In November 2022, the world was taken by storm with the release of OpenAI’s service ChatGPT, a generative AI system. Overnight, one hundred million people started using it, making Sam Altman, the CEO of OpenAI, a household name. This sparked a race among companies to build a better system, with OpenAI itself seeking to outdo its flagship model GPT-4 with a successor, GPT-5. However, despite the initial hype and excitement, it soon became clear that generative AI may not live up to its promises.
At its core, generative AI relies on fill-in-the-blanks technology, also known as “autocomplete on steroids.” While these systems are great at predicting what may sound good or plausible in a given context, they lack a deeper understanding of what they are saying. This has led to issues with “hallucination,” where the system asserts false information without fact-checking itself. As a result, generative AI has been plagued with errors and inaccuracies, making it a poor product in itself.
As the initial excitement and hype surrounding generative AI faded, 2024 became the year of disillusionment. The lack of profits and high operating losses for companies like OpenAI, estimated at $5 billion, have raised doubts about the potential success of generative AI. Customers have also expressed disappointment with the limited capabilities of ChatGPT compared to the high expectations set by the initial hype.
Moreover, the race to build bigger and better language models has resulted in companies reaching similar levels of performance, with no clear competitive advantage. This lack of a “moat” has led to dwindling profits and even price cuts for companies like OpenAI. In a move to stay relevant, Meta has even started offering similar technology for free.
Despite demoing new products, OpenAI has yet to release any major advances that could justify the name GPT-5. If they fail to do so by the end of 2025, the enthusiasm for generative AI may diminish, and the entire field could suffer. As an expert tech journalist, it is clear that generative AI may not live up to its initial promises, and the future of the field remains uncertain.
