If you cheat, money will follow

The time is advanced to 2016.

Google's AlphaGo competes with Lee Sedol in the Go game...

In the second game, on step 37, AlphaGo made a strange move.

At the time, it was considered a very stinky move.

As a result, after review, we found that this was a key step in laying the foundation for victory.

And this step adopts an idea that has never been seen before in the history of human Go.

"According to this logic, once consciousness is formed, coupled with massive data, it will far exceed human IQ." Mr. Xu said.

Li Fei: "...!"

It's normal for Li Fei to be nervous.

Everyone is very nervous.

Including Iron Man Musk.

As early as 2010, Musk paid attention to artificial intelligence, or general large models.

In North America, they like to call it AGI (Artificial General Intelligence).

At that time, Musk met Hassabis at an investment fund conference organized by a billionaire.

Hassabis is the founder of DeepMind.

The two met at a meeting and had a great conversation after the meeting, but they regretted meeting each other so late.

Musk invited Hassabis to visit his SpaceX company.

Hassabis happily went because he needed money to train the model.

While eating in the SpaceX cafeteria, Musk revealed his ambitions to Hassabis.

He solemnly introduced the Mars colonization plan.

He believes that this plan is the key to solving a series of problems such as overpopulation, energy crisis, and war crisis on the earth.

Hassabis strongly agrees with Musk’s point of view.

But the premise is: "The robot that destroys the earth will not chase it from the earth to Mars..."

A word to awaken the dreamer.

This completely touches on Musk’s intellectual blind spot.

Musk agrees.

He believed that this danger was likely to happen, and he was as anxious as the domestic Internet giants competing for business in the wet market.

Then Musk made a bet on DeepMind.

Four years later, in 2014, DeepMind was acquired by Google for $6.5 million.

Another year later, in 2015, at Musk’s birthday dinner.

Musk and his friend Larry Page turned against each other.

According to Musk, it was the safety issue of artificial intelligence that became “the last straw that broke the friendship between the two.”

Musk believes that artificial intelligence research should be public welfare and non-profit, and become a tool for human beings rather than replacing humans.

Page believes that humans will eventually integrate with AI and become intelligent creatures, competing for resources, and the most powerful intelligent creature will win.

Musk believes that intelligent life is meaningless because humans are already extinct.

Page believes that Musk is a speciesist.

Musk suddenly felt like he had seen through the other person, and they parted ways in shame.

At the end of the same year, Musk participated in the creation of OpenAi and invested US$100 million, becoming the co-CEO of OpenAi with Altman.

Back to the topic.

When it comes to the safety issues of large models, Liu Chaoyang is well-informed and proposes:

"We can reach out to Hinton, Suzkovi and Krichewski..."

"They study convolutional neural networks, and they must have a unique understanding of this algorithm."

"After I contacted them last time, although they did not agree to our invitation, we have been keeping in touch through the Internet..."

"I think as long as we show them what we've developed, they might agree to come and take a look."

Whether it is Hinton, Suzkowei or Krichevsky, they are all pure researchers.

He is also a typical "effective altruist".

"Effective Altruists" is a pan-academic research group.

Created by AI scientist Yudkovsky.

The main point is that on the one hand, we want to build a great, universal, and universal tool through AI...

Help mankind conquer cancer and solve difficult problems such as climate warming;

On the other hand, they believe that AI will definitely attack humans if it is too smart and must be supervised.

"Effective altruists" and AI fusionists are naturally in conflict, but they are not completely diametrically opposed.

After all, the “effective altruists’” ideals of AGI require financial support from AI fusionists.

AGI is not something you can just talk about, and convolutional neural networks are not unique and complex algorithms.

The key is that computing power and data are still needed to support it.

In short, whoever has the money to accumulate computing power and data will be ahead.

At this time, Hinton was 70 years old and still working at Google.

However, his disc herniation problem is very serious and he cannot stand or sit for long periods of time.

He is quite critical of Google's commercial ambitions.

He has told the media many times that "I regret my life's work."

At the same time, I regret developing AI, which may bring disaster to mankind.

Sutskvi, student of Hinton.

Currently working at OpenAi, he is the chief scientist of OpenAi.

He doesn't take money seriously and has always had high vision for AGI.

Otherwise, he would not have given up his high-paying + stock position at Google in 2015 and cut his salary to join OpenAi.

Currently, OpenAi is in financial difficulties.

Because Musk failed to seize power, left the company, and withdrew US$1 million along the way.

CEO Altman is in frequent contact with various investment funds, mainly targeting Microsoft.

Microsoft is very interested, but doesn't like that OpenAi is a non-profit organization.

They hope that OpenAi will be a for-profit organization so that their investment will be returnable and valuable.

Altman actually also wants to turn OpenAi into a profit-making organization, such as setting up a profit-making company under OpenAi.

Only in this way can Altman obtain the shares and go public for cash.

He does not want to work for free and purely contribute to humanity.

Krichevsky, another student of Hinton, has contributed as much to convolutional neural networks as Sutskevi.

However, his life ideal is not here.

After the Lake Tahoe auction, he received US$4400 million of the US$1320 million, plus US$5 million from working at Google for five years, and he has achieved financial freedom.

In 2017, he chose to retire and stay away from artificial intelligence.

As for why you retired so young?

Probably not a simple financial freedom can explain it.

The next question comes.

Who is going to contact the three masters and apprentices of Hinton?

Chang Le, Liu Chaoyang and Mr. Ren are not suitable.

North America probably wants them to go.

Mr. Xu? Not enough cards.

"I'd better go." Brother Ma volunteered.

As the head of a social empire worth 6 trillion Hong Kong dollars and about 7600 billion US dollars, he has enough cards and even more power to speak.

Nor will it be a thorn in the side of North America.

It is also influential enough.

Because social networking, short videos, games, and e-commerce are all just successes at the application level. They are like castles made of sand that will collapse at the drop of a hat.

"Okay, Mr. Ma, thank you for your hard work." Liu Chaoyang said:

"I will make an appointment with them."

"It's okay, I just want to meet the great master in the AI ​​field." Xiao Ma said.