Over the last ten years, artificial intelligence has advanced at an exponential rate, with each new wave of advancements opening up intriguing new possibilities for uses and capabilities. The most significant of these waves has definitely been the emergence of a single, all-purpose AI model, such as LLMs, which can be used to a wide range of tasks, including scientific reasoning, picture comprehension, and code development.
The level of realism with which these activities are completed is defining and developing a whole new generation of technological applications. Though considering the possible impact is exciting, this unexpected accomplishment does raise a very unsettling concern regarding the AI environment going forward.
How will AI play out in the future?
The introduction of models like ChatGPT has revolutionized our perception of what artificial intelligence is capable of, as has the growing expense of creating such models.
Have these broad artificial intelligence models been the essential factor behind all notable technological advancements and AI products?
It is now a widely held opinion that broad artificial intelligence models will rule the landscape. But more likely, numerous entities will be involved in advancing the AI ecosystem. Additionally, a multitude of very useful AI systems will surface, apart from a single general AI model. These AI systems will drive new technological advancements in AI and have a complicated structure, utilizing many AI models, APIs, etc.
Specific AI systems rather than all-purpose AI models will be responsible for handling well-defined, high-value workflows.
Read more from Maithra Raghu (Samaya AI), Matei Zaharia (Databricks), Eric Schmidt (Schmidt Futures)
https://maithraraghu.com/blog/2023/does-one-model-rule-them-all/