Ten thousand I am across the heavens

Most information and resources related to hacking techniques are not open source. It's also very difficult to obtain such information on the open internet. Once everyone knows that this stuff can be sold for money, even if someone manages to buy it, they won't choose to share it for free.

This means that, theoretically, even if you could buy pirated software, you still have to pay for it. His classmates, wanting to improve their hacking skills, often needed to purchase large quantities of programs they needed online.

Mo Chengzi was quite familiar with the prices of these items, as this body had purchased quite a few before. Otherwise, he wouldn't have achieved such a good ranking in this category.

The original owner of this body was not a genius; these things in front of him were not something one could learn just by wanting to.

Okay, now it's his turn, and he's quite experienced in this area. First, he opens the error message interface and begins to carefully read through the programs he previously used. He wants to understand how they work. This isn't difficult for him. Although the original owner hadn't read these materials, he generally remembered the interfaces and functions of these programs. By working backward from the clues, reconstructing their operating principles isn't difficult.

He quickly grasped the basic operating principles of these devices and began fixing the bugs that had appeared earlier. These were the effects of a virus infecting a portion of his software after the counterattack. Based on these effects, he could deduce the enemy's attack methods. How were they damaging his machine?

"Looking at those programs that look like they've been hit by bullets, most of them crashed because they received a huge amount of data in an instant. So in this virtual world, ordinary bullets probably represent conventional data attacks. Using massive amounts of junk data to make the program malfunction is one of the simplest, most brutal, and also most effective attack methods."

The performance of my equipment was severely damaged, not just because of the data breach, but also because the attack data contained virus-infected data. So, while the bullets looked identical, some were actually poisoned, and others were tampered with.

What truly broke through the defenses were these bullets with special effects. After the virus spread, it severely hindered the operation of the entire program, which caused a serious bug in the program on my body.

Now that we know the source of the damage, it's not difficult to deduce the type of damage and counter it. Hmm... however, at this level, surpassing this world is still beyond my current capabilities. Judging from the content the original owner downloaded, this world's information technology level, while not high, isn't something I can easily crush with just some random code.

I need to create some coding tools for myself that are ahead of their time. I need to improve my efficiency.

An AI-assisted compilation system must be developed first. Even a rudimentary one would improve my compilation efficiency by hundreds of times. Otherwise, relying on manual coding, I simply cannot catch up with those classmates who buy ready-made solutions.

With this in mind, Mo Chengzi began searching for AI models online. Sure enough, he discovered that the vast majority of artificial intelligence products on the market were closed-source. A search of the history of internet development revealed that in its early days, most AI systems were open-source. Users could download large models and deploy them locally. Usage was limited only by computing power. What content to generate and how much content to generate depended entirely on the user's ideas.

At that time, various artificial intelligence programs were still under development, requiring extensive testing by testers and modifications by independent developers. This led to continuous updates and iterations. However, as times have changed, these artificial intelligence programs have matured and entered the commercial realm, rendering open source obsolete.

Almost overnight, major companies began reclaiming copyrights from privately owned AI models, which were also subjected to cyberattacks. Internally installed backdoors almost entirely collapsed during a single update. Except for a very few servers that were completely shut down, virtually all privately owned AI models were replaced immediately.

Meanwhile, major companies have also launched affordable and user-friendly commercial-grade AI systems. Because they don't require local deployment, the pressure on computing power is virtually nonexistent, and there are no configuration requirements. Furthermore, to initially capture the market, they were made extremely inexpensive.

It's fair to say that during that period, computing power was poured in regardless of cost. This massive, all-out effort forcefully squeezed out the space for locally deployed AI models, completely eliminating the survival space for small businesses that wanted to use locally deployed models to create AI applications.

Almost all AI models today are monopolized by major companies. They are all closed-source models running in the cloud. Ordinary people have no access to the models themselves, let alone the ability to use or modify them. They must pay exorbitant prices, subscribe to memberships, and purchase usage time to select and use the various AI models released by these large companies.

And this doesn't seem like the time to conquer the market. The price is very low. After almost achieving a monopoly, the price of individual use of AI models is quite expensive.

These individual hackers can't even use it. Firstly, they're just ordinary students; unless it's for exams, they can't afford such expensive equipment in their daily training. Secondly, because it's cloud-based, all computing power is subject to company review. Even if they had the funds to deploy such models for hacking, if their work is found to be detrimental to the company or potentially cause financial losses, their accounts will be frozen immediately, even with methods to bypass the review process and local decoding. This significantly impacts efficiency, and the scope of content that can be bypassed is limited. Many useful functions will be restricted.

These restrictions make it difficult for amateur hackers to access artificial intelligence, which in turn widens the technological gap between companies and the general public year by year. Commercial-grade AI can write its own programs and optimize them repeatedly, making it many times more efficient than humans. With the adjustments made by top software engineering teams, its progress is something amateur hackers simply cannot keep up with. However, even so, there is still a huge gap between their technical level and that of AI outside the Great Firewall.

This is also why a few top-tier hackers can still cause big trouble for companies. They can risk their lives outside the firewall to try to acquire knowledge about artificial intelligence. There are even legends of a lucky top hacker who once brought back a newly created artificial intelligence from outside the firewall. Now he has become one of the legends of Night City. No one knows where he is now, only that he is still active on various online platforms.

Mo Chengzi was aware of this; his memory couldn't possibly hold the complete contents of an entire artificial intelligence model. Even if he could conjure thirty hands, typing out an artificial intelligence model by hand was a pipe dream.

Just when he was at a loss about his initial model, the system, which had been quiet for a while, suddenly popped up a red dot.

"Loading the Internet scrap collection system, loading, loading complete!"

Mo Chengzi: "" He opened the Internet Scavenging System and saw a bunch of URLs. After opening them, he found that they were links to some cloud storage companies that were on the verge of bankruptcy. This immediately caught his eye. This stuff was extremely useful to him.

The cloud storage company, which is on the verge of bankruptcy, was once very successful. In fact, it stored a large amount of prohibited pirated data. Now that it is on the verge of bankruptcy, it is resorting to any means to stay alive, relaxing its review standards and allowing a large number of black videos, black Super Dreams, and black programs to be stored normally.

The system identified some of the most unscrupulous companies for him, and the original model that Mo Chengzi wanted could be obtained simply by purchasing a membership. Although the membership originally cost a hefty 300 yuan, and was quite insidious—it had to be canceled within three minutes or it would automatically deduct the cost for a twelve-month subscription—it was still a hundred times cheaper than the official version!

Five hundred gigabytes of data were quickly packaged and downloaded to Mo Chengzi's hard drive. This was just the most basic model, but it was also the source of all artificial intelligence. Almost all existing models were upgraded from this original model.

"Tsk tsk, the system is still sensible! There are probably only a handful of people in the world who possess this now, and it has actually fallen into my hands."

The reason why very few people in the world possess this device is not because of its high level of secrecy, but because it has no economic value anymore, just like how very few people in the 21st century possess mobile phones.

Without upgrading the model, the electricity it consumes alone cannot be recouped by the content it produces, let alone the cost of server rentals. This is why no one has tried to monetize it.

Mo Chengzi did a quick calculation and realized that the original model couldn't generate any revenue. Fortunately, he knew that the Shanhe Republic actually possessed a model training method that was hundreds of times more efficient than existing methods, even for this old, original model. It could be evolved with minimal computing power and resources.

These things are actually of little use to the development of the Republic of Mountains and Rivers. But there's always a group of people researching them. These guys are a bit like people with doomsday syndrome. They fantasize that one day they might encounter a scenario where internet technology regresses by 100 years, and only they retain the memory. Like a group of post-apocalyptic survival fanatics, they research how to build the internet from scratch.

Although Mo Chengzi wasn't a die-hard enthusiast in this field, some of these guys' research had already gone quite far. When it became popular, he had also come into contact with the related technologies, and although he only borrowed the research ideas, it was enough to be relevant here.

Training a model naturally requires a lot of computing power, and the computing devices on his body are all civilian-grade, equivalent to a personal mobile phone on Earth, not even comparable to a personal computer, so of course they can't be used for training.

However, the school's servers are still open to students. The rental price is also slightly cheaper than commercial-grade servers.

Mo Chengzi decisively rented the school's commercial-grade server on the spot, and at the same time, reluctantly spent 5000 yuan to purchase a one-week temporary membership for the company's artificial intelligence service. He then started unlimited data sessions.

I created a data interface to connect this original model with the latest cutting-edge artificial intelligence models.

One of the core aspects of this technology is plagiarism. It allows you to directly reverse engineer and copy each other's original code from the conversations between artificial intelligences, and thus reverse engineer the training process.

This technology was originally designed for human use, requiring a large amount of data on people and human activities to be input into the model for training. However, since more advanced, finished artificial intelligence already exists, training with the AI's own training data is more efficient and effective than manually searching the entire internet for data.

The relevant technology has not yet been developed in this world. Therefore, the company has not implemented any regulations to prevent the theft of model code.

This gave him the opportunity to rapidly upgrade the model's performance within a short training period, jumping from version 5.0 to version 7.0. It was around version 6.5 onwards that artificial intelligence began to gradually become closed-source, with content no longer publicly available. Only after this version was it considered ready for commercial application.

Although mainstream AI versions used by major companies are now at version 10.0, far surpassing the intelligence of version 7.0, the device he has has barely reached a commercial level. It provides some assistance during his use, and that's about it. He still has six days of membership left. While the school server rental is expensive, his limited funds are far from enough to support long-term rentals; he'd basically need to invest all his income just to rent it for one day. It's only thanks to his overpowered abilities that he has no daily expenses; otherwise, he truly couldn't afford the server.

His behavior, however, was seen as a lack of ambition by his classmates. Everyone else spent the day trying to outmaneuver their opponents, starting over after each failure, and upgrading their equipment.

He seemed completely discouraged after his first failure. He didn't even bother repairing his equipment or formatting the virus-infected parts. He just sat there looking lost in thought, silently typing away at his code. Many students whispered that he was broke and taking on freelance work to earn money. They said he was basically using the school's server computing power to make a living.

[To be continued]

Please add this to your favorites and vote for it! (End of Chapter)