Genius is just the threshold for me!

"no."

Lin Weiliang shook his head and opened the detailed backend code: "Xingdong Company's AI chips are in mass production. In the past two months, we have obtained a batch of AI chips and, together with CPU chips, built and designed a land warfare scenario with 300 units."

However, it was found that it was fine to simply let the warhead system perform simulations and then let humans make decisions.

If you let the system make its own decisions, the decisions it makes will inevitably be wrong from time to time.

The algorithms and systems of the Warhead system are very different from those of single combat systems such as drone swarm systems.

If it's a swarm system, let alone a batch of AI chips and CPUs working together, even ordinary CPU chips combined with Lin Dong's algorithm can handle a swarm of three thousand swarms.

The Warhead system integrates a variety of different equipment and troop types.

300 units correspond to 300 different pieces of equipment and personnel, and each piece of equipment and personnel has a different combat style and system.

It involves multi-system collaboration and joint operations, which means that the hardware performance requirements for system computation and simulation are not on the same level.

"Is this really that weird? Dad, did you guys make a mistake when you wrote the algorithm...?"

Upon hearing this question, Lin Dong took Lin Weiliang's notebook and quickly flipped through the backend code. After reviewing all the code and algorithms, he found absolutely no problems.

It is even several levels more advanced than the original algorithm he provided.

It is clear that Comrade Lin Weiliang, the team from the National University of Defense Technology, and Dr. Feng He have put a lot of effort into optimizing and improving the system.

"Can we see the data for the entire system?"

Lin Dong confirmed that there were no problems with the code and algorithm, then turned to look at Lin Weiliang.

Since the code algorithm is fine, the problem must be with the data.

After all, AI has the ability to learn autonomously. Algorithms are only the core and foundation, and external functions still rely on data training and learning.

It's like a child with exceptional talent who could become a world-class scientist.

But if he is taught all sorts of opportunism and trickery from a young age, he will definitely not grow up to be a scientist, but just a hoodlum.

"Yes, I have the highest authority."

Lin Weiliang nodded, accessed the intranet, and jumped to the Warhead system at the National University of Defense Technology via a military line. He then entered his access password letter by letter and number.

Di-

The entire data module of the warhead system, as well as the various modules and access simulation information, are all displayed on the screen.

Lin Dong checked the access information, carefully examined each functional module, and finally opened the specific data...

Ok?

With just one glance, Lin Dong frowned, then quickly dragged the mouse, and a large amount of data flowed past his eyes like a waterfall.

"There's something wrong with these data..."

Lin Dong closed the data interface, opened the historical data of the Warhead System, and checked the previous update iteration records.

"Nine cycles in two months, Dad, are you guys working this hard?"

Lin Dong flipped through the records one by one, and when he got to the second record of iterations, the mouse suddenly stopped.

"Found the problem?" Lin Weiliang's eyes widened when he saw Lin Dong stop.

"The data here is starting to go awry..."

Lin Dong highlighted one of the data sets, examined it carefully, and then looked at the data from the first iteration. Suddenly, he frowned and asked, "Dad, have you been letting the system iterate through deep learning all along?"

"right!"

Lin Weiliang nodded: "Establish the relevant rules, and then let the system learn on its own."

"Let's revert to a previous save and start over."

Lin Dong shook his head: "AI does not inherently possess autonomous decision-making capabilities. If Zhanlu is to make decisions, it cannot rely solely on AI's own deep learning. It must combine traditional manual optimization with intelligent algorithms."

In this process, in addition to the rules of the battlefield, the experience and knowledge of combat experts, as well as human combat psychology, can also be incorporated.

Furthermore, avoid making all-encompassing decisions at once; instead, adopt a phased and scenario-specific approach. This will ensure more stable and realistic decision-making in the warhead.

"Understood."

Lin Weiliang suddenly realized: "I was using the experience of the wolf pack system and the bee swarm system to train the warhead before, and I forgot about the differences between the two."

After so many years of hard work and study, he became the King of AI. Lin Weiliang understood Lin Dong's meaning.

That is, because the War Skull needs to make comprehensive decisions, it must guide its evolution step by step and integrate human experience and growth, rather than letting it learn autonomously.

The drone swarm system and the robotic dog's Wolf Warrior system, being single-system platforms, can continuously iterate and evolve by allowing the systems to learn autonomously, compete with each other, and step on each other's feet.

This is because the underlying AI system lacks the ability to make decisions. There are two other types of systems: one is intelligent decision-making, and the other is the difference in results caused by game theory.

"Different AIs have different characteristics, and therefore require different training methods and data."

Lin Dong gave a brief explanation and returned to his seat. Just as he sat down, an email popped up in Wang Xingguo's inbox.

Ding-dong-

"...An email from an external organization?"

Lin Dong noticed that the icon was different from the one in the intranet emails, so he moved the mouse and clicked on the email content.

[Chief Engineer Shen of AVIC Science & Industry]: Chief Engineer Wang/Chief Engineer Jiao, hello. One of our institute's cruise missile projects is encountering difficulties with the engine.

I just saw your paper on miniaturized missile engines published on the intranet. Could you please tell me if your technology can meet the performance requirements of this missile?

Just then, the phone next to me vibrated a few times, and a message popped up.

【001】: The unlock password for AVIC Science & Technology's email data. There are three sets, in the order of one, three, two. You need to be quick. The file has a security time lock set. It will expire if it is not unlocked within five minutes outside their internal network environment.

[001]: Regarding the engine issue, if you see no problems, just proceed with it.

[56]: Should we proceed directly? Are you sure there are no problems?

【001】: No problem, I don't know how to handle it either. Just send a copy to 624.

【56】: Okay then.

Lin Dong replied with a question, and upon looking at the password sent by Wang Xingguo, he couldn't help but mutter to himself.

"It's a time lock again."

"Why are you being so cautious about missiles?"

As he spoke, Lin Dong opened the attachment, entered the three sets of 18-digit passwords in the order of 1, 3, 2, and then pressed the Enter key.

Di-

The confidential data file was opened, and a large amount of required information and data filled the entire screen in an instant.

Codename: Eagle Strike

Specifications: Supersonic cruise anti-ship missile, launchable from multiple platforms including submarines, destroyers, and small surface vessels, with a maximum takeoff weight of 2.1 tons and a 300 kg warhead.

The requirements are a range of 1000 to 1500 kilometers, a cruising speed of Mach 5 to 8, a terminal speed of Mach 10, and terminal anti-ship tracking capability...

"This is……"

Upon seeing this performance parameter table, Lin Dong immediately thought of a more complete code name.

——YJ-19

That hypersonic anti-ship missile that debuted in 25, using an air-breathing scramjet engine, with a terminal speed of Mach 10 to Mach 12!

"It is said that this uses rocket ignition for takeoff, followed by solid fuel scramjet... Well, solid fuel has a higher energy density, so the terminal speed can be increased even further."

Lin Dong muttered to himself and looked at the specific data below.

For them, when conducting research, all superficial parameters and drawings are one-sided; only concrete data can reflect the most realistic situation.

"This data requirement... the super-charged engine of Hot No. 4 should be fine, but we need more specific data."

After reviewing the data and pondering it for a while, Lin Dong opened the email reply interface and replied to the email in the same tone and wording as Wang Xingguo. At the same time, he did not forget Wang Xingguo's instructions to copy a copy to Jiao You.

[Wang Xingguo]: Chief Engineer Shen, the preliminary technical analysis of the solution can meet the requirements, but more specific data is needed for further verification.