What I'm reading: Sparks of AGI

Today I'm introducing a new feature of this blog. Every week I will highlight something I've read in the previous week that may interest readers of this blog. These will vary from books to academic research papers, online articles, to newspaper stories. There will always be a brief discussion about why the recommendation is interesting. 

Let's start "what I'm reading" with a recently published paper from Microsoft Research titled: "Sparks of Artificial General Intelligence: Early experiments with GPT-4" by Sébastien Bubeck et al. The title explains why I find the paper exciting, but its strength lies in the methodology the researchers use to evaluate the performance of GPT-4. You needn't be a computer scientist or ML specialist to understand this paper. The paper is summarised by the authors in this manner: 

"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. Given the breadth and depth of GPT-4's capabilities, we believe that it could reasonably be viewed as an early (yet still incomplete) version of an artificial general intelligence (AGI) system."

That really sums up why this paper is worth a read. If you are impressed by ChatGPT, you haven't seen anything yet!