Mad AI Disease

Large language models require training data sets in order to continuously improve. However, given the rate at which models are growing we are soon going to run out of training data. And synthetic data is not the solution we thought it might be.