Artificial Intelligence and Cryptocurrencies: A Revolutionary Alliance That Promises New Uses and Innovations.
The time has come to take stock of the situation and look to the future.
It's an understatement to say that the field of artificial intelligence took a radical turn in November 2022 with the advent of ChatGPT. This artificial intelligence, based on a type of model known as LLM (Large Language Model), revolutionized the field. Since then, LLMs and other generative AIs have gone from strength to strength. Unsurprisingly, the links with the field of cryptocurrencies were not long in coming. It therefore seems essential to me to explore with you the origins of these AIs, the progress made in recent years, and the developments that are already emerging in this field and that of cryptos.
Quantum AI: A Technological Revolution at the Crossroads Between AI and Quantum Computing.
Quantum AI is at the heart of this newsletter, which I've been developing for a few months now, as it is the fruit of the encounter between quantum computing and artificial intelligence. Quantum AI promises to overcome the limitations of classical AI by harnessing the unparalleled power provided by qubits.
Generative AI: 70 years of history and development
The beginnings of artificial intelligence are closely linked to those of computer science. This goes back to the 1950s, with the father of modern computing: Alan Turing. He is best known for breaking the code of German transmissions. According to historians, his work alone helped shorten the Second World War by two years. But the work of the English mathematician and cryptographer didn't stop there.
In 1950, shortly after his death, he combined science and philosophy to imagine the beginnings of artificial intelligence. Turing was convinced that machines would one day be able to reason, and that artificial intelligence would become a reality. As a reminder, the first computers didn't see the light of day until 1946, and the advent of giants such as IBM didn't begin until the early 50s. For his part, Alan Turing proposed a revolutionary experiment: the Turing test. This was a series of tests designed to measure a machine's intelligence.
“I bet that within fifty years, there will be no way of distinguishing between the answers given by a man or a computer, on any subject whatsoever.”
Alan Turing was right. Indeed, in a scientific publication dated March 31, 2025, several researchers revealed that given the right starting instructions, GPT-4.5 was judged to be human 73% of the time, passing the Turing test with flying colors.
"When prompted to adopt a humanlike persona, GPT-4.5 was judged to be the human 73% of the time: significantly more often than interrogators selected the real human participant. The results constitute the first empirical evidence that any artificial system passes a standard three-party Turing test."
But there was a long way to go before we got to this point. Not surprisingly, the first generative models were relatively simple by today's standards. These included Hidden Markov Models, Restricted Boltzmann Machines, and Variational Autoencoders.
However, it wasn't until 2014 that a major event shook up the field. Ian Goodfellow and his colleagues introduced Generative Adversarial Networks (GANs). These are based on two distinct neural networks. A revolutionary approach in which the duality between these two networks enables them to evolve and adapt.
A Trapped-Ion Quantum Processor Generates the First "Truly Random" Number.
You've probably all waited for the expression “it's random” several times in your life. The phrase is frequently used to describe unpredictable situations, but true randomness, in physics as in mathematics, has long remained an elusive concept.
ChatGPT, Midjourney, and their rivals: 2022, the pivotal year for artificial intelligence
The year 2022 will have been pivotal in the evolution of artificial intelligence. Whether through the emergence of Chatbot like ChatGPT or generative AI like Midjourney.
Midjourney: the visual revolution
It all began in July 2022 with the emergence of Midjourney, the first generative AI to open its doors to the general public. Whether realistic images, paintings, cartoons, or landscapes, the possibilities seem endless. Creations made with Midjourney quickly invaded the Internet. In just a few clicks, users can generate images via the Discord integration of the generative model. It's a world first, and just the beginning.
ChatGPT: the democratization of LLM
Shortly afterwards, OpenAI took the industry by storm with the unveiling of ChatGPT. A conversational artificial intelligence. Everyone is completely flabbergasted by its results. Of course, it tends to hallucinate and can completely miss the point of any question put to it. However, this is the first time that such a conversational level has been achieved by an AI.
In practice, ChatGPT is based on a family of models known as LLMs. As a reminder, an LLM (Large Language Model) is an artificial intelligence model trained on immense quantities of text, capable of understanding, generating, and manipulating natural language coherently and contextually. The craze was so great that ChatGPT was the fastest IT project in history to reach the million-user mark. In just 5 days.
Artificial intelligence and the emergence of the open-source world
The open source sector has rapidly caught up with the OpenAI and Midjourney giants. We have seen the emergence of alternatives such as Stable Diffusion, Llama, and Mistral. Open source alternatives that open the door to new possibilities, such as hosting and local operation of these AIs. And total control over personal data management, a central component and recurring question among users of solutions such as ChatGPT.
With Zuchongzhi 3.0, China Unveils a Quantum Computer 10¹⁵ Times Faster Than Existing Supercomputers.
The technological rivalry between the USA and China is intensifying, particularly in quantum computing, which is now considered a major technological challenge. Recently, Microsoft unveiled Majorana 1, the first quantum chip based on a topological architecture
New areas of AI evolution
If recent years are anything to go by, AI developments are not going to stop there. Several areas are currently being actively explored. How can we not mention the subject of Autonomous Agents, these AIs that no longer simply analyze and respond, but are capable of acting autonomously? A field that is already developing in cryptos, with agents, for example, analyzing crypto news just-in-time and alerting the community in real time, or even acting autonomously among themselves, as in the first transaction carried out between two AIs last December.
Edge AI is another fast-growing field. It involves running AI algorithms directly on devices close to the data source, such as sensors, cameras, and connected objects. So, rather than sending all the data to a central server or the cloud, it is processed locally, in real time, with very low latency, while enhancing confidentiality since sensitive data does not leave the device.
In March 2025, Inception Labs unveiled a new approach that could gain momentum in the future: diffusion language models (dLLMs). These differ from current models in that they are generated in two stages:
Global draft: Rapid generation of a rough text structure.
Iterative refinement: Progressive improvement of content over several passes.
According to initial results, this would increase generation speed by a factor of 10 while improving consistency on long texts.
Artificial Intelligence and Crypto: A booming field
As we briefly touched on, AI is already being integrated into the cryptocurrency field. As a result, we've seen the emergence of numerous autonomous agents with a variety of functions. Some are simply connected to large data sources and analyze this data in real time, sharing their opinions via social networks. One example is Aixbt, an AI-driven crypto market intelligence platform developed by Virtuals. It analyzes trends, discussions, and market sentiment in real time from multiple sources such as social networks (notably X/Twitter), opinion leaders, and on-chain data.
In January 2025, platforms introducing AI agents capable of analyzing trends and refining investment strategies appeared. These rely on a battery of information specific to the crypto market to help investors with their investment strategies.
Other projects, such as Fetch.ai, aim to create autonomous agents equipped with crypto wallets that apply investment strategies. In this way, AI optimizes asset allocation, manages risk, and adjusts positions according to market conditions, often across multiple Blockchains and DeFi platforms. For their part, crypto miners are also moving into the AI era. For example, Galaxy Digital recently converted an entire mine hitherto dedicated to mining to AI model training.
As you can see, we're not done talking about AI yet. Indeed, their emergence is set to redefine many aspects of our society. And it will also have a major impact on the crypto ecosystem, opening the door to new uses that we can't even imagine today.