In the aftermath of the recent upheaval in OpenAI’s leadership, a final revelation emerged, acting as an epilogue to the chaotic situation: a Reuters report disclosed a supposed groundbreaking development at the startup known as “Q-Star” or “*Q.” This discovery related to AI, was reportedly a catalyst for internal conflicts. OpenAI staff expressed concerns to the board about a “powerful artificial intelligence discovery that they said could threaten humanity.”
The reported Q* program purportedly enabled an AI agent to perform “grade-school-level math,” a significant breakthrough that could have implications for advancing artificial general intelligence (AGI). The undisclosed details surrounding the Q program led to speculation and discussions online about its true nature.
Some speculation suggests a connection to Q-learning, a machine learning (ML) form. Q-learning is a subcategory of reinforced learning (RL), where AI agents learn through trial and error to achieve specific goals. In this context, it’s suggested that the program’s alleged ability to perform math operations may be linked to Q-related reinforced learning.
While some experts express scepticism about AI’s current capability to solve math problems, others caution that it might not necessarily translate into broader AGI breakthroughs even if achieved. As reported by the MIT Technology Review:
Researchers have for years tried to get AI models to solve math problems. Language models like ChatGPT and GPT-4 can do some math, but not very well or reliably. We currently don’t have the algorithms or even the right architectures to be able to solve math problems reliably using AI. Deep learning and transformers (a kind of neural network), which is what language models use, are excellent at recognizing patterns, but that alone is likely not enough.
Wenda Li, an AI lecturer at the University of Edinburgh
