GPT-4 and the rise of general intelligence

In a recent paper, researchers have investigated an early version of OpenAI's GPT-4, a large language model that demonstrates a remarkable level of general intelligence. This new generation of LLMs is setting new benchmarks in their ability to tackle complex tasks across domains such as mathematics, coding, vision, medicine, and more. The performance of GPT-4 is so strikingly close to human-level that it is being considered an early and incomplete version of artificial general intelligence (AGI).

We contend that (this early version of) GPT-4 is part of a new cohort of LLMs (along with ChatGPT and Google's PaLM for example) that exhibit more general intelligence than previous AI models.

…

We demonstrate that, beyond its mastery of language, GPT-4 can solve novel and difficult tasks that span mathematics, coding, vision, medicine, law, psychology and more, without needing any special prompting. Moreover, in all of these tasks, GPT-4's performance is strikingly close to human-level performance, and often vastly surpasses prior models such as ChatGPT.

As we continue to make strides in AI, it's crucial to reflect on how these leaps in technology might shape our future. The authors of the paper acknowledge the need for a new paradigm beyond next-word prediction, suggesting that we might still be at the beginning of a much larger transformation.

More in this paper.