Expert analyses, academic papers, and news articles

OpenAI Advances Towards AGI with O1 Model, Expected to Automate Jobs by 2025

Sep 23, 2024, 04:35 PM

OpenAI is making significant strides towards achieving Artificial General Intelligence (AGI). Noam Brown, an AI researcher at OpenAI, suggests that AGI potential is becoming more apparent as models begin to outperform humans in areas such as mathematics, coding, Go, and chess. OpenAI recently unveiled the O1 model, also known as Project Strawberry/Q, which represents a new level of AI capable of thinking and reasoning before responding. Dane Vahey from OpenAI states that while GPT-3 was as smart as a 4th grader and GPT-4 reached high school level, the O1 model is capable of performing at a superhuman level, outperforming humans more than 50% of the time. Former employees and other AI labs forecast that AGI could be fully realized within the next 1-4 years, with some projecting even more advanced AI, known as Artificial Superintelligence (ASI), by 2027. OpenAI's AGI is expected to automate most jobs by 2025.