Reborn in 2010, the AI ​​in my mind is too powerful.

If Chen Yiran wants to take revenge on Icahn, it's definitely not going to work through capital. Trying to outmaneuver the dogs of Wall Street with capital is like a starving dog trying to escape to the toilet—it's suicide.

Therefore, we still need to focus on technological aspects.

Icahn's frantic actions stemmed from his belief that Mingdao's stock price was indeed being suppressed too low, and he was confident he wouldn't lose money.

Mingdao's three major businesses, particularly the EDA business in manufacturing, are indeed in dire straits. Industry insiders believe that it may be completely wiped out in another two years.

However, the remaining two businesses, PCB design and embedded systems, while not the industry leader, are still ranked third or fourth in the industry.

Chen Yiran planned to secretly get a head start and take a gamble, hoping to use Jasmine's powerful capabilities to quickly break into these two businesses and then sell the more mature products to Zuken and Siemens before Icahn.

This would prevent the two companies from acquiring any more of Mindray's business lines, resulting in a huge loss for Icahn's investment in Mindray.

Moreover, this has a side benefit: if this really works out, then Mingdao will be a company known worldwide as a doomed entity. When things have fallen on hard times, Chen Yiran can still acquire Mingdao's EDA business.

The price at that time might be surprisingly low, since only his company would have the incentive to acquire it. Other companies wouldn't be able to compete with FlowEDA.

Chen Yiran began searching for relevant information online.

His understanding of PCB design basically stopped at the term "circuit board," which he only heard in videos. His knowledge of it was just a green board with a bunch of dots soldered on.

Embedded systems were something I had absolutely no idea what.

Fortunately, we have Jasmine. Technical skills are interconnected, and Jasmine had already accumulated enough knowledge during her previous studies. Although these two business lines are quite different, she still knows a little bit about them.

At least knowing what to learn is already quite an achievement.

Under Jasmine's guidance, Chen Yiran began intensively searching on Google.

Finally, a very long list of things to watch was compiled, with the IPC-2221 General Design Standard and the IPC-2222 Rigid Printed Circuit Board Specification being the top priorities.

These can be purchased from the Institute of Electrical and Electronics Engineers' (IEEE) paper repository. It's outrageous; just two PDFs, and they're nearly 500 pages long.

Then he went to Amazon and bought the e-book of Howard Johnson's "High-Speed ​​Digital Design".

This book, published in 1993, is a classic in the field of signal integrity. It specifically discusses high-speed signal integrity, covering everything from transmission line theory to the parasitic capacitance of vias and the ground bounce effect, with a dense array of formulas.

Industry peers regard it as the bible of digital circuit design, the best reference book for high-speed digital circuit design from beginner to expert, and it is recognized as the most valuable textbook in the fields of signal integrity and EMC.

After that, you need to look at the publicly available information from various manufacturers, such as Samsung and Micron's optical networks, various preset information for DDR3, and the routing guidelines for PCIe and SATA.

In short, after a whole day, Chen Yiran, who was dizzy and lightheaded, was able to say to outsiders that she was an expert in the relevant field.

Chen Yiran felt he had reached his limit, but still asked Jasmine first, "How should I put it? I'm so sleepy and need to take a nap. Can you give me a general idea of ​​what we can handle? Let me tell you, what you learned before can always be vaguely reflected in me a little bit."

Now I feel like it's really working. For example, with these things, although I still don't understand them at all and can't explain why, I just feel like I can tell you can handle it.

Is it possible that you simply trust me? But based on my general feeling, it seems fine. However, I can't be sure, as I haven't been thinking at full speed.

Jasmine's words were enough. After thinking for a moment, Chen Yiran first called Li Feifei of Kai Deng and explained his requirements.

Li Feifei was clearly surprised to hear Chen Yiran's request: "Why don't you release the news of our reconciliation yet? Is there something you need on your end? We are truly cooperating in good faith. If you have any concerns, we can discuss it further."

Chen Yiran said weakly, "It has nothing to do with you. Our cooperation is fine. How about this: our core lawyer can discuss the cooperation agreement first, but we can agree to formally implement it next year. I need a few months; I have some business issues on my side that I need to resolve."

At the same time, I hope we maintain the current ongoing litigation, so that it doesn't appear that we've reached a settlement. Any excess legal fees incurred by your company can be billed, and I'll offer you a discount when we cooperate next year, okay?

Li Feifei understood; Chen Yiran was trying to scam them. Although he didn't know the specifics, they were about to become business partners, so he had to give him some face. He generously said, "No need, we can afford the legal fees. This is fine too; we can go to Xinshi these next few months and see if we can extort some money from them."

Regardless of the official details, we hope to finalize the cooperation agreement as soon as possible and sign it. Frankly, only after this is confirmed can we formulate our fiscal year plan for next year.

"Okay, in detail. Have your legal counsel communicate with my lawyers. Gao Dewei will handle the handover himself. But it must be kept confidential. Try to avoid meeting in person for any communication. I've heard that Wall Street isn't honest and often sends people to keep an eye on you."

"Don't worry, we'll go abroad and find a place to sign the contract. It'll be so sudden that even if someone follows us, they won't be able to tell."

After hanging up the phone, Chen Yiran thought about her company's situation again. If she suddenly recruited people in related industries now, it would definitely not be able to hide it from outsiders.

Others might not be sure what he's up to, but Icahn will definitely figure it out. If he accelerates the sale of the business, his revenge plan will fail, and the subsequent software won't fetch a good price. Even if the acquiring party has plenty of cash, it's unrealistic to make two large acquisitions in quick succession.

When I told Jasmine about it, she quickly came up with a very good idea.

[I met Lin Xiao at the DAC Design Automation Conference before; his research focuses on mixed-signal layout optimization. We could register a company in China and have him build a small team there. Finding relevant personnel at Tsinghua University shouldn't be difficult, and they don't need to be exceptionally skilled. They just need to be able to assist me.]

Yeah! I thought his skills were pretty good before, and I was planning to recruit him to the US after he graduated. Now we can just have him lead a team in China first, doing some support work, and then bring him to the US later. No matter how good Icahn is, he can't reach China.

Relieved, Chen Yiran didn't bother going to the bedroom and fell asleep on the sofa in the living room.

In the following days, Chen Yiran began to make crazy improvements to the existing PCB and embedded technologies at an astonishing speed.

Of course, he was only responsible for drinking glucose; the main thing was jasmine.

Jasmine has been very happy these past two days, feeling like she's in a casino.

In His words, He was very good at these two technologies, so much so that He could handle them with ease without much thought.

Essentially, all PCB design specifications can be translated into mathematical constraints. Differential pair traces must be of equal length and within ±2 mils. The total length of DDR3 data lines cannot exceed 2.5 inches, adjacent layer traces cannot be parallel for more than 500 mils, and the characteristic impedance of signal lines must be controlled within 50 ohms ±10%.

These are rules that engineers would manually check one by one in traditional tools, but I can directly write them as cost functions and insert them into Grey. Of course, to sell them separately, we would need to develop a completely new software.

"Okay, but we can't use a bunch of small animals this time. Those are just extras; we'll sell them later. Let me think... let's use fruit. We'll call this PCB software 'White Lychee'."

Yes, you can start thinking about the follow-up issues now. This is much easier than chip placement. Chip placement requires consideration of variables such as heat dissipation, timing, and Vt mixing.

PCB routing doesn't have that many variables; it mainly involves physical constraints and signal consistency.

By running a PCB constraint test on DeepFlow using the white-labeled PCB tracer, the routing time for a 12-layer high-speed digital board can be reduced from several days to less than twenty minutes. Furthermore, signal integrity simulation no longer needs to be performed separately after routing; we can revolutionize the process by simulating while routing.

"I regret what you just said! If I'd known it was this easy, I would have done this first and made some money. This would have been a piece of cake, making four or five hundred million dollars for nothing. What's wrong with you? Why didn't you warn me sooner? Embedded systems? Is embedded systems really that easy?"

【╮(╯▽╰)╭, you were the one who said we should move towards chip manufacturing. The chip industry is already at the pinnacle of the technology chain. Now, doing anything else feels somewhat backwards compatible to me.】

Embedded systems aren't too complicated. I can directly generate a precisely cyclic software processor model based on the specifications. The entire bus, peripherals, and interrupt controller are simulated in parallel on the GPU. Running dozens of virtual development boards simultaneously on a single Fermi GPU is no problem. This is revolutionary; although we're targeting Mindray, I feel we're going to kill off all the other players in the industry.

"what?"

Traditional virtual prototyping runs purely on the CPU, with four cores being roughly the maximum. Debugging, recompiling, reloading, and switching back to simulation takes five to ten minutes in a single cycle. Our approach is different; it runs dozens of instruction set emulators in parallel on the GPU, allowing a single card to emulate hundreds of ARM cores, reducing compilation to testing to tens of seconds.

Moreover, it doesn't require actual silicon wafers; development can still proceed even if the chips don't flow back. The current industry leader is Intel's Wind River; they'll go crazy. Of course, Mingdao is in an even worse situation. According to the data we found, they are the third largest in the industry, relying on low profit margins and high volume to sell 3 billion units, yet only generating a little over 100 million in revenue annually.

"Alright, let's get it done quickly."

Chen Yiran continued to lie flat, ordered a bunch of food, and started soaking in the pool.

Drinking glucose all the time is efficient, but eating more delicious food will help soothe his tired nerves.

Chen Yiran chatted with Jasmine while eating ice cream. According to Jasmine, Chen Yiran reading novels or watching dramas while working would affect his computing power.

So while working, Chen Yiran could only chat with Jasmine to pass the time.

"Embedded systems are used so widely, how come the third-largest company in the industry can sell for 3 billion?"

Yes, but they all target very low-end industries, so it's all about low profit margins and high volume. For example, mobile phone parts, or even components for small electronic devices. They're not even as good as Microsoft, which makes ATMs and cash registers in addition to PCs.

Not to mention industry leaders like Intel's Wind River and Green Mountain, the industry's second-largest player. These are used in the aerospace and military industries; the F-35 and F-22 both use their technology.

"Oh, is that so?"

Chen Yiran took a sip of Monster Energy drink, casually replied, and then suddenly jumped up.

"No, no, you mean this thing has military applications???!!!"