, the main body of the paper might as well have been written in Greek.
But he had studied the abstract meticulously, word by word.
In the abstract, the author stated plainly that he was proposing a completely new path to solving the "generative large model hallucination problem"—an issue that had both academia and industry stumped.
Seeing the lingering hesitation on Terence Tao’s face, Bob decided to turn up the heat.
"The author uses differential manifolds as the main framework. Professor Tao, in this field, besides Professor Qiu Chengtong over in Huaxia, you’re the most qualified person in the world to review it!"
Bob’s tone carried a perfectly measured dose of flattery. "Considering our many years of collaboration at JAMS, would you just take a look? Even a quick scan of the full text?"
Hearing this, Terence Tao let out a long sigh.
To be honest, on any other day, hearing Bob put it so earnestly, he would have certainly made time to take a look.
But today, it just wasn’t possible.
Since the first half of the year, the top brass in the USA had, without warning, slashed research funding for the entire University of California System, and the lab he directed was not spared.
The biggest reason he had agreed to Jask’s invitation to fly to Silicon Valley and help test the mathematical capabilities of a new model was Jask’s promise to provide his lab with a generous amount of funding over the next three years.
For Terence Tao, that money was a godsend.
Therefore, he really didn’t want to—and couldn’t afford to—bail on Jask.
However, Bob seemed to read his mind and immediately seized the opportunity to press his case. "Professor Tao, I know you’re on your way to see Jask. But just think, what if...
I’m just saying, what if this paper I brought today actually contains some insightful ideas that could help with the training and development of large models? It might even lead to a deeper collaboration between you and Mr. Jask!"
At these words, a flicker of interest finally sparked in Terence Tao’s deep-set eyes, hidden behind his gold-rimmed glasses.
It was true that in the last couple of years, nine out of ten of the so-called "AI-focused" papers sent to him by JAMS and other journals had been a waste of his time. He had even tactfully suggested to the editorial staff before that they needed to be more stringent in their preliminary screening.
But today, seeing Bob’s unprecedented conviction, and thinking back on his words...
Terence Tao was genuinely tempted.
He glanced at the time on his phone, mused for a moment, then hesitated no longer.
He first stepped aside and made a call to his assistant, giving a single, concise order: "Reschedule my flight. Push it back by at least an hour."