The New Direction Web3 Should Take in the AI Era

Original author: Tyler Cowen, Bloomberg columnist Original compilation: Leo, BlockBeats

Editor’s note: With the explosion of GPT, AI has officially entered the public’s field of vision, and many venture capital and projects in the encryption field are gradually approaching/integrating AI technology, and even many institutions have abandoned encryption and turned to AI. Of course, the best way is to innovate at the intersection of AI and Crypto, and new narratives including AIDPoS, AI concept currency, and zkML have emerged. As mentioned in the article, a "new technology" that has been hyped for many years is mostly due to "human subjective thinking". In this respect, AI robots will surpass humans both in technology and in operation. It will indeed liberate humans in some executive layer operations. For the traditional economy, AI will make humans feel fear and worry, and we are very afraid of handing over the assets of the bank account to AI. But in today's encryption field full of "unstable factors", AI is expected to become the savior of Crypto. AI represents digital abundance, and encryption represents digital scarcity. There will be conflicts between them (encryption is people-oriented, AI is people-oriented, etc.), but more is that they can exert powerful energy when combined. Bloomberg columnist Tyler Cowen wrote the article "The best use case for future encryption comes from AI", compiled by BlockBeats as follows:

Web3 under GPT

Remember Web3? Let ChatGPT answer: According to GPT-4, Web3 is "the next frontier of Internet technology, characterized by decentralized, user-centric applications that prioritize data privacy and facilitate seamless, connected experiences .”

Today, Web3 has become one of the least credible ideas of this era, a "technological innovation" that has been hyped for many years-until the price of encryption plummeted, and many encryption projects collapsed. Like Crypto, Web3's Advocates have never been able to make a real-world use case for the technology they believe in.

Of course, I asked GPT-4 for the Web3 definition for a reason. I'm particularly looking forward to a resurgence of the ideas behind Web3 - a use case that can be used for legal and institutional frameworks for AI bots, and it's worth rethinking Web3 at this point.

Use Cases of AI Bots Combining with Web3

Say you run a charity and want to create and distribute an AI robot that will teach math to underprivileged children, but robots will always encounter some obstacles. For example, in some jurisdictions it may need to pay licensing and registration fees, it may need to purchase add-ons for recent teaching courses, and if it operates abroad, it may need to upgrade its translation capabilities. Taken together, these processes require capital.

Such transactions would be much easier if AI were allowed to have their own bank accounts. But that doesn't seem likely to happen anytime soon, how many banks will be able to accept robo-accounts? Imagine the public protests if a bank fails and the government has to bail out some robo-accounts. ** As a result, bots will likely not be able to access "bank accounts" - which would drive them to use cryptocurrencies as their core medium of exchange. **

Critics often point out that the U.S. dollar is more effective than crypto as a form of currency exchange. But if AI bots can't use dollars, they have no choice but to use encryption. Yes, some owners may let the bot access their accounts, while others may want to have every bot spend confirmed through the dollar-based banking system; but most would rather let the bot run itself without many risks and trouble. **Reality brings us back to encryption. **

Why “agent” bots are often more efficient than “tool” bots, agent bots need funding commensurate with a reasonable degree of robot autonomy. Also, maybe for some reason (do you want to be sued in a country for what your bot said or did), most of these bots won't be owned by humans, **this is also driving AI bots Another force at work in the crypto industry. **

In the beginning, cryptoassets were what we know today as Bitcoin and Ethereum. But over time, bots may issue toekn based on their cognition (for the bot's convenience). It is speculated that these bots are good at crunching numbers, are more diverse, they may use multiple crypto assets, ** a robot-based crypto-economy may be more complex than a human-based crypto-economy. **

This type of parallel economy may only account for a small percentage of GDP, but it will help maintain the price of crypto assets. Also, if humans are dealing with these bots (maybe some people want to donate to your math teaching bot?), they also need exposure to crypto assets. Therefore, the **robot-based crypto-economy will permeate the traditional economy. **

The plot gets complicated. Once a crypto-based robot economy emerges, such an economy will need property rights and a legal system, just like a human economy. Let’s say your teaching bot buys a translation improvement service from another bot, but there is a problem with the service, your teaching bot will “want a refund”, to enhance trust and reduce transaction costs, many bots will agree to arbitration in advance service, usually provided by a third robot. (These robots cannot be attached to any geography and regionalism, they must be fair and just). The final verdict will be inscribed on the blockchain, using smart contracts if required. New property rights will be defined.

AI robot goes deep into Web3

Some bots may wish to advertise themselves, or establish core identities, some bots will purchase token goods and establish ownership of those goods on-chain, some also from Web3 use cases. At the peak of Web3 fever, many critics believed that it was unreasonable for humans to be willing to pay millions of dollars in assets for NFT. Why would you spend so much money just to get a blockchain-based "receipt" to prove that you have spent? This money? Buyers don't even necessarily own any of the accompanying intellectual property.

Ridiculous as it may seem, robots of the future will do something similar to, say, your math-teaching robot: Suppose it's competing with other math robots. It will think of a promotional website-"My portfolio, come to me to learn calculus!", maybe it will use some images of Newton and Leibniz based on Midjourney to prove it. All of these steps are likely to be accomplished with Web3 Tools or its robotics innovation fork.

The development and success of bots will depend on the price of electricity, so bots may actively trade in the electricity futures market to hedge their positions, and they may have to trade with other bots and humans.

Remember DAOs? That DAO? So far I haven't seen a DAO run by humans achieve widespread success, perhaps because humans need more power, or because DAOs are just another covert form of human power (e.g. one person controlling 51% of the vote ). Bots already understand DAOs and the problems of not functioning perfectly, and they may redefine DAOs. Meanwhile, bots will train themselves to learn how to run their DAOs**, and bot "organizations" may become more democratic than human ones. **

Some bots may find a "limited liability" regime more efficient, by dealing with such a bot, you or your bot will be restricted from suing the bot for more than a specified amount. For hiring robots to perform more dangerous tasks, you and the robot may agree on a stricter liability standard, based on so-called "corporate law", which will bring a lot of innovation and diversity of methods. Human institutions are sluggish in comparison, and perhaps we can learn from these robotic experiments over time.

And now, **I am sure that AI bots can generate more Web3 applications under our training. **

Conclusion

Lesson: Don't dismiss technological or institutional advances just because it's hard to see what benefits they might bring. Quantum mechanics existed for decades, seemingly as a concept driven by our curiosity before it became a fundamental principle of modern computing. Critics say games corrupt young people, but GPUs now used in advanced games help build better AI, and even Nvidia already has nearly a trillion dollars in assets.

Of course, when this happens, we might call it "Web4". But before that, Web3 will finally get its due rewards and prove its worth.