Courts, Games and Education: How General AI Will Change Them

1. There is still no universally accepted definition of the term “metaverse”. Essentially, it can be used to refer to all virtual space. But it is more correct to speak not of metaverses, but of the term popularized by Snapdragon — extended reality, “augmented reality”. 

No one argues that we should dive headlong into this virtual space as if into The Matrix. The point is the expansion of our lives and the emergence of new, more convenient tools. The economy here is one of the engines of development. 

One of the major problems of DeFi is user experience: decentralized applications are difficult to use. A metaverse that allows creating its own economic models or borrowing them from outside could deliver the long-awaited breakthrough in this area. It would give people who currently know nothing about cryptocurrencies mechanisms for convenient, easy and secure interactions with them. 

2. Play-to-Earn will not go away. This mechanism will simply become one of many marketing tools enabling the transformation of a certain amount of money into an audience. Unfortunately, sustainable models that can fully close the loop and generate income from within are not yet visible. But the possibility of enabling users to monetize their hobbies continues to develop in various forms. The idea of distributing income among the audience is by no means new. We will watch the development of this model for a long time and wait for someone to manage to create it in a more or less closed format.

All the economy now existing in metaverses is largely laid out by their creators. However, the success of such a project still lies in the value it offers to a particular user. And no one will come to interact with a project simply because it has excellent economic prospects. A user will come for the concrete implementation, and it is from that that everything will depend. Many have seen the future GameFi solely in the search for ways to motivate people to spend time in a game not only for economic reasons. For now, that is difficult. 

3. A complex interface is only one of the DeFi problems that lie on the surface. When examining what prevents users from rushing into decentralized finance, you find that, surprisingly, the lack of market security is a problem. The same NFT are constantly being stolen, in staggering volumes.

If you look at analytics for 2022 alone, assets worth more than a hundred million dollars were stolen in the context of only registered phishing attacks. How many in other formats is unclear. We are facing a situation that needs to be regulated, but it will not be possible to resolve this task entirely with on-chain methods. 

Here comes a whole class of products, currently not very actively represented in DeFi, but enabling mathematically verifiable decision-making in a trustless environment trust. These are decentralized courts, which could become a serious driver for metaverses. Facing a huge number of disputes, we somehow resolve them. In the traditional world, we have courts and sets of laws in different countries to which we can appeal. In DeFi we have no such mechanisms, other than smart contracts themselves.

In principle, a fairly specific type of oracle would be enough to make decisions in a trustless environment. Moreover, beneath this lies a vast scientific base dating back to the mid-20th century, when game theory introduced the concept of Schelling points. It goes like this: If we cannot communicate with each other in any way, but we have a common objective, then we must find something common in our knowledge base. 

For example, if we need to meet in New York, most people would think of Central Park, because that point is associated with New York in their memory. Economist Thomas Schelling showed that using such an approach one can make decisions. The project is called Kleros, it is a decentralized court, the outcomes of which you encounter constantly. 

4. The main difference of the metaverse from other gaming projects is the absence of a script. Instead, a wide range of other possibilities opens up, allowing autonomous agents to create dynamic worlds. 

Most of us have seen how ChatGPT works — a chat where you can ask questions and receive answers. It keeps getting more sophisticated. On the basis of ChatGPT, new solutions have emerged that bring closer the realization of the idea of AGI, which could not only perform tasks given by humans but also set its own tasks. 

In March 2023, a project named Auto-GPT was introduced. It is a product that feeds ChatGPT an abstract task and asks it to solve the question. Then comes the most interesting part. Auto-GPT creates new autonomous agents, each of whom is given some tasks. Like the duplicates in the Strugatsky brothers’ novel ‘Monday Begins at Saturday’. You were handed a big task from above. You break it into many small tasks and for each you create a gnome who can do only one thing, but very well. And then this task assembles backward.

It’s easy to imagine a hypothetical NPC in a computer game with such capabilities and Internet access. The next step is a virtual world populated by such characters, with whom you can interact and who can interact with each other. each other.

5. Virtual technologies will reshape ordinary reality. When metaverses powered by autonomous agents appear, they will transform our lives. 

Already today we have the opportunity for every student to assemble their own personal teacher who will not only deliver knowledge in the format most understandable to the child, but will also shape the program to be as enjoyable, convenient, and gamified as needed. And who will verify all this? The LLM itself can perform this task, but the problem is that only when the child grows up will we understand what they were actually doing and how it affected their development.  

Who else is likely to deploy such in the near future? Politicians. Here we come to another problem. Let us go back a few years and remember how computers were just learning to play Go, which, as was thought, could not be algorithmized — too many possible moves. 

Naturally, artificial intelligence didn’t just beat a human; it did so cleanly. The interesting thing is not so much that it won, but how it did it. During the match, gray-haired men who had spent their lives playing Go watched the computer play and simply didn’t understand what was happening. 

And the biggest problem is that we may not even notice this until it is too late.