Chapter 412: The Problem Has Been Solved
When Jonathan and Lin Ran’s dialogue frequently mentioned Old John Morgan, he was living a life like MacArthur’s in Tokyo.
Ever since Yanjing and Washington reached an agreement, Old John Morgan had been staying in Tokyo, at most flying to Shanghai for a face-to-face meeting with Lin Ran.
The route from Tokyo to Shanghai is too close.
Old John Morgan did not live in any embassy or official residence, but instead booked the entire top floor of the Imperial Hotel.
Ever since Tokyo changed prime ministers, he met daily with Japan’s financial magnates leaders, faction elders, and new cabinet ministers, trying to exert his own influence on this country.
And as he led Wall Street conglomerates to repeated successes in Taipei’s financial market, especially after the recent harvest that could be called a perfect storm, his influence in the global financial world had reached its peak.
Old John Morgan felt that his previous life was simply too boring; this was what life was. Using a allusion he recently learned from ancient Chinese texts, his cooperation with Lin Ran was: seeking a kindred spirit in high mountains and flowing water.
Of course, if his days were enjoyable, that meant someone else was not; the unhappy ones were Taipei’s financial practitioners. Taipei’s local companies, from finance to employment to the consumer market, were comparable to the Asian Financial Crisis.
Even more fatal was that, compared to just being targeted economically like in the Asian Financial Crisis, this time they faced double blows economically and politically, double blows with no end in sight.
After receiving Lin Ran’s phone call, Old John Morgan sat across from Lin Ran in Lin Ran’s office the very next morning. Lin Ran poured red wine into the glass in front of him, “Please!”
Old John Morgan picked up the wine glass, closed his eyes and smelled it: “Lafite? How did you know this is my favorite among high-end red wines? Though it seems all Chinese people like this one, but the vintage isn’t great; it’s not the legendary 1959 or 1961?”
Lin Ran smiled, because our first meeting was at the Russell Erskine Hotel, Huntsville’s high-end hotel, where Old John Morgan drank Lafite at the time.
As for why 1961 and 1959 are legendary vintages.
In the spring of the former, Bordeaux suffered an extremely severe frost that destroyed a large number of grapevine buds. The subsequent flowering period was also cold and rainy, leading to extremely low grape yields. However, from August to the October harvest season, the weather became perfect, resulting in this batch of grapes having flavor compounds at an unprecedented, extremely concentrated level.
The latter had hot, sunny, and very dry weather throughout the entire growing cycle. That year’s grape yield was normal, but the sugar content was extremely high, and the fruit flavors were extremely rich.
As for 1960, it was a terrible vintage for the grapes needed for wine; even the wines produced by Lafite winery were lackluster.
“Lafite, 1960.” Lin Ran raised his glass and said faintly, you don’t need to know the reason.
“Professor,” Morgan said after finishing the red wine, his face beaming with irrepressible pride: “I must say, last week was our highest return investment this year—no, this decade.”
He picked up the encrypted tablet on the table and pulled up the latest battle results report.
“The events around Japan completely shattered the last and most stubborn psychological defense line of the Taipei financial market, which is the regional joint defense expectation.” Morgan’s tone was like doing financial report analysis.
Globally, the only ones qualified to hear Old John Morgan do financial report analysis were Lin Ran and big T.
“When the market finally realizes that no one will come to defend them, capital flight is no longer just outflow, but an avalanche.
Over the past five days, the Taiwan Stock Weighted Index has cumulatively fallen 12%, foreign capital net outflow has exceeded 20 billion US dollars, and the New Taiwan Dollar to US Dollar exchange rate has broken through the 32 mark.”
“Our alliance fund,” he smiled: “through perfect operations in the derivatives market, has gained an additional 6.5 billion US dollars in profit in this week.
Their stabilization funds are now like a desperate gambler, throwing their last chips on the table, but unfortunately, we can see all their cards.”
“Congratulations, Mr. Morgan.” Lin Ran’s tone was calm without waves, “It seems the task of dismantling the stage has basically been completed. Next, just wait for you all to take away what should be taken away and give us what should be given to us.”
Morgan nodded: “Of course, we won’t stop any of their local companies from relocating to the mainland.”
Old John Morgan used “mainland.”
Lin Ran continued: “Elon in Washington has encountered a little trouble. He successfully exposed the old NASA’s corruption and incompetence to the sunlight.
But he discovered that after demolishing a dangerous building, he found himself standing on a pile of ruins, without even a usable brick in hand. He needs a general contractor who knows how to build skyscrapers, a partner who can reintegrate, digest the remnants of companies like Boeing and Lockheed, and forge a new, efficient commercial empire.”
Old John Morgan laughed loudly: “Professor, Elon is something. I didn’t expect him to think of asking you to be the lobbyist.”
He restrained his smile and said seriously: “Of course, Professor, we certainly need to talk with Elon. Whether it’s us, Boeing, Lockheed, Raytheon, the capital behind them all has a large number of orders from NASA. Musk can mess around, but he can’t affect our business.”
Lin Ran nodded: “Including Collins Aerospace, Pratt & Whitney, and a series of other companies.”
These are all Raytheon’s subsidiaries, and Raytheon and its subsidiaries are also among NASA’s most important suppliers.
And behind Raytheon Company stands the Adams family in Old John Adams Morgan.
Old John Morgan: “We’ve long known Musk would come to talk to us; it’s just a matter of time, and who holds the initiative.
He doesn’t want to lose the initiative, so he found you, you found me, I go find him, and he’ll think the initiative is in his hands.”
After Old John Morgan finished speaking, his gaze sharpened for a moment, then softened again: “But Professor, for the sake of this glass of Lafite today, I’ll send someone to talk to him. It’s just that the vintage is a bit regrettable; 1960 isn’t a great year.”
At the same time, with Spring Festival approaching, Lin Ran thought, why do we always concentrate manpower to solve major problems at this time of year?
Last time it was solving the mass production breakthrough of the Shockley-Queisser limit for photovoltaic limits, making the photoelectric conversion efficiency of single-layer solar energy exceed 33.7%. Now, photovoltaic modules with new structures achieving 60% photoelectric conversion efficiency are about to enter large-scale mass production. Shareholders in China’s stock market photovoltaic sector are all eagerly waiting, calculators nearly worn out, waiting to see who fires the first shot.
This time, it’s to solve the lens problem for lithography machines.
All along, the biggest difficulty in China’s lithography machine project has been the light source and mirrors.
Without Zeiss selling to ASML that projection lens system, which accounts for nearly 40% of the machine’s cost and consists of more than a dozen perfect mirrors, China’s EUV lithography machine, even if it solves 99% of the problems, cannot go into production.
And with Canon’s FPA-1200NZ2C 5nm NIL lithography machine already installed and debugging completed in Shanghai, Lin Ran knew the time to solve the lithography machine problem had arrived.
Organized by Chinese officials, personnel and venues were arranged in advance at the Shanghai Microelectronics Institute, including Canon’s NIL lithography machine also placed at the Shanghai Microelectronics Institute.
Only after all the relevant people arrived did Lin Ran enter.
“As everyone knows, our biggest gap with EUV lithography machines is in the lenses. This is not something we can overcome in a short time with manpower, time, and will.
Lenses with errors at the picometer level, and more than ten of them, with coupling relationships between each one—this is a physical limit, a chasm in materials science and precision processing, not something willpower alone can overcome.”
The meeting room was silent.
Everyone already knew what Lin Ran said. Whether in internal meetings or reporting upwards, they had repeated it countless times; this is our biggest difficulty, which can only be ground out slowly with time.
Pile it up with quantity to achieve a good result.
Do we still need you to come and say it?
However, in the semiconductor field, although Lin Ran is not an expert in semiconductor production and manufacturing, as a top expert in artificial intelligence and the proposer of the left-right brain chip concept, he is definitely not an outsider.
“Perhaps our thinking has been wrong from the beginning.
Why do we have to polish a perfect lens? Can we print a perfect lens?”
“Print?” The experts present looked at each other; they seemed to grasp a bit of inspiration.
Lin Ran walked to the curtain and pulled up a structural diagram of a metasurface lens.
“Traditional lenses rely on geometric optics refraction. We polish the curvature of glass so that light bends paths when passing through media of different thicknesses, ultimately converging to a point. This is a physical shaping process. But metasurface lenses rely on wavefront optics phase modulation.” He explained.
“We don’t need to change the path of light; we just need to change the pace of the light waves.
Each nanometer antenna on this planar substrate acts like a phase delayer.
When a beam of parallel light waves passes through it, some parts are delayed by a quarter wavelength, others by half a wavelength.
By precisely controlling the phase delay at every point, we can reshape a plane wave, after emission, perfectly into any shape we want, such as an ideal spherical wave, and focus it perfectly.”
Lin Ran continued:
“Polishing lenses is an empirical physical problem testing centuries of process inheritance.
Designing this phase delay matrix is a pure mathematical problem testing computing power and algorithms, which happens to be my strength.”
Among the experts present, one who understood mathematics asked weakly:
“General Manager Lin, I’ve thought about what you said. For a 300mm diameter lens, if we arrange 5nm-level antennas on it, we will face more than 10 to the 14th power, that is, one hundred trillion independent computing units.
Each unit has multiple variables like shape, dimensions, rotation angle, etc.
This is a typical NP-hard problem with a nearly infinite solution space.”
He continued, lowering his head to calculate on his notebook:
“Our fastest supercomputer now takes months to simulate a nuclear fusion reaction. To use traditional electromagnetic simulation and optimization algorithms on this supercomputer to find the global optimal solution for the perfect phase function that Comrade Lin Ran wants would probably take a thousand years of nonstop running.”
Lin Ran clapped lightly: “Well said. What’s your name?”
The expert raised his head: “My name is Wei Zhe.”
Lin Ran grinned: “Good name; it differs by just one character from TSMC’s current chairman’s name.”
Wei Zhe touched his head embarrassedly: “I majored in mathematics as an undergraduate, switched to optics for my master’s, and have always been interested in mathematics.”
Lin Ran nodded and walked to the podium: “Exactly, he said it very well.
What if we can use mathematical transformations to turn this from a search problem into a solving problem?
The difficulty we face now is finding the optimal arrangement for trillions of independent nanometer antennas in real space.
This computational load is astronomical.”
But the essence of optics is waves.
Any complex wave can be decomposed into a series of simple plane waves in Fourier space.
The perfect focusing function we want is actually a very concise and elegant mathematical expression in Fourier space.
So the key to the problem is no longer how to arrange the antennas, but whether we can find an efficient algorithm to build that unique, deterministic mathematical bridge between the physical structure in real space and the target function in Fourier space?”
The experts present felt like they were listening to a book from heaven; only Wei Zhe vaguely grasped it.
“If I had no confidence, I wouldn’t have called everyone here.
In astrophysics, for handling telescope image distortion, there’s an algorithm called phase recovery algorithm. Combining ideas from unitary transformations in quantum computing, I’ve developed a brand new algorithm.”
Lin Ran clicked the mouse, and the PPT switched to the next page: “Iterative Fourier Transform Constraint Algorithm, IFTCA
This algorithm no longer blindly brute-force searches like a headless fly; its logic is more like solving a Sudoku puzzle.
Brute-force search tries all numbers 0 to 9 in every cell until finding the answer.
Our IFTCA algorithm provides the computer with a set of logical rules.
Simply put, we first define the desired answer in Fourier space, project this ideal answer back to real space through an inverse transform, obtaining a preliminary but error-filled antenna structure.
We then use physical constraints to correct this structure, erasing all answers that violate physical laws, then project this corrected, physically realistic structure back to Fourier space through a forward transform to see what it looks like now.
Finally, we compare the result with our originally desired perfect answer, calculate the error, use this error as the correction parameter for the next iteration, and repeat the entire process.
Through thousands or tens of thousands of iterative corrections between ideal and reality, this algorithm doesn’t traverse the entire solution space; it follows the path of steepest gradient descent, deterministically and convergently, toward that unique optimal solution satisfying both optical ideals and physical reality.
So Engineer Wei, with this algorithm, we don’t need a thousand years.
With the help of existing supercomputers, solving the perfect matrix for the entire lens takes only three months.”
Just as the audience below was in an uproar, Lin Ran continued:
“And we’ve already calculated it long ago, just waiting for Canon’s NIL lithography machine to arrive, waiting for everyone to arrive, ready to start working.
The lens is just one link; there are many more links waiting for everyone to solve.
Additionally, to avoid repeated Dragon’s Roar incidents, including the recent Huawei technology leak, we’ve specially invited everyone here to work concentratedly for a while.”
Wei Zhe below the stage was stunned, because only his mathematical literacy allowed him to understand how complex designing this algorithm was—not much easier than a supercomputer calculating for a thousand years. He finally understood why the man was called a god.
