News
Grading Artificial Intelligence with Cryptocurrency Practices
Summary: Grading Artificial Intelligence with Cryptocurrency Practices Cryptocurrency and artificial intelligence, these two things seem completely different. They are both products after the birth of computers. That is their commonality. Bitcoin and Ethereum were born with silver spoons in their mouths because they have advanced attributes, which are both productivity and production relations. In the early ...
Grading Artificial Intelligence with Cryptocurrency Practices
Cryptocurrency and artificial intelligence, these two things seem completely different. They are both products after the birth of computers. That is their commonality. Bitcoin and Ethereum were born with silver spoons in their mouths because they have advanced attributes, which are both productivity and production relations. In the early days of artificial intelligence, it was more about technical research, but after ChatGPT, it caused people to worry that it would threaten human security. Some scientists have signed a petition calling for a ban on the development of artificial intelligence, just like the ban on human cloning. Cloning technology is a productive force, and the results are predictable; artificial intelligence is different. The scary thing about it is its "uncertainty." What will the future of artificial intelligence be like? Because these uncertainties belong to the field of production relations, it is not a simple technology, nor can it be prohibited.
Both are solutions to the problem of uncertainty
We introduced Bitcoin's method of handling uncertainty problems in the article "Learning from Satoshi Nakamoto to Handle Uncertainty Problems" (chainless.hk). Cryptocurrency has been exploring how to use computers to simultaneously solve the uncertainties of production efficiency and production relations since the 1980s. The concept closest to artificial intelligence is called smart contracts. Of course, the English terms for artificial intelligence and smart contracts are very different.
The similarity between them is that they are both automatic machines. Cryptocurrency studies how to obtain specific results in an uncertain environment, while artificial intelligence gives uncertain results in an uncertain environment. Which one is more difficult? Of course, it is more difficult to achieve specific results in an uncertain environment. Only when this problem is solved can humans safely use artificial intelligence as a tool. Because once the outcome is determined, there is no risk.
Under uncertain conditions, there are generally only selective results, and it isn't easy to have deterministic effects. Selectivity is an easy problem for humans, but it isn't easy to obtain exact results using machine intelligence methods. For example, a Byzantine general problem has stumped humans. Another example is the conversion of Ethereum from a POW consensus mechanism to an equity consensus mechanism (POS). The new mechanism is more influenced by people than the original POW consensus. How do we reach a consensus when someone is doing evil? If it is easy to use humans to verify, if They are found dishonest and do bad things, punish them, but this is difficult for machines. It took the Ethereum community over 5 years to solve this conversion; OpenAI has been run for just over 5 years.
POS is just a problem with blockchain. PoS first appeared in Peercoin in 2012; the history of GPT can also be traced back to 2012. At that time, Google Artificial Intelligence Laboratory (GCP) proposed an idea at an internal meeting to enable machines to "learn" language and grammar. In 2013, they released a GPT-based language model (Transformer) that learns language by encoding input.
These are two routes for humans to explore uncertain problems, but they face a common problem: How can humans trust the results of machines? Do you believe that machines will not do evil?
Setting the standard for artificial intelligence with cryptocurrency practices
Artificial intelligence creates fear in humans because it has no exact solution, while cryptocurrency is the opposite. How are cryptocurrencies made? Although there are many cryptocurrency projects, their methods are roughly the same. The summary process is as follows:
Rules for establishing procedures
Write open-source programs
Conduct program audits
Whether the results of the testnet test program meet the goals
Officially launched and executed automatically, without human intervention during the execution process
Get publicly verifiable results
According to the above process, in an uncertain environment, it is entirely automated by the machine, without human intervention, and the machine automatically produces the results. Human intervention only changes the program. Without changing the program, there will be no risk. In this process, machines are just man's faithful tools. However, when artificial intelligence can change the program independently, its risks will be completely uncontrollable, and it will deviate from the original intention of developing artificial intelligence. How will this be done?
When there are issues of uncertainty, the above process should have a review unit and scope.
When encountering dangerous problems, there should be apparent prompts.
Grading Artificial Intelligence Projects Using Cryptocurrency Practices
1. Five-star standard
When humans deal with uncertainty, they have adopted a classification approach in the past. This is just like the classification of movies. Movies above Category III are not suitable for children. Artificial intelligence also needs to be graded, which is based on the successful experience of cryptocurrency.
As the exact result is subject to the above cryptocurrency process, no approval is required. If it complies with the process, Give it five star.
2. Four-star standard
Its process is subject to third-party scrutiny if it's not made public. Since it is for civilian use, it is best to use it in the same way as the cryptocurrency community. It can be discussed openly through volunteers, go through a testnet program, and conduct risk assessments. Give it four-star.
3. Three-star standard
If it can only accept professional audits, it will be one level worse than without community discussion. Give it three-star. But civilians can use it, and the audit agency must bear joint liability.
4. Two-star standard
If it fails to pass the professional audit, give it two-star, indicating the risks and not for civilian use.
5. One-star standard
Artificial intelligence products produced under black-box operations will be given one-star, with a reminder of the risks, and not for civilian use.
This is the star rating ChatGPT deserves now.
Sam Altman is a sensible man
Sam Altman's suggestion in the US Congressional inquiry is excellent. He hopes the White House will have an artificial intelligence expert group to study this matter. How to do? Learn the practices of cryptocurrency.
The exploration of cryptocurrency in the production relationship tells us how an intelligent machine should usually serve humans.
Many innovations in cryptocurrency were ahead of our time. None of our past theoretical frameworks are adequate to explain current economic phenomena. Artificial intelligence not only changes productivity but also affects production relations. Humanity's cryptocurrency exploration is the best reference for the subsequent development of artificial intelligence.
Cryptocurrency and artificial intelligence reinforce each other
Finally, a friend wanted me to comment on artificial intelligence and cryptocurrency. I assess that they are all computer explorations of dealing with uncertainty. This valuable exploration has gone through the efforts of several generations of elites and about forty years.
In terms of technical difficulty, artificial intelligence is complex. Regarding product difficulty, cryptocurrency is more complicated and involves more areas. "If Satoshi Nakamoto had not developed Bitcoin, no one would have developed it within ten years." This sentence was said by Mr. Wei Dai, the natural person of Satoshi Nakamoto (chainless.hk). The technical level of similar artificial intelligence products is not much different, which is also the result of the natural progress of human technology.
Artificial intelligence is corporate product development involving limited projects and people; cryptocurrency is community development, with tens of thousands of development projects, hundreds of thousands of developers, more pitfalls, and more experience. So, it represents the thoughts of the new era of humanity.
The reason why artificial intelligence is so popular is that the market for uncertain problems is vast and gives people unlimited imagination. But for issues that require certainty, such as intelligent machine translation, it isn't easy to increase the accuracy from 97% to 98%, while cryptocurrency has 100% accuracy. Therefore, the requirements of cryptocurrency are higher and solved more cleverly. The path explored by cryptocurrency is enough to bring a demonstration effect to artificial intelligence.
Only by fully and accurately understanding cryptocurrencies and by completely opening up your mind to research cryptocurrencies can you keep up with the development of the times.
