There are reports that a new model from OpenAI’s, Q* (pronounced Q Star), is capable of solving basic math problems, and is being viewed as a potential breakthrough in the quest for Artificial General Intelligence (AGI). This development, as reported by Reuters, coincides with the executive upheaval at OpenAI, including the controversial firing of CEO Sam Altman.
After initial silence, OpenAI acknowledged internally the existence of project Q* and a preceding letter to the board, as confirmed by a spokesperson. Mira Murati, a long-time executive, informed staff about certain media stories without confirming their accuracy.
Insiders at OpenAI see Q* as a significant step in achieving AGI. Initially the reports focus on Q* being able to perform math at a grade-school level, which on the surface sounds trivial, but many believe that this is the first big step into an AI that can learn, improve and even create the data to evolve and train new models. This breakthrough could accelerate the development of models that are not just equal, but potentially superior to human intelligence.
Whether this was the impetus for Altman’s conflict with the board is not known, but commentary on OpenAI’s forums and among AI commentators has fueled speculation about this potential development.