The richest man started taking advantage of loopholes after joining Alibaba.

The silence lasted for about twenty seconds.

Twenty seconds is nothing in normal times, but in room 607 it felt like twenty minutes.

The sound of water flowing from the radiator filled the silence.

The head of the National Security Bureau remained silent.

His pen was placed next to the checklist, with the cap off.

The dark blue folder was still closed, with six horizontal lines replacing the six question marks.

The man on the left turned the screen back to face him.

His fingers were on the keyboard, but he wasn't typing.

He was looking at something on the screen, his gaze sweeping from top to bottom as if he were reading a long passage of text.

The woman on the right has her hand on the closed laptop.

Her thumb unconsciously rubbed against the computer case, making a very soft sound.

Then the man on the left looked up.

He didn't look at Lin Che.

He was looking at the head of national security.

The two stared at each other for about two seconds.

The head of the National Security Bureau nodded slightly.

The man turned his head and looked at Lin Che.

This was the first time he had looked directly at Lin Che since joining 607.

His gaze was steady, devoid of hostility or curiosity.

It is an evaluative perspective.

Then he spoke.

"Narratives cannot be used as a basis for judgment."

His voice was not loud, his speech was slow, and he pronounced each word very clearly.

His Mandarin is standard and he has no accent.

"We need to see."

Six characters.

Narratives alone cannot serve as a basis for judgment; we need to see the facts.

This is the first thing the male member of the technical assessment team has said since joining 607.

The two clauses together contain fifteen characters.

There were no formalities, no preamble, no "excuse me" or "could you please?"

Lin Che nodded.

"Can."

He bent down and took out a third folder from the briefcase at his feet.

Real-time demonstration solution.

Third edition, final draft.

He opened the folder and pulled out an A4 sheet of paper, which contained a flowchart of the Abyss cutting process.

It's simple, just three steps: input public data, system calculation, and output prediction results.

The flowchart was hand-drawn with a black pen; the lines aren't perfectly neat, but the labels are very clear.

He placed the paper on the table and pushed it so that all three people could see it.

"My first step is to use the historical data you provided to conduct a backtracking verification."

He glanced at the checklist in front of the head of national security.

"I can use Abyss to backtrack to each of the six records in your table."

"Enter publicly available data prior to that point in time and see if the system's prediction matches the actual trend."

The head of national security pushed the comparison table to the center of the table.

Six horizontal lines pointing upwards.

Lin Che took his laptop out of his briefcase.

It's silver-gray, and the screen is a size larger than two ThinkPads.

He turned on the computer, entered the password, and waited a few seconds.

A dark-background interface appeared on the screen, with a data input field on the left and an output panel on the right.

There is a line of small white text at the top of the interface.

Abyss v3.7 | Offline Mode | Last sync: 2022-12-08

This is the first time Abyss has appeared on the screen in room 607.

It's not a screenshot, not a printout, it's alive.

Lin Che turned the computer 180 degrees so that the screen faced the three people opposite him.

Article 3. March 2021. South American soybean futures.

He started filling in numbers in the data input field.

Publicly available data prior to March 1, 2021.

Brazil's monthly rainfall report, South American soybean export statistics, historical prices of the Chicago Board of Trade, and global shipping indices.

Four sets of data, each followed by a source label.

It took about forty seconds to fill it out.

He pressed the calculation key.

The output panel on the right side of the screen flashed, and a line of text and a number appeared.

Forecast: 87.3% probability that South American soybean futures will see a single-day increase of more than 3% within the next 14 days.

The head of the national security agency leaned closer to look at the screen.

His thin-rimmed metal glasses reflected the light from the screen.

"87.3%."

He repeated the number in a low voice.

It wasn't a question, it was confirmation that I had seen it correctly.

Lin Che cleared the data in the input field and started filling in the second field.

July 2021. Rare earth sector.

Input data: Ministry of Industry and Information Technology's rare earth production quota announcement, Inner Mongolia environmental protection inspection schedule, monthly rare earth permanent magnet index, and year-on-year growth of rare earth production in northern China.

Forty seconds.

Calculation.

Forecast: 79.1% probability of a concentrated portfolio rebalancing signal appearing in the rare earth sector within the next 10 days.

Article 3.

January 2022. US dollar exchange rate.

Input, operation.

Forecast: Probability of a trend reversal in the US dollar index within the next 21 days: 91.6%

91.6%.

This is the highest number out of the six.

It was also the one with the longest lead time, eleven days in advance.

The man on the left stopped looking at his screen from the second line onwards.

He stared at Lin Che's computer, watching every input and every output.

His fingers were on his keyboard, but he wasn't typing.

He's watching.

The woman on the right reopened her laptop.

She wasn't taking notes.

She is comparing.

On her screen, Lin Che couldn't see the specific content, but he could see that the screen was divided into two columns, left and right.

The data in the left column scrolls upwards, while the data in the right column remains stationary.

She was comparing Abyss's output with the data she had brought herself.

Lin Che then proceeded to implement the fourth, fifth, and sixth points.

All six items have been completed.

The six predicted probabilities, ranging from 79.1% to 91.6%, all point in the same direction: Abyss's predictions match the actual trend.

more importantly.

Abyss's predicted timeline corresponds perfectly with the "abnormal trading time" recorded by CR-1247.

It's not a coincidence, it's a subset correspondence.

Every abnormal transaction recorded by CR-1247 can be found in the corresponding output record in Abyss's prediction log.

The trades were not based on insider information, but on Abyss's predictions.

CR-1247 has been changed from "evidence" to "a subset of the Abyss prediction log".

The person in charge of national security looked away from the computer screen.

He didn't look at Lin Che; he looked at the man on his left.

The man nodded.

The gesture was even smaller than the usual nod from the head of national security.

The head of national security glanced at the woman on the right again.

The woman's fingers remained on the keyboard; she neither nodded nor shook her head.

Her gaze remained fixed on her own screen.

The head of national security picked up a pen and wrote a few more words in the lower right corner of the comparison table.

It's next to the two words I wrote before.

Lin Che still couldn't see clearly.

All six verifications have been completed.

The radiator in room 607 seems to be making a bit louder than before.

The desktop was covered with files.

What happens next is no longer up to Lin Che.