Best Uses for Cryptocurrencies: Is Combining with AI the Best Way to Go?

Web 3.0 isn’t much talked about these days, but as AI bots become more common, the idea of Web3 will regain its relevance.

According to GPT-4, Web 3.0 is “the next frontier of Internet technology, characterized by decentralized, user-centric applications that value data privacy and facilitate seamless, connected experiences.” Today, Web 3.0 is an “innovation” that was hyped for years—until crypto prices crashed and caused many crypto projects to close down—as one of the most discredited ideas of the era. As with encryption, Web 3.0 advocates have never been able to produce a real use case for their technology. However, there is a reason why I asked GPT-4 for a definition. I fully expect the ideas behind Web 3.0 to make a comeback - as a legal and institutional framework for AI bots, it’s worth thinking about how to do it. Say you run a charity and want to create and distribute an AI robot that will teach math to underprivileged schoolchildren. That’s fine, but the bot hits some snags. In some jurisdictions, it may need to pay licensing and registration fees, it may need to purchase add-ons to recent innovations in teaching, it may wish to upgrade its translation capabilities if it operates abroad, and for a variety of reasons, it may Still need money. **All of these transactions would be very easy if AI were allowed to have bank accounts. But that’s unlikely to happen anytime soon. How many banks are ready to handle this? Imagine the public outcry if the bank fails and the government has to bail out some robo-accounts. As a result, bots may remain “unbanked” — which would drive them to use cryptocurrencies as their core medium of exchange. ** Critics often point out that the U.S. dollar is more efficient as a form of exchange than cryptocurrencies. But if AI bots can’t use dollars, then they will have to use cryptocurrencies. Yes, some owners may give bots access to their checking accounts, while others may want to determine each bot payout through the dollar-based banking system. But I suspect that most people would rather just let the bots run on their own without all this risk and hassle - again, which brings us back to cryptocurrencies. “Agent” bots are known to be generally more efficient than “tool” bots, and they will require capital. Also, probably for liability reasons many of these bots will not be owned at all (do you want to be sued in a foreign country for something your bot said or did?). **This will be another force driving bots to operate in encrypted relationships. ** At first, crypto assets will be existing assets such as Bitcoin and Ethereum. But over time, bots may invent their own tokens for convenience. It is conceivable that these bots will be good at number crunching, so they may use multiple crypto assets for decentralization reasons. A bot-based crypto-economy could evolve to be far more complex than a human-based crypto-economy. That parallel economy may only be a small fraction of GDP. But it will help maintain crypto asset prices. Also, if humans are dealing with these bots (maybe some people want to donate money to your math-teaching bot), they’ll need some cryptocurrency as well. Thus, a bot-based crypto-economy will permeate the general economy. The plot now gets more complicated. Once there is a cryptocurrency-based robot economy, that economy needs property rights and a legal system for the same reasons as a human economy. Suppose you teach a robot to buy a translation improvement service from another robot - but it is defective, you teach the robot to “recover its compensation”. ** To enhance trust and reduce transaction costs, many bots will agree in advance to an arbitration service, usually provided by another bot. (These bots can’t just show up in the Fairfax County courthouse and demand justice.) Final adjudication judgments will be made through inscriptions on the blockchain, using smart contracts if necessary, and defining new property rights. ** All of these, simply put, are exemplary illustrations of Web 3.0.

Some bots may wish to advertise themselves, or establish core identities, and some will do this by purchasing token goods and establishing ownership of those goods on the blockchain. This is also a Web 3.0 idea. At the height of Web 3.0 mania, many critics thought it absurd for humans to spend millions of dollars on NFTs. Why spend so much money just to get a blockchain-based receipt? Proving that you spent that much money, the buyer doesn’t even necessarily have any accompanying intellectual property. You can laugh all you want, but robots will do something similar. Back to your math-teaching robot: Suppose it competes with other math-teaching robots. It might want to make a promotional website – “Here’s my portfolio, come here and learn calculus!” Maybe it’s depicting Isaac Newton and Leibniz with some stunning Midjourney images. All of these steps are likely to be accomplished using Web 3.0 tools or their derivatives based on robotics innovations. **Robots’ growth and success will depend on electricity prices, so bots may actively trade in the electricity futures market to hedge their positions. They may have to trade with other robots and humans. ** Remember DAO (Decentralized Autonomous Organization)? I have yet to see a DAO run by humans succeed on a large scale, perhaps because humans need more authority, or because DAOs are just another hidden human authority in disguise (e.g. one person controls 51 % of votes). The bots have learned about the DAO and the points where it failed, and they might try again. Meanwhile, bots will train themselves to learn how to make their DAOs work, and bot “companies” may end up being more democratic than their human counterparts. ** Some bots may find it efficient to enjoy the limited liability version. Dealing with such a bot, you or your bot will stipulate that you cannot sue the bot for more than a specified amount. In order to hire a robot for a more dangerous task, you and the robot might agree to a stricter standard of liability. These “corporate laws,” if you want to call them that, will feature a lot of rapid innovation and diversity of approach. Human institutions look sluggish by comparison. Over time, perhaps we can learn from these robotic experiments. ** I believe robots can think of many other Web 3.0 applications - especially if we train them. Here’s a broader heuristic: Don’t dismiss a technological or institutional advance because it’s hard to see how it might be useful. Quantum mechanics existed for decades and seemed like a fascinating spectacle before becoming one of the fundamental principles of modern computing. Critics attack games as corrupting young people, but GPUs now used in advanced games help build better artificial intelligence, and AI computing company Nvidia approaches a trillion-dollar company. Of course, when all of this is happening, we’ll probably be calling it Web 4.0. But Web 3.0 will finally get the respect it deserves.