茶杯狐 -CUPFOX

Chapter 101: Origin

Not until the sun rose and the morning rays spilled over the earth did Xiang Kun finally stop editing the PRD. Looking at the plan that had entirely deviated from his initial idea, he couldn’t help but rub his brows, as if his old problem had recurred…

The Al assistant he envisioned, considering the numerous functional modules and level of complexity, had turned out to be a large-scale project. Some functions were even not yet mature with no available technology for reference. It would pose a challenge even to giant corporations like Baidu, Google, and Microsoft, let alone to him alone.

If he had been presented with such a requirement document at work a month ago, he would have probably bounced it back to the product manager, suggesting that he rewrite it when sober.

But this document was his own creation, and the requirements were all his…

That’s why he was somewhat persistent, wanting to give it a try, to see to what extent he could achieve.

However, when he returned to what he was really good at, he felt more directly the changes in his brain’s way of thinking and computing speed, thereby making him more excited.

The current are already quite intelligent. For instance, if you asked, “What’s the weather like tomorrow?”

The Al first converts your voice into text through speech recognition technology, then analyzes your intention, extracting keywords to “fill the slots”.

There are two “slots” in this sentence, one is the time “tomorrow”, and the other is the event “weather”.

So it executes the command, calls up the weather forecast program, looks for tomorrow’s weather, and provides the result.

Some Al would even continue the conversation by asking, “Are you going out tomorrow?”

If a new keyword is received, a new round of slot filling begins, assisting the user in handling various tasks.

Ultimately, the Al relies on keywords to trigger preset feedback. While the rate of correctly identifying the actual intention is getting higher and higher, it all seems more and more “alive” as the feedback is becoming increasingly comprehensive and human-like.

However, the present Al is still far from achieving self-conscious artificial intelligence.

Xiang Kun did not have the unrealistic expectation of creating a self-conscious artificial intelligence all by himself. He just wanted to make use of

“reinforcement learning” combined with “deep learning” methods so that the

Al assistant could continuously help him improve the “Vampire Mutation Model”. Based on the output results of the model, the Al would purposely conduct searches on the internet for medical, biological, and physical studies and even news, selecting potentially useful information for him to review. Subsequently, it would improve future search rules and decision-making according to his actual choices.

In terms of lifestyle and training, the Al assistant would also provide suggestions based on the ‘Vampire Mutation Model” and spontaneously point out some problems he might overlook. It would propose new directions to try, assist him in conducting experiments, and record parameters.

Moreover, he hoped that the Al assistant could help him analyze and monitor his daily activities and online operations for any abnormalities that might lead people to notice something unusual. It would then provide solutions to help him keep his mutation a secret.

After spending an entire day, Xiang Kun had a rough framework for implementing this and marked the majority of the functionalities along with numerous technical difficulties to be tackled.

However, the deeper he delved into it, the more he realized that this project was not something he could complete on his own in a short amount of time.

Although there were many open-source codes and datasets available online for various functions, a lot of time would still be required to integrate them. Not to mention the many functions that he would have to develop from scratch.

While he was indeed far more capable now than he used to be, even claiming that one of him could match ten of his old selves would not be an overstatement, and he could work continuously 24 hours a day.

But many constraints were still beyond his current capabilities. To illustrate, his existing hardware facilities wouldn’t support the planned development, requiring not only a new computer but also a larger bandwidth, higher computing power of cloud services, and even customizing some terminal devices.

“It looks like I either need to find a shortcut, or get extremely rich, just like Tony Stark or Bruce Wayne…” Xiang Kun rubbed his bald head, murmuring to himself.

But he didn’t plan to go back to the initial need and make a rudimentary Al assistant to use for now. After fantasizing about a grand feast, seeing the cold hard bread in front of him, he couldn’t bring himself to eat it except for extreme hunger.

He planned to start with the voice synthesis functionality. After all, it was essential for any Al assistant, no matter the final design.

Just like in video games, before distributing specific attribute points, players always 11Ke to generate tne race nrst. ror an WitnOUt a pnyslcal entity, its voice can indeed be regarded as its face.

The voice synthesis technology has developed to a fairly mature stage today. Apart from the market giants, there are numerous open -source toolkits available online for direct use.

However, Xiang Kun’s obsessive-compulsive disorder struck again. Sometimes he was dissatisfied with the smoothness of the voice, other times with the timbre of the existing female voice.

At 11 a.m. on September 13, Xiang Kun received a delivery from Jingdong. Thus, he temporarily halted his struggle with synthetic voice and began to unpack the parcel with anticipation.

The magnetic field detector and microscope were delivered together. After verifying that the items were intact and undamaged, Xiang Kun prepared to start testing with the magnetic field detector.

This detector could measure low frequencies from 5Hz to 400Khz and high frequencies from 300MHz to 3000MHz. It also featured a magnetic field testing mode with a precision of IV/m and 0.011-1.

First, he brought the handheld detector close to various electronic devices to confirm its readings were accurate and the device was functional.

Then he brought the detector near the Canary’s cage. The canary, since he had opened the cage two days ago, had stayed inside without ever leaving, still “behaving” itself within its territory.

Looking at the canary that had let out two chirps, Xiang Kun pondered for a moment and decided to go to the kitchen and use the remaining two rabbits for his experiment.

Unfortunately, during the entire process when he used “intimidation” to exert various effects on the rabbit, the detector’s readings showed no significant changes.

“Was my previous speculation incorrect? Or is it just that the detector I bought is too cheap, and the sensitivity is not high enough?” Xiang Kun couldn’t help but mutter a few sentences, then placed the detector aside and started fiddling with the microscope.

He didn’t immediately examine his own cells or blood. Instead, he first drew a tiny amount of blood from the rabbit. Without preparing a smear, he just dropped a bit of saline and considered the slide as prepared, ready for observation.

The microscope he bought came with a monitor linked to the eyepiece which could be directly connected to a computer screen. So, he observed directly on the computer screen.

There was nothing special about the rabbit’s blood. It was only for the purpose of getting familiar with the microscope that he hadn’t used since high school, as well as for a simple reference, so he didn’t observe it much.

Next, he located several microscopic images of human blood online, watched some videos of others observing human blood with a microscope, again for reference.

Eventually, he took his own blood and started the observation.

Xiang Kun did not prepare a smear, wanting to observe his blood more naturally. He knew that dramatic changes would occur in his blood within dozens of minutes..