The Evolution of AI and Cryptocurrencies
In a reflection on the evolution of artificial intelligence and its intersection with cryptocurrencies, an intriguing comparison emerges between two significant iterations of OpenAI’s language model: GPT-3 and its successor, GPT-4o. Back in April 2022, GPT-3 made a statement about Bitcoin’s all-time high, inaccurately claiming it was around $20,000. This misstep serves as a perfect illustration of the limitations prevalent in earlier AI models, especially when compared to the enhanced capabilities of GPT-4o in May 2025, which accurately noted Bitcoin’s peak at nearly $69,000 in November 2021.
Factors Driving Bitcoin’s Rise
This previously reported high was driven by various factors, including pandemic-related financial stimuli, a surge in institutional investment, and rampant speculation, which fueled an unprecedented market rally. The inaccuracy of GPT-3 highlights how much the field has advanced over a short span of just a few years.
The Journey with GPT-3
The journey with GPT-3 began during its closed beta testing, where observations were made during a week spent engaged in conversations with the AI, an experience colored by the uncertainty of the time as the appointee was recovering from COVID-19. The dialogues traversed numerous subjects, notably Bitcoin, reflecting the AI’s ambition in understanding complex concepts, sounding almost sentient while fundamentally limited in knowledge.
The Transformation of AI Landscape
Three years on, the AI landscape has transformed significantly. While GPT-3’s foresight into the potential convergence of blockchain and AI in fields like healthcare and governance resonates well today, many of its other predictions now appear naïve. For instance, it was not aware of the massive surge in non-fungible tokens (NFTs) or the substantial growth of cryptocurrencies like Solana that occurred post-pandemic.
Innovations in GPT-4o
Contrastingly, GPT-4o boasts a suite of innovations. With a marked reduction in error rates attributed to improved methods such as retrieval-augmented generation and enhanced fine-tuning, it presents factual information in real-time, demonstrating a newfound ability to access and interpret live data. These enhancements allow it to perform tasks that were previously unimaginable, such as citing sources directly and creating reference lists across numerous formats.
Advancements in Human-AI Interaction
Furthermore, advancements in human-AI interaction have transitioned the emphasis from text-based inputs to voice commands, revolutionizing how journalism and information dissemination take place. In an April 2025 interaction, querying GPT-4o yielded sharp and responsive answers on complex subjects like tokenomics in under 300 milliseconds. This speed and sophistication mark a transformative change in editorial processes.
The Future of AI in Journalism
As this technology matures, the potential for AI to reshape journalism remains a hot topic. When posed with a question about Tesla’s Bitcoin purchase in 2022, GPT-4o could not only refute inaccuracies from the past but also provide detailed context referencing real-world events, showcasing its enhanced capabilities over its predecessor.
Human Elements in Storytelling
Reflecting on this transition, if GPT-4o had the chance to address GPT-3, it might express gratitude for laying down the groundwork for further exploration, acknowledging the visionary ideas that still hold value while affirming the significant strides made in AI technology since then.
Despite these advancements, the irreplaceable qualities of human reporters—ethics, intuition, and critical thinking—stand firm as fundamental to the profession. The narrative, reflections, and judgments remain essential human attributes, underscoring that while AI can assist and enhance efficiency, it cannot replace the core of human storytelling.
