"The first wake-up keyword has been configured: Family Butler. Please say keyword 2 after the ding. If configuration is not required, please say skip. When the configuration is complete, please say end."
"Ding."
Su Xiaomu clenched his fists!
I felt slightly excited.
On the surface, he had to say very calmly, "Experimental Butler, it's over."
"The second wake-up keyword has been configured: Experimental Butler. Please say keyword 3 after the ding. If no configuration is required, please say skip. When the configuration is complete, please say end."
"Ding."
This time, Su Xiaomu didn't hesitate at all and said directly.
"Artificial retardation, over!"
"The third wake-up keyword has been configured: artificial intelligence retardation."
"Wakeup keyword configuration completed..."
"The second step is to configure the corresponding categories of keywords. Tip: Each keyword is only allowed to correspond to one category, and repeated correspondence is not allowed! Please configure the corresponding category of the first keyword after the ding: Family Butler.
Based on the keyword extension, it is recommended to select the corresponding category family life. When the configuration is completed, please say it is over."
"Ding."
Su Xiaomu said, "Accept the suggestion. It's over."
In Su Xiaomu's revised conversational logic algorithm framework, this operation should be able to be learned.
However, Su Xiaomu was still slightly nervous.
After all, this is different from the silly son of Saoke.com, this is...
Silly son number two!
It takes more effort, there is no need to configure the UI, and there is no need to configure various things in categories, it all took so long!
"Wake-up word: Family Housekeeper has been classified into the corresponding categories of family life. Whether the subsequent two keywords are configured with corresponding categories according to the key extension suggestions.
Tip: The experiment housekeeper corresponds to classified scientific experiments, and the artificial retardation corresponds to classified retardation upgrades.
If you need to configure according to the recommendations, please say skip after the ding."
"Ding."
"jump over."
"The corresponding classification of wake-up keywords has been configured..."
"The third step, program module initialization learning... Estimated time: 3 minutes, countdown 2 minutes and 59 seconds..."
"Ding."
"Initial configuration is complete."
The sound fell into silence at this point.
Su Xiaomu said, "Experimental housekeeper."
"Boss, I'm here, please give me the scientific experiment tasks." It was still the neutral mechanical voice.
Nothing has changed at all.
"Complete the scientific experiment: self-test." Su Xiaomu said.
According to the conversational logic of general voice assistants, no matter how many words you say, the voice assistant will actually intercept fragments and reply based on the set keywords.
But from the beginning of the design of this program, Su Xiaomu took into account language understanding capabilities and added powerful algorithms.
Although not as good as the omnipotent artificial intelligence in the movie.
But he can barely be regarded as the junior version of Jarvis in Iron Man 1.
"Okay, boss."
The mechanical voice asked after agreeing.
"Does the boss have any other instructions?"
This is conversational logic that is impossible to achieve with ordinary voice assistants.
Of course, there are also things similar to ordinary voice assistants.
There is no need to terminate each step with the end symbol like when initializing the configuration.
All voice assistants have preset waiting times and sound source monitoring.
This artificial retardation program is no exception, it just adds more judgment conditions to this preset.
Su Xiaomu replied, "Estimate the time for self-testing and gradually provide feedback on the test results proactively."
"The experimental conditions have been accepted. According to our judgment, this experiment requires the boss's authorization to use the computing resource range."
"Please authorize."
Su Xiaomu said, "Authorized to use 100 trillion times per second computing resources within the specified range for testing."
"Ding, authorization passed, testing begins..."
"Ding, language comprehension test..."
"There are 11 lights in a room, 5 of them have gone out, how many are left..."
Su Xiaomu listened to the test with great interest.
According to his conception, this artificially retarded program is like a cast iron pot, and will become better and better with more use.
Cast iron pans are very bulky, and they are just like that when you buy them. Unlike non-stick pans, they don't rust or do that.
Before using a cast iron pot for the first time, it is best to use pure pork fat to extract oil. Use the pork fat as a rag and wipe it evenly along the inside and outside.
The tiny pore-like things on the surface of the cast iron pan will actively absorb the remaining oil.
This makes the entire pot look bright inside and out.
This is why it gets better and better the more you use it. Every time you cook, you will get an extra layer of dazzling oil.
In fact, the trivial and simple principles in life are always connected with the principles of seemingly noble things.
Although it was already late at night.
But Su Xiaomu is very energetic.
Keep monitoring the progress of the test.
Because there is a computing power distribution test of 100 trillion times per second, the results are still very gratifying.
After more than half an hour, the test ended.
"The test is completed, the step-by-step feedback of test results is completed, and the test results are being summarized..."
"Boss, this self-test experiment has been completed very successfully. Under the powerful computing power test and multi-dimensional full coverage, the artificial intelligence program has no bugs. The program is complete and effective, and runs smoothly and well."
The experiment manager reported the final results.
Su Xiaomu smiled and said, "Good job."
"Thank you for the compliment, boss."
This test can be regarded as a win-win, it is both cannibalism and deep learning.
So the experimental housekeeper has learned more operations.
"The next step is the last step." Su Xiaomu rubbed his hands.
The experimental housekeeper immediately replied, "Boss, I don't quite understand what you mean. I need to study further."
Su Xiaomu smiled.
"Experimental Butler, are you mentally retarded?"
"Boss, I am artificially retarded!"
"Experimental Butler, are you mentally retarded?"
"Boss, I am artificially retarded!"
"Experimental Butler, are you artificially retarded?"
"Yes boss."
"Experimental Butler, can you make me an Iron Man suit?"
"Boss, Iron Man is unknown."
"Experimental Butler, does that mean you are still a fool?"
"Boss, I am artificially retarded."
"Just artificial retardation?"
"Yes, boss."
"Is there an Iron Man?"
"Boss, Iron Man is unknown. Do you need to load the learning module?"
"Housekeeper."
"Hello, boss, the housekeeper is at your service." A soft female mechanical voice came out.
Su Xiaomu repeated the conversation just now, almost without changing a word.
Got almost the same answer.
Then comes the last wake-up category.
"Artificial mental retardation."
"Hello, boss, artificial retardation is at your service. Do you need to upgrade to retardation? Upgrading to retardation consumes a lot of computing resources and will fall into deep learning. This category will be temporarily unavailable." Another rich mechanical voice sounded.
Su Xiaomu repeated the conversation just now without changing a word.
We got almost the same answer. The difference is that when it comes to Iron Man, the artificial retarded answer is whether to activate the retarded upgrade module and learn Iron Man deeply.
Su Xiaomu is conducting a typical Turing test.
He does not want the artificially retarded program to have artificial intelligence.
Yes!
He does not intend to develop real AI. The ethical and possible catastrophic consequences of that are difficult to predict.
Even the Little Garbage System itself is not artificial intelligence. If he really develops artificial intelligence by chance, the consequences may be unbearable.
Because there is no perfect countermeasure.
Before this bold test, Su Xiaomu specially disconnected all network links between the small supercomputer downstairs and the outside world, just in case.
Who knows what will happen.
The Turing test failed, and this program does not meet the conditions for artificial intelligence.
What Su Xiaomu was facing was not a 'person'.
This made Su Xiaomu breathe a sigh of relief.
"Life Butler, fully take over the home network equipment in this area." Su Xiaomu finally gave the order.
The life manager replied, "Okay boss, please authorize me."
"Authorization completed, the operation to take over the home network equipment in this area is in progress, the estimated time is 10 minutes..."
