Apart from the US election this month, the industry's hottest topic is the narrative wave of "AI + Crypto".
In his book Sapiens: A Brief History of Humankind, Yuval Noah Harari puts forward a relatively novel view: he believes that the formation of history is more driven by various "coincidences" rather than the "inevitability" of history rationalized by later generations. This concept runs through the whole book, and Harari examines the unpredictable and even almost random events that affect human history.
The recent unexpected rise of a meme coin called "GOAT" seems to confirm Harari's point of view. GOAT became famous because of a semi-autonomous AI agent on the X (formerly Twitter) social platform, a robot with the social account @truth_terminal (literally "truth terminal"), which kept tweeting to promote GOAT until the price of the meme coin skyrocketed, and truth_terminal also entered the public eye. So far, AI and Crypto have successfully created sparks.
Readers are naturally confused: What is the intersection between AI and Crypto?
Answering this seemingly simple question is actually not easy. This article will unravel this issue from three aspects:
What is AI? How did AI emerge? What stage of development is AI in today?
The essence of Crypto and its infrastructure: blockchain technology
AI + Crypto: The Emergence of Decentralized AI (DeAI)
AI in “Back to the Future”
Most of the AI concepts that are well known to the public today originate from ChatGPT launched by OpenAI in November 2022. But discussions surrounding AI are not new in academia.
In the classic sci-fi movie "2001: A Space Odyssey" in the 1960s, AI was brought to the big screen as the protagonist: HAL 9000, an AI that was forced to shut down at the end because it posed a threat to the lives of human astronauts. Even back then, AI was not new - the academic community generally viewed the spirit test as the earliest model of artificial intelligence.
Due to the slow development of computer technology, AI experienced a period of stagnation and low tide in the next few decades, and public attention also waned until OpenAI launched ChatGPT in 2022, which gave artificial intelligence a new lease of life. AI has been able to reach this point today thanks to the rapid development of computer technology, especially breakthroughs in machine learning algorithms, deep neural networks, and natural language processing. These advances, coupled with unprecedented increases in data processing speeds and the availability of large data sets, have enabled AI capabilities to reach a level of complexity that was once thought to be unattainable. The geometric upgrade of software and hardware capabilities has also contributed to this, especially the breakthroughs in the fields of GPUs and TPUs, which have brought long-dormant AI back into the spotlight.
But as people use ChatGPT for tasks ranging from research and brainstorming to copy editing, new concerns arise: What happens if OpenAI controls a large amount of data and knowledge? If a private company can own and control such a large amount of data and knowledge, is humanity in danger?
There have been numerous concerns and accusations against OpenAI. Even Elon Musk, a founding member of OpenAI, has publicly criticized OpenAI many times, saying that it is the opposite of "open". From data collection to algorithms and data usage, it is all opaque black box operations. This is also the biggest challenge facing centralized AI, which is called the "black box problem" in the industry.
This raises a fundamental question: How to make AI truly open source? And this is exactly the core reason why Crypto, and more precisely Blockchain technology, enters the narrative.
The rise of decentralized AI
The decentralized nature of blockchain technology makes open source and transparent AI possible.
It is generally believed that with blockchain technology, AI can be truly open source and transparent by leveraging several key technologies, as follows:
Decentralized storage ensures data integrity;
Smart contracts provide transparent model access and audit trails;
On-chain federated learning allows for collaborative model training without sharing raw data;
Distributed computing networks decentralize AI processing;
Cross-chain bridges and oracles introduce real-time, reliable off-chain data;
Privacy protection methods such as zero-knowledge proofs ensure the security of sensitive information.
The term "decentralized AI" means that everyone can contribute to every stage of AI from development to deployment, serving as an alternative and counter-measure to closed and centralized platforms such as OpenAI.
However, there is still a lack of industry consensus on the standard definition of decentralized AI. The only consensus is that blockchain can support the vision of DeAI, and cryptocurrencies play a key role in token economics (how token economics and AI projects work together will be further discussed in future articles).
Ethereal narrative? Hardcore innovation?
Since both AI and Crypto industries are suspected of being highly bubbled, it is only natural to raise this question.
Based on my observations in the industry, I believe 2025 will be the year when decentralized AI shines.
Unlike many over-hyped blockchain projects or protocols, decentralized AI is here to stay. This innovation combined with immediacy makes it attractive to both blockchain veterans and firm believers in the world of AI.
More and more blockchain and AI talents are beginning to realize the potential of decentralized AI to reshape the industry. The interest and momentum are growing, which will inevitably turn DeAI from a niche to mainstream and rigid demand, paving the way for hard-core technologies and innovations that may be born in the next few decades. But of course, the road will not be smooth sailing.
Author: Max Li, Founder and CEO of OORT and Professor of Columbia University
Originally published in Forbes:
