Xueba's computing system

Chapter 268 Customized Training

Finally, the model training of the Xianxia novel dataset is finally completed.

Why did Lin Yuan insist on using dialogues from fantasy novels? It was to test the performance of DS-LLM.

Because the AI large language model has already undergone sufficient training, if it is a regular conversation, Lin Yuan needs to spend a lot of effort to measure the actual performance of the model.

Therefore, for the sake of simplicity, Lin Yuan directly switched to a completely unconventional dialogue logic to test the performance of the model. And Xianxia novels are naturally the best training material because of their large number.

Lin Yuan immediately loaded the trained model weight file, and then entered a request - list for me a daily conversation scene between cultivators.

After a brief thought, paragraphs of text kept appearing in the model's dialogue window. . .

[Clouds gather over the mountain pavilion, and three figures emerge from the mist.

The cultivator surnamed Li held up the teacup, and the green jade ring on his finger glowed faintly: "Junior Brother Zhao, there has been a faint sound of tides in your sea of qi recently. It seems that you have cultivated the Qingmu Jue to the third level."

The young man surnamed Zhao smiled bitterly and shook his head: "I was almost attacked by the evil spirit of the Earth Dog when I was breaking through the barrier yesterday. If it weren't for the deacon elder giving me this bottle of Ice Soul Pill..." As he spoke, he took out a carved jade bottle from his sleeve. The green pill pattern was chilling in the morning light.

"Hmph, you people in Danxia Peak are really generous." The female cultivator in plain clothes flicked the sword box on her waist, and seven bronze swords suddenly surrounded her. "In order to cast this Beidou sword formation, Lin even exchanged the three hundred year old Huangjing in front of the cave for star sand."

. . . . . . 】

Lin Yuan read it for a long time with great interest.

If you only read the excerpts without looking at the context, he found that the fairy-tale excerpts written by AI were really quite strong. At least they were much more interesting than many random pseudo-fairy-cultivation novels.

As for Xianxia novels, the most important thing is to highlight the Xianxia flavor. The names of the skills and elixirs, as well as the characters' speech, behavior and purpose of their actions, made Lin Yuan feel that this model was not put together randomly.

"It seems that the jobs of those authors who write clichéd articles are in jeopardy." Lin Yuan just had a vague feeling that many process-based jobs are destined to be replaced by AI in the future.

Especially for pure service industries that are independent of equipment operation, such as customer service, copywriting, and formulaic writing, this seems to be an irreversible trend.

He suddenly remembered his entrepreneurial journey as a Tesla food delivery company.

The scene before him was so similar to his previous food delivery entrepreneurial process.

In the food delivery industry, once AI is introduced to standardize various processes, the next step is to introduce an assembly line mechanism, and then individuals are replaced by giants.

Now that AI has been introduced into the service industry, once the AI model is powerful enough and standards are established for those specific scenarios, the rest can be mass-produced in an assembly-line manner.

For example, the core logic of a waiter soliciting customers at the entrance of a restaurant is not difficult. It seems that real people are more humane, but in such a low-income position, the emotions and work commitment of real people are uncertain factors.

Robots, on the other hand, have no emotions.

The robot can formulate different welcome words for different groups of people based on the algorithm. It can introduce its own healthy dishes to the elderly, its own popular special dishes to the young, and introduce itself to the children and give them various small toys.

In short, the diners’ demand model is not difficult and can be easily covered by a large language model.

In this case, the waiter who solicits customers at the door will no longer be needed.

Next comes the waiters who serve the dishes. Since this type of waiter not only serves the dishes, but also has to meet some requests from diners, such as urging the dishes to be served or adding soup, they also need to have a conversation function.

Then if a large language model that is powerful enough can cover such needs, a large number of restaurant waiters will no longer be necessary positions.

Next is the kitchen, and there are people washing vegetables in the back kitchen.

If we expand further, it seems that all kinds of counter consultation positions no longer have any need to exist.

For example, positions like bank teller.

This type of position seems to have regional characteristics, because the dialects in different places are different. What's even more tricky is that some dialects have unique sentence structures and word logic, which cannot be simply translated.

The most troublesome thing is that the simple conversation logic is obviously not enough for the elderly in rural outlets. However, the AI big language model supports further training.

Lin Yuan had already thought of it. Just like what he had just done, he could collect the local dialect conversations and make a data set. Then this data set would just need to cover the conversation characteristics of people of all ages.

For example: we can specifically collect the conversation logic characteristics of middle-aged and elderly people, and then feed the data set to the big model, and then train a dedicated weighted language model with local characteristics for each place.

Even the virtual world of immortals can be trained to develop a unique dialogue logic, not to mention the various regions that already exist in reality.

And after that?
Lin Yuan continued to imagine.

The elderly will find the robots covered with artificial silicone leather covers on the windows more friendly, because these robots will not give them the cold shoulder, but will patiently guide them over and over again.

Then, unmanned 24-hour banking was born.

If this logic is applied to all service industries, then a large number of people will no longer be needed.

Lin Yuan was a little dazed. He couldn't tell whether this was good or bad.

Because if this logic continues to develop, it will lead to a terrifying logic - would robots be better at managing humans?

This was not Lin Yuan's imagination, because based on his experience, he had no confidence in creatures like "humans". At least for robots, for silicon-based life, there would only be bugs, but no lies.

Lies are the cancer of human society.

For a moment, Lin Yuan felt like his brain was about to explode.

"Oh, why should I have a compassionate heart?" He could only laugh at himself like this.

In the food delivery industry, he can block the path for giants to crush individuals in advance by opening up algorithms and introducing the concept of algorithm tax.

But what about other industries? There are so many sub-industries in the world, can they be blocked? What's more frightening is that even if Longguo blocks it, what if other countries open it up?

Once foreign countries completely open up, human power will be insignificant under the powerful productivity enabled by AI.

Just think about it, compared to a robot chef with perfect algorithms, human chefs are nothing.

Don’t talk about the unique characteristics of human cooking. Once you pass the Turing test, you’ll be screwed.

As long as there is a curtain hung in front of you, as long as the AI algorithm is perfect enough, and as long as the mechanics and electronic controls keep up with the development of AI, it is impossible for you to know whether there is a person or a robot behind the curtain.

It's like a drink prepared according to the Coca-Cola recipe in a one-to-one ratio. How do you distinguish it from Coca-Cola?

Suddenly, Lin Yuan thought of a question - why can't humans be replaced?

He remembered the sentence - the mission of carbon-based is to give birth to silicon-based?
"Bullshit!!!" Forget it, he was too lazy to think about it in the end.

Lin Yuan's feeling at this moment was that hundreds of years ago, when humans were facing the vast ocean, who knew what was hidden in the ocean ahead, who knew what was on the distant continent, and who knew what the great geographical discoveries would bring.

At this moment, he was standing at the moment when human industrial civilization began.

Industrial civilization has brought beauty, but also chaos.

This world is going to change eventually.

Lin Yuan finally withdrew his thoughts. He was not sure if he could live to witness that moment. So, he should finish the short-selling task first.

Mane and Xiao Qianqian are the most down-to-earth at the moment.

(End of this chapter)