Lotus 1-2-3 once held a commanding presence in the world of spreadsheets, similar to the position OpenAI's ChatGPT occupies in today’s AI landscape. Yet, as history shows, dominance in the tech sector is a chapter, not the entire story. OpenAI has led the charge with GPT-4 for over a year, but the field is crowded with contenders, each vying to set the new standard.
In this competitive environment, different strategies emerge. Anthropic is targeting organisational applications of LLMs. This reminds me of Microsoft’s strategy that toppled 1-2-3 from its perch in favour of Excel. But this is unfolding in a different era where the distinction between professional and personal use is increasingly blurred, so we'll have to wait and see what happens.
But it would be foolish to assume that OpenAI has been sitting and waiting for everyone to catch up, as Sora, the amazing text-to-video model, recently illustrated. The conversation around AI’s evolution was stirred anew when Lex Fridman asked OpenAI's Sam Altman about the arrival of GPT-5. Altman’s response, “We will release an amazing new model this year. I don't know what we'll call it,” suggests that innovation remains a constant priority for OpenAI.
However, introducing new models is just one facet of the broader evolution reflected in Abernathy and Utterback's diffusion of innovation model. The model suggests that the AI industry is still in an early stage of product innovation. The integration of long-term memory, advanced machine learning, and realistic human-like avatars into LLMs is anticipated in the near future, but the exact trajectory is still unfolding.
Therefore, Organisations that want to adopt AI must maintain a delicate balance between leveraging the current 'out-of-the-box' capabilities and avoiding premature investments in customisations and extensions to models that may quickly evolve or become obsolete.
Comentarios