Today is Januari 1, 2024. I wish all of our customers and returning blog readers a happy new year on our journey towards AGI and beyond! It’s a very exciting Journey, even though the year has only just began. o1 came out.
OpenAI released a preview of its o1 model on September 12, 2024, followed by the full version on December 5, 2024
Wikipedia The o1 model is designed to enhance complex reasoning tasks by spending more time “thinking” before generating responses, making it particularly effective in areas such as science, coding, and mathematics. It did so very well! So good that OpenAI already announced the next iteration of the product: o3 (o2 is already in use by a telecom provider).
Remember how ChatGpt started in 2022? At that time the model 3.5 was estimated to function at the level of a smart Elementary Schooler. Today ChatGpt with o1 is estimated to function at the level of a smart PHD student. Since it’s release it has extensively. been tested and the results are more than encouraging. For example in the field of medical Diagnosis it appears to score better then real Doctors. On average, real Doctors make mistakes in Diagnosis of 50%. o1 outperforms this score by a staggering 90%! A very good example that shows that wer’e heading straight ahead to AGI!
When will it happen? We don’t know yet but the o1 results are telling us that the current rate of new developments should make it possible by. the end of the coming year or the start of the next year 2026!