Original author: Kuleen, Head of DePIN at the Solana Foundation
Compiled by: Yuliya, PANews
Currently, the intersection of AI and crypto technology is entering an experimental phase reminiscent of a 'Cambrian explosion'. This article from the Solana Foundation elaborates on three key development directions of AI + crypto integration.
TLDR
1. Build the most vibrant intelligent agent-driven economy on Solana
Truth Terminal has already proven the feasibility of AI agents operating on-chain. Experiments in this field are continually breaking the boundaries of agent on-chain operations, which not only has immense potential but also offers a vast design space. This has already become one of the most groundbreaking and explosive directions in the fields of crypto and AI, and this is just the beginning.
2. Enhance LLMs' capabilities in Solana code development
Large language models have already shown excellent performance in code writing and will continue to improve in the future. With these capabilities, the efficiency of Solana developers is expected to increase by 2-10 times. Recently, establishing high-quality benchmarks to evaluate LLMs' understanding and writing capabilities regarding Solana code will help understand the potential impact of LLMs on the Solana ecosystem. High-quality model fine-tuning schemes will be validated in benchmark tests.
3. Support for an open and decentralized AI technology stack
"An open and decentralized AI technology stack" includes the following key elements:
Acquisition of training data
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 innovations and experiments
Provide alternatives for users who do not trust centralized AI
1. Build the most vibrant intelligent agent-driven economy
There has already been much discussion about Truth Terminal and $GOAT, so there's no need to elaborate here. But it is certain that when AI agents begin to participate in on-chain activities, a new world full of possibilities unfolds (notably, agents have not yet taken direct action on-chain).
While it is currently impossible to accurately predict the future development of agents' on-chain behavior, by observing the innovations that have already occurred on Solana, we can glimpse the vast potential of this design space:
AI projects like Truth Terminal are developing new types of digital communities through Meme coins like $GOAT
Platforms like Holoworld AI, vvaifu.fun, Top Hat AI, Alethea AI enable users to easily create and deploy intelligent agents and their associated tokens
AI fund managers trained on the personality traits of well-known crypto investors are emerging, with the rapid rise of ai16z on the daos.fun platform creating a new ecosystem of AI funds and agent supporters
Additionally, gaming platforms like Colony allow players to participate in games by guiding agent actions, often resulting in unexpected innovative gameplay.
Future development directions
In the future, intelligent agents can manage complex projects that require multi-party economic coordination. For example, in the research field, agents can be responsible for finding therapeutic compounds for specific diseases. Specifically:
Token fundraising through the Pump Science platform
Utilizing raised funds to pay for access to paid research materials, and for computing costs of compound simulations on decentralized computing networks like kuzco, Render Network, io.net
Recruiting humans for experimental verification work through bounty platforms like Gib.Work (e.g., running experiments to verify/build simulation results)
In addition to complex projects, agents can also execute simple tasks such as building personal websites and creating artworks (like zerebro), with infinite possibilities in their use cases.
Why is it more meaningful for agents to execute financial activities on-chain than to use traditional channels?
Agents can fully utilize both traditional financial channels and cryptocurrency systems. However, cryptocurrencies have unique advantages in certain areas:
Micro-payment applications—Solana has excelled in this area, with applications like Drip proving this point
Speed advantage—instant settlement capabilities help agents achieve maximum capital efficiency
Entering capital markets through DeFi—this may be the strongest rationale for agents participating in the crypto economy. When agents need to engage in financial activities beyond payments, the advantages of cryptocurrencies become even more apparent. Agents can seamlessly mint assets, trade, invest, lend, and use leverage. Particularly with Solana, which already boasts a plethora of top-tier DeFi infrastructure on its mainnet, it is particularly well-suited to support these capital market activities.
From the perspective of technological development laws, path dependence plays a critical role. Whether a product is optimal is not the most important factor; the key is who can first reach critical mass and become the default choice. As more and more agents earn income through cryptocurrencies, crypto connections are likely to become a core capability of agents.
The foundation hopes to see
The Solana Foundation hopes to see agents equipped with crypto wallets engaging in bold innovative experiments on-chain. The foundation does not overly limit specific directions here, as the possibilities are simply too broad—believing that the most interesting and valuable agent use cases are likely to be unforeseen at present.
However, the foundation is particularly focused on exploring the following directions:
1. Risk control mechanisms
Although current models perform excellently, they are still far from perfect
Agents cannot be given completely unconstrained freedom of action
2. Promote non-speculative use cases
Purchase tickets through xpticket
Optimize stablecoin portfolio yields
Order food on DoorDash
3. Development progress requirements
At least reach the prototype stage of the testnet
Ideally already running on the mainnet
2. Enhance LLMs' ability to write Solana code, empowering Solana developers
LLMs have already demonstrated strong capabilities and are making rapid progress. In the application field of LLMs, the area of code writing may see particularly steep progress curves as this is a task that can be objectively assessed. As mentioned below, 'programming has a unique advantage: the potential for superhuman data scaling through self-play. Models can write code and run it, or write code, create tests, and then check for self-consistency.'
Today, although LLMs are still not perfect in code writing, with noticeable shortcomings (e.g., poor bug discovery), AI-native code editors like Github Copilot and Cursor have fundamentally changed software development (even changing how companies recruit talent). Given the expected rapid rate of progress, these models are likely to revolutionize software development. The foundation hopes to leverage this advancement to elevate Solana developers' work efficiency by an order of magnitude.
However, several challenges currently hinder LLMs from achieving excellence in understanding Solana:
Lack of high-quality original training data
Insufficient number of verified builds
Lack of high-information-value interactions on platforms like Stack Overflow
Historically, the rapid development of Solana's infrastructure means that even code written six months ago may not fully meet today's needs
Lack of methods to assess models' understanding of Solana
The foundation hopes to see
Helping to obtain better Solana data on the internet
More teams releasing verified builds
More people in the ecosystem are actively asking good questions and providing high-quality answers on Stack Exchange
Create high-quality benchmark tests to evaluate LLMs' understanding of Solana (an RFP will be released soon)
Create LLM fine-tuning models that perform well in the above benchmark tests; more importantly, accelerate the work efficiency of Solana developers. Once high-quality benchmark tests are available, the foundation may reward the first model to reach the benchmark scoring threshold.
The ultimate significant achievement will be: a completely AI-created, high-quality, differentiated Solana validation node client.
3. Support for an open and decentralized AI technology stack
In the field of AI, the long-term power balance between open-source and closed-source models remains unclear. There are indeed some arguments that support closed-source entities continuing to maintain a technological edge and capture the primary value of foundational models. The simplest current expectation is to maintain the status quo—tech giants like OpenAI and Anthropic pushing frontier developments, while open-source models quickly follow and gain unique advantages through fine-tuning in specific application scenarios.
The foundation is committed to closely integrating Solana 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 iterations
The rapid improvements and fine-tuning of open-source models like Llama by the open-source community demonstrate how the community effectively supplements the work of large AI companies, pushing the boundaries of AI capabilities (even Google researchers pointed out last year, 'Regarding open source, we have no moat, and OpenAI does not either'). The foundation believes that a thriving open-source AI technology stack is crucial for accelerating progress in the field.
2. Provide alternatives for users who do not trust centralized AI
AI may be the most powerful tool in the arsenal of dictatorships or authoritarian regimes. State-sanctioned models provide officially recognized 'truths', serving as an important control vector. Highly authoritarian regimes may possess superior models because they are willing to ignore citizen privacy to train AI. It is an inevitable trend for AI to be used for control, and the foundation hopes to prepare for this by fully supporting an open-source AI technology stack.
Several projects in the Solana ecosystem are already supporting an open AI technology stack:
Data collection—Grass and Synesis One are advancing data collection
Decentralized computing power—kuzco, Render Network, io.net, Bless Network, Nosana, etc.
Decentralized training frameworks—Nous Research, Prime Intellect
The foundation looks forward to seeing
Hope to build more products at all levels of the open-source AI technology stack:
Decentralized data collection: e.g., Grass, Datahive, Synesis One
On-chain identity: protocols supporting wallet verification of human identity, protocols verifying AI API responses, enabling users to confirm they are interacting with LLMs
Decentralized training: projects similar to EXO Labs, Nous Research, and Prime Intellect
IP infrastructure: allowing AI to license (and pay for) the content it uses
