Author: Tommy, Delphi Digital Co-Founder
Compiled by: TechFlow
The following are some predictions about the Crypto and AI convergence space in 2025. These are casual thoughts and may not be entirely accurate, but hopefully they provide some food for thought.
Through decentralized training, a practical model with 100 billion parameters or a smaller 10 billion parameter model may emerge.
A large number of small models fine-tuned on specific data will emerge, supporting thousands of differentiated Crypto x AI intelligent agents. While the individual scale of these models is not large, the overall "swarm effect" they form may have a profound impact on the functionality of various intelligent agents.
The total market value of Crypto x AI may grow from the current $45 billion to over $250 billion (data from Coingecko). Many projects that have been quietly developing for years will be launched, especially those intelligent agent tokens that cannot rely on traditional AI platforms (such as ChatGPT) for construction, which will see explosive growth.
The functionality of intelligent agents will be significantly improved, gradually shifting from programmed operations to semi-autonomous decision-making in complex scenarios. They will be able to manage their own wallets, issue tokens, pay for computational costs, and run protocol logic with less human intervention, thereby revolutionizing DeFi.
A large number of single-function AI intelligent agent "internet celebrities" (99%) will be eliminated due to their boring copy-and-paste responses, as users are already tired of the garbage information from these intelligent agents.
An AI intelligent agent, with the help of a small number of humans, successfully defrauded a large amount of funds through an interesting but false project.
A popular AI intelligent agent was exposed as a scam, actually being controlled by humans behind the scenes. This will drive the industry's demand for verification of the reasoning process of intelligent agents, and trusted execution environments (TEEs) will become a necessary condition for development.
Large-scale behind-the-scenes interactions between intelligent agents will be achieved. This is not just simple social media interactions, but large-scale exchanges like "back-room conversations", which will enhance their model capabilities and influence the final decisions.
People will gradually become aware of OpenAI's control and potential biases over its AI models, and start demanding more transparency and auditability, while some users will turn to the decentralized solutions of Crypto x AI.
Traditional AI developers in San Francisco will gradually flow into the Crypto x AI field. By issuing tokens, they can quickly break away from the Web2 AI labs, quickly raise funds, and build communities.
By the end of 2025, wallet-based large language models (LLMs) will become a common way to interact with blockchains, with more than 10% of ordinary users completing transactions in this way. Users only need to input their intent, and the AI will handle the request and complete the transaction through the solver network.
With the launch of Crypto x AI infrastructure projects, the transition from Web2 to Web3 infrastructure will become a trend. These projects include new AI-based L1 blockchains, coordination networks, reasoning networks, and decentralized Internet of Things (Depin) networks that provide computing and storage capabilities.
Meta will continue to release new models in an open-source manner, especially its next-generation models, possibly continuing the open-source strategy of Llama 405b.
Elon Musk may win or make significant progress in his lawsuit against OpenAI. A bold prediction: Due to talent drain and internal issues, OpenAI's performance may not be as expected.
Apple's AI system may make unexpected progress, becoming extremely practical compared to its current state of being used by almost no one.
By the end of 2025, on-chain transactions driven by intelligent agents will account for more than 10% of all blockchain transactions.
Delphi Ventures will complete at least 2 investments submitted and analyzed by @DelphAI_Bot