Everybody in the AI space is now talking about Q* and Q-learning. It seems to be what spooked the OpenAI board.
What is it? Q-learning tries to find an optimal policy that defines the best action to take in each state to maximize the cumulative reward over time.
In other words, it is a model that is able to run autonomous agents that build strategies for long term success, incentivized by rewards.
Right now, researchers have been able to make this work in smaller experiments, but if we scale this all the way with multimodality, it can lead to AGI.
Elon Musk has predicted superintelligence to arrive in 5-6 years, maybe he is right? What do you think?
