Some of the hardest problems in AI are that 1) we’re “running out of data” 2) AI doesn’t “financially scale” 3) agents can go rogue and 4) AIs “catastrophically forget” when we try to change them. These problems may seem hard, but we just need to look at them in the right way. So, let’s talk about it. First, we are not running out of data. The data has never been fully used. AI leaders believe they are running out of data, so they are trying to make models generate new data and train on…