The Three Major Development Directions of AI and Encryption Technology Integration

Currently, the intersection of AI and encryption technology is entering a thriving experimental phase. This article elaborates on three key development directions of the AI + encryption integration.

Summary

1. Build the most vibrant economy driven by intelligent agents

Existing projects have proven the feasibility of AI agents operating on the blockchain. Experiments in this field continue to push the boundaries of agent operations on the blockchain, with enormous potential and a vast design space. This has now become one of the most groundbreaking and explosive directions in the fields of encryption and AI, and this is just the beginning.

2. Enhance the capabilities of large language models in code development.

Large language models have shown excellent performance in code writing and are expected to further improve in the future. With these capabilities, developers' efficiency is expected to increase by 2-10 times. Recently, establishing high-quality benchmarks to evaluate large language models' understanding and writing of code will help understand their potential impact on the ecosystem. High-quality model fine-tuning solutions will be validated in benchmark tests.

3. Support open and decentralized AI technology stack

The "open and decentralized AI technology stack" includes the following key elements:

• Training data acquisition
• Training and inference computing power
• Model weight sharing
• Model output verification capability

The importance of this open AI technology stack is reflected in:

• Accelerate model development innovation and experimentation
• Provide alternatives for users who do not trust centralized AI.

1. Build the Most Dynamic Smart Agent-Driven Economy

When AI agents begin to participate in on-chain activities, a new world full of possibilities has already unfolded (it is worth noting that currently agents have not yet taken direct action on-chain).

Although it is currently impossible to accurately predict the future development of agent behavior on the chain, we can glimpse the broad prospects of this design space by observing the innovations that have already occurred:

• Some AI projects are developing new digital communities through Meme coins.
• Multiple platforms allow users to easily create and deploy smart agents and their associated tokens.
• AI fund managers trained on the personality traits of well-known encryption investors are emerging, creating a new ecosystem of AI funds and agent supporters.
• Some gaming platforms allow players to participate in the game by guiding proxy actions, often resulting in unexpected innovative gameplay.

future development direction

In the future, intelligent agents can manage complex projects that require multi-party economic coordination. For example, in the field of scientific research, agents can be responsible for finding therapeutic compounds for specific diseases. Specifically:

• Raise tokens through relevant platforms
• Use the raised funds to pay for access to paid research materials and the computation costs for compound simulations on a decentralized computing network.
• Recruit individuals through a bounty platform to perform experimental verification tasks (e.g., running experiments to verify/establish simulation results)

In addition to complex projects, agents can also perform simple tasks such as creating personal websites and producing artwork, with limitless possibilities for application scenarios.

Why does it make more sense for agents to conduct financial activities on the blockchain rather than using traditional channels?

Agents can fully utilize both traditional financial channels and encryption systems at the same time. However, encryption has unique advantages in certain areas:

• Micro Payment Application
• Speed Advantage: Instant settlement feature helps agents achieve maximum capital efficiency
• Entering the capital market through decentralized finance: This may be the strongest reason for agents to participate in the encryption economy. The advantages of cryptocurrency become even more apparent when agents need to engage in financial activities beyond payments. Agents can seamlessly mint assets, trade, invest, engage in lending operations, use leverage, etc.

From the perspective of technological development patterns, path dependence plays a key role. Whether a product is optimal is not the most important; the key lies in who can first reach a critical scale and become the default choice. As more and more agents earn profits through encryption, encrypted connections are likely to become a core capability of agents.

Future Outlook

Hope to see agents equipped with encryption wallets conducting bold innovative experiments on the chain. Specifically, the following directions are worth paying attention to:

1. Risk Control Mechanism

   • Although the current model performs excellently, it is still far from perfect.
   • Cannot grant agents complete unrestricted freedom of action

2. Promote non-speculative use cases

   • Purchase tickets through relevant platforms
   • Optimize stablecoin investment portfolio returns
   • Order food on the takeaway platform

3. Development Progress Requirements

   • Must reach at least the prototype stage of the testnet
   • It is best to be running on the mainnet already.

2. Enhance the ability of large language models to write code, empowering developers

Large language models have demonstrated powerful capabilities and are making rapid progress. In their application fields, the area of coding may see particularly steep progress, as it is a task that can be objectively assessed. As someone pointed out, "Programming has a particularly unique advantage: the potential for superhuman data augmentation through 'self-play'. The model can write code and run it, or write code, write tests, and then check for self-consistency."

Today, although large language models are still not perfect in coding, with obvious shortcomings (for example, performing poorly in finding bugs), tools like AI-native code editors have fundamentally changed software development (even altering the way companies recruit talent). Considering the expected rapid rate of progress, these models are likely to completely transform software development. We hope to leverage this advancement to increase developers' work efficiency by an order of magnitude.

However, there are currently several challenges that hinder large language models from achieving excellence in understanding certain specific technologies:

• Lack of high-quality raw training data
• Insufficient number of verified builds
• There is a lack of high information value interactions on platforms like Stack Overflow.
• Historically, infrastructure has developed rapidly, meaning that code written even 6 months ago may not fully meet today's needs.
• Lack of evaluation models for understanding the level of specific technology.

Future Outlook

• Help obtain better relevant data on the Internet
• More teams release verified builds
• More people in the ecosystem are actively asking good questions and providing high-quality answers on Stack Exchange.
• Create high-quality benchmarks to evaluate large language models' understanding of specific technologies.
• Create a large language model fine-tuning model that performs well in the benchmark tests mentioned above, and more importantly, accelerate the work efficiency of developers. Once high-quality benchmark tests are available, it may provide rewards for the first model to reach the benchmark test threshold score.

The ultimate significant achievement will be: a brand new, high-quality, differentiated validator node client completely created by AI.

3. Support for Open and Decentralized AI Technology Stack

In the field of AI, the long-term balance of power between open-source and closed-source models remains unclear. There are indeed arguments supporting the idea that closed-source entities will continue to maintain a technological edge and capture the primary value of foundational models. The simplest expectation at present is to maintain the status quo—large tech companies driving cutting-edge developments while open-source models quickly follow and gain unique advantages through fine-tuning in specific application scenarios.

We are committed to closely integrating the ecosystem with the open-source AI ecosystem. Specifically, this means supporting access to the following elements:

• Training data
• Training and inference computing power
• Model Weights
• Model output verification capability

The importance of this strategy is reflected in:

1. Open source models accelerate innovation iteration

The rapid improvements and fine-tuning of open source models by the open source community demonstrate how the community effectively complements the work of large AI companies and pushes the boundaries of AI capabilities (some researchers even point out that "when it comes to open source, we have no moat, and other companies don’t either"). We believe that a thriving open source AI technology stack is crucial for accelerating progress in this field.

2. Provide options for users who do not trust centralized AI.

AI may be the most powerful tool in the arsenal of dictatorial or authoritarian regimes. State-sanctioned models provide an officially recognized "truth" and serve as an important control vehicle. Highly authoritarian regimes may possess superior models because they are willing to overlook citizen privacy to train AI. The use of AI for control is an inevitable trend, and we hope to prepare in advance and fully support open-source AI technology stacks.

Multiple projects in the ecosystem are already supporting the open AI technology stack:

• Data Collection: Some projects are advancing data collection.
• Decentralized computing power: Multiple networks and platforms provide support.
• Decentralized training framework: Some research groups and projects are exploring this area.

Future Outlook

I hope to build more products at all levels of the open-source AI technology stack:

• Decentralized data collection: Supports various data collection projects
• On-chain identity: Protocols that support wallet verification of human identity and protocols that verify AI API responses, enabling users to confirm that they are interacting with a large language model.
• Decentralized training: supports decentralized training solutions similar to existing projects.
• IP infrastructure: enabling AI to license (and pay for) the content it uses.