Ask HN: What would you do with an AI model capable of continuous learning?
Let's say, hypothetically, that you had a model that did not need to re-train in order to incorporate new information in its weights. What would you do with such a model? Understand that it's not the model but the algorithm that's valuable? I wouldn't have the scale or compute to leverage a model capable of continuous learning, but there will be a company or two that will give me a few billion to give them the secret sauce. I’d stop asking it to update session logs have it watch a bunch of stock charts at once, see if it could find edge. turn mercury to gold? Honestly with the latest mythos news I am a bit concerned about when this happens. But I would test it as a CEO. I would use it to create and maintain a human personality identification model. The goal is to define and create a metric I would call distance from self. The use case is to identify, with precision, a scale that has sociopaths on one end and people immune from bias on the other. I don’t care why adults behave the way they do. I just care to identify adults based upon their potential behavior. Certain careers should be protected from intentional corruption.