Customized training

Finally, the model training for the Xianxia novel dataset was completed.

Lin Yuan insisted on using dialogues from Xianxia novels to test the performance of DS-LLM.

Because AI Large Language Models have already undergone sufficient training, if it were conventional dialogue, Lin Yuan would have to spend a lot of effort to accurately measure the model's actual performance.

Therefore, for simplicity, Lin Yuan directly switched to a completely unconventional dialogue logic to test the model's performance. And Xianxia novels, due to their immense quantity, were naturally the best training material.

Lin Yuan immediately loaded the trained model's weight file and then input a request: "List a daily dialogue scene between cultivators for me."

After a brief thought, lines of text continuously appeared in the model's dialogue window...

【Cool pavilions among the mountains, where floating clouds gather, and three figures arrive, treading through the mist.

The cultivator surnamed Li raised a teacup, a faint light swirling from the green jade thumb ring on his finger: "Junior Brother Zhao's Qi Sea has recently shown signs of tidal resonance; it seems he has cultivated the 'Azure Wood Art' to the third stage."

The young man surnamed Zhao smiled wryly and shook his head: "Yesterday, during my breakthrough, I was almost consumed by the Earth Fiend Qi. If not for the elder in charge bestowing this bottle of Ice Soul Pills..." As he spoke, he took out a carved jade bottle from his sleeve, the emerald pill patterns emanating a chill in the morning light.

"Hmph, your Danxia Peak is quite generous." The female cultivator in plain clothes lightly flicked the sword box at her waist, and seven phantom bronze miniature swords instantly circled around her. "To forge this Big Dipper Sword Array, I even exchanged the three-hundred-year-old Huangjing in front of my cave dwelling for star sand."

...............

Lin Yuan watched with great interest for a long time.

If he only looked at the snippets without context, he found that the Xianxia fragments written by the AI really had a strong flavor. At least, they were much more flavorful than many miscellaneous pseudo-cultivation texts.

For Xianxia novels, the most important thing is to highlight the Xianxia flavor. The naming of cultivation techniques and elixirs, as well as the characters' speech, demeanor, and objectives, made Lin Yuan feel that this model was not just randomly piecing things together.

"It seems the rice bowls of those authors who write generic stories are in jeopardy." Lin Yuan vaguely felt that many procedural jobs would inevitably be replaced by AI in the future.

Especially pure service industries that do not involve device operation, such as customer service, copywriting, and formulaic writing—this seems to be an irreversible trend.

He suddenly remembered his entrepreneurial journey with Tesla Food Delivery.

The scene before him was strikingly similar to his previous food delivery venture.

In the food delivery industry, once AI was introduced to standardize various processes, it was followed by the introduction of an assembly line mechanism, and then individuals were replaced by giants.

So, now that AI has intervened in the service industry, once AI models become powerful enough and standards for specific scenarios are established, couldn't the remaining work also be mass-produced in an assembly-line fashion?

For example, the core logic of a server attracting customers at a restaurant entrance is actually not difficult. While a real person might seem more personable, in such low-income positions, a real person's emotions and work commitment are uncertain factors.

Conversely, robots have no emotions.

Robots can formulate different welcoming remarks for different groups of people based on algorithms. They can introduce healthy dishes to the elderly, popular specialty dishes to young people, and even offer various small toys to children.

In short, the demand model of diners is not difficult; it can be easily covered by Large Language Models.

In that case, the server at the entrance who attracts customers would no longer be needed.

Next are the servers who deliver food. Since these servers not only deliver dishes but also need to fulfill diner requests such as hurrying up the food or adding broth, they also need to have conversational capabilities.

Then, sufficiently powerful Large Language Models can also cover these needs, making a large number of restaurant servers no longer essential positions.

Next is the kitchen, and also the people who wash vegetables in the back kitchen.

If extended further, various counter consultation positions seem to have no reason to exist.

For example, positions like bank tellers.

These positions seem to have regional characteristics because local dialects differ, and even more frustratingly, some dialects have unique sentence structures and word usage logic that cannot be simply translated.

The most troublesome part is that simple dialogue logic is clearly insufficient when rural branches face elderly grandpas and grandmas.

However, AI Large Language Models support further training.

Lin Yuan had already thought of it: as long as he collected local dialect dialogues to create a dataset, just like he had done, and this dataset covered the conversational characteristics of people from all age groups.

For example, specifically collecting the dialogue logic characteristics of middle-aged and elderly people, then feeding this dataset to the Large Model, and subsequently training a specialized language model with local characteristics for each region.

Even a virtual Xianxia world can be trained with a unique dialogue logic, let alone the various existing regions in reality.

What then?

Lin Yuan continued to imagine.

Grandpas and grandmas would find the robots at the counter, covered in artificial silicone skin, more approachable because these robots wouldn't give them attitude and would patiently guide them again and again without getting annoyed.

Then, unmanned 24-hour banks would be born.

If this logic were applied to all service industries, wouldn't a large number of people become redundant?

Lin Yuan was somewhat dazed; he couldn't tell if this was good or bad.

Because this line of logic, if developed further, would lead to a terrifying conclusion—would robots manage humans better?

This was not Lin Yuan's baseless imagination, because from his experience, he really didn't have much confidence in the creature called "human." At least for robots, for silicon-based life, there would only be bugs, but no lies.

Lies are the cancer of human society.

Lin Yuan felt his brain was about to explode for a moment.

"Alas, why must I have a compassionate heart?" He could only mock himself like this.

In the food delivery industry, he could block the path of giants crushing individuals in advance by using open-source algorithms and introducing the concept of an algorithm tax.

But what about other industries? There are so many niche industries in the world; can they all be blocked? Even more terrifying, even if Dragon Country blocked it, what if other countries opened it up?

Once other countries fully open up, under the powerful productivity boosted by AI, human labor is minuscule.

Tell me, in front of an algorithmically perfect robot chef, what is a human chef?

Don't talk about human cooking having a unique touch; as long as it undergoes a Turing Test, it will all be utterly shattered.

As long as there's a curtain in front of you, as long as the AI algorithm is sufficiently perfect, and as long as mechanics and electronic control keep pace with AI's development, you cannot possibly know whether it's a human or a robot behind the curtain.

It's like a beverage prepared one-to-one according to the Coca-Cola recipe; how do you distinguish it from Coca-Cola?

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

He remembered that saying—is the mission of carbon-based life to give birth to silicon-based life?

"Nonsense!!!" Forget it, he finally stopped thinking about it.

Lin Yuan's feeling at this moment was like that of humans centuries ago facing the vast ocean, having no idea what lay hidden in the sea ahead, what was on the distant continents, or what the Age of Discovery would bring.

At this moment, he felt like he was standing at the dawn of human industrial civilization.

Industrial civilization brought beauty, but also chaos.

This world, after all, was about to change.

Lin Yuan finally pulled his thoughts back. He wasn't sure if he would live to witness that moment. So, he decided to finish the short-selling task at hand first.

Money, small money, was the most grounded thing at this moment.